PMID int64 35.7M 36.4M | Title/Abstract stringlengths 22 8.78k | MeshTerms stringlengths 7 101 | SemanticTypes stringlengths 4 64 |
|---|---|---|---|
36,236,751 | studi classif metal object fluxgat magnetomet cube structur war unexplod bomb remain underground faulti bomb serious threaten safeti peopl abil accur identifi target crucial subsequ mine work deep learn algorithm use recogn target signific improv recognit accuraci compar tradit recognit algorithm measur magnet moment target includ geomagnet angl paper resnet base recognit system present classifi metal object type first fluxgat magnetomet cube arrang structur fmcas magnet field featur collector construct util eight fluxgat magnetomet sensor array structur provid mm separ sensit unit magnet field data acquir along east west survey line northern side measur target use fmcas next locat type target modifi creat databas magnet target model increas divers train dataset experiment dataset construct construct magnet flux densiti tensor matrix final enhanc resnet use train data classif recognit recogn accord test find valid set group method recognit accuraci percent recognit accuraci rate percent recal rate percent precis rate percent target largest magnet moment best recognit impact experiment find demonstr enhanc restnet network effici classifi metal item provid new idea underground metal target identif classif | D000465;D016571 | T170 |
36,236,750 | optic strain gaug prototyp base high sensit balloon like interferomet addit manufactur optic strain gaug base balloon like interferomet structur form bent standard singl mode fiber combin printer piec present demonstr use measur displac interferomet simpl compact size easi fabric low cost repeat sensor base interfer core clad mode caus fiber curvatur light propag curv balloon shape interferomet region portion releas core limit coupl clad balloon axial displac result artwork construct sensor head sandwich two cantilev displac dimens associ micro bend alter sensor respons function displac determin use wavelength shift intens chang interrog techniqu therefor optic strain gaug good option applic structur displac need examin sensor present sensit nm displac measur rang mm strain sensit pm | D005336;D007368 | T059;T090 |
36,236,749 | robot non destruct test non destruct test ndt evalu nde common refer vast group analysi techniqu use civil medic industri sector evalu properti materi tissu compon structur without caus damag | D065287;D012371 | T061;T090 |
36,236,747 | framework inclus unmodel contact task dynam industri robot paper present method includ unmodel dynam load robot end effector algorithm collis detect general understand robot oper context approach reli applic previous develop modif dynam time warp algorithm well univers applic algorithm identifi kinemat paramet entir process appli arbitrari robot configur requir identif dynam paramet paper address two main categori contact task unmodel dynam determin base whether extern contact forc consist profil end effector base coordin conclus repres exampl analys paper applic task load manipul press bend crimp first type forc applic drill screwdriv snap fit bolt rivet assembl latter categori result present paper base realist test measur obtain industri robot | D012371 | T090 |
36,236,745 | noncontact sleep heartrat monitor method use continu wave doppler radar base differ quadrat sum demodul search algorithm continu wave doppler radar advantag simpl structur low cost low power consumpt attract extens attent detect human vital sign howev respir heartbeat signal mix echo phase amplitud differ two signal larg becom difficult measur heartrat hr interfer respir stabli accur paper differ quadrat sum demodul method propos accord mix characterist respir heartbeat demodul heartbeat featur extract help easi detect breath signal combin constrain nearest neighbor search algorithm realiz sleep hr monitor overnight without bodi movement restraint consid differ vital sign characterist differ individu irregular sleep movement h sleep data nine night collect three subject compar ecg base hr refer equip exclud period bodi turn hr error within time experi confirm method tool long term hr monitor play import role sleep monitor smart elder care smart home | D011825;D012815 | T066;T073 |
36,236,744 | environ adapt object detect framework autonom mobil robot object detect essenti function mobil robot allow carri mission effici recent year various deep learn model base convolut neural network achiev good perform object detect howev case robot carri mission particular environ util model train without consid environ robot must conduct task degrad object detect perform lead fail mission poor model accuraci occur class imbal problem occurr frequenc object class train dataset signific differ studi propos systemat solut solv class imbal problem train multipl object detect model use model effect robot move various environ carri mission moreov show experi propos multi model base object detect framework environ context awar effect overcom class imbal problem result experi cpu usag decreas latenc decreas object detect accuraci increas averag | D012371 | T090 |
36,236,743 | shot fine grain imag classif via gnn tradit deep learn method convolut neural network cnn high requir number label sampl case cost obtain label sampl high obtain enough sampl solv problem shot learn fsl use current typic fsl method work well coars grain imag data well fine grain imag classif work proper assess class similar inter class differ fine grain imag work fsl framework base graph neural network gnn propos fine grain imag classif particular use inform transmiss gnn repres subtl differ differ imag moreov featur extract optim method meta learn improv classif experi three dataset cifar cub dog shown propos method yield better perform indic propos method feasibl solut fine grain imag classif fsl | D016571 | T170 |
36,236,742 | data driven predict control exoskeleton hand rehabilit subspac identif studi propos control method data driven predict control ddpc hand exoskeleton use activ passiv resist rehabilit ddpc model free approach base past system data one strength ddpc constraint state ad control perform control design featur control algorithm elimin essenti problem rehabilit robot term easi custom safe repetit rehabilit task plan within certain constraint experi carri design hand rehabilit system repetit various therapi task real time experi result demonstr feasibl effici propos control approach rehabilit system | D000067190;D012371 | T090;T074 |
36,236,739 | deep learn encrypt traffic classif unknown data detect despit widespread use encrypt techniqu provid confidenti internet communic mobil devic user still suscept privaci secur risk paper novel deep neural network dnn base user activ detect framework propos identifi fine grain user activ perform mobil applic known app activ snif encrypt internet traffic stream one challeng countless applic practic imposs collect train dnn model use possibl data therefor work exploit probabl distribut dnn output layer filter data applic consid model train ie unknown data propos framework use time window base approach divid traffic flow activ segment app activ identifi observ fraction activ relat traffic test shown dnn base framework demonstr accuraci identifi previous train app activ averag accuraci identifi previous untrain app activ traffic unknown data framework employ | D000077321;D063731 | T170;T066 |
36,236,738 | hyperspectr face recognit adapt parallel svms partial hidden face scenario hyperspectr imag open new opportun mask face recognit via discrimin spectral inform obtain hyperspectr sensor work present novel algorithm extract facial spectral featur differ region interest perform comput vision techniqu hyperspectr imag particular histogram orient gradient appli algorithm uwa hsfd dataset extract facial spectral featur set parallel support vector machin custom kernel base cosin similar euclidean distanc train fli classifi unknown subjectsfac accord distanc visibl facial spectral featur ie region conceal face mask scarf result draw optim trade recognit accuraci compress ratio accord facial region occlud | D000066499 | T041 |
36,236,737 | wearabl sensor technolog predict core bodi temperatur systemat review heat relat ill rang heat exhaust heatstrok affect thousand individu worldwid everi year character extrem hyperthermia core bodi temperatur cbt usual gt c declin physic athlet perform cns dysfunct eventu multiorgan failur measur cbt shown predict heat relat ill sever current measur method practic use high acuiti high motion set due invas obstruct natur excess cost noninvas predict cbt use wearabl technolog predict algorithm offer potenti continu cbt monitor earli intervent prevent hri athlet militari intens work environ thus far lack peer review literatur assess efficaci wearabl devic predict analyt predict cbt mitig heat relat ill systemat review identifi studi repres total distinct algorithm predict core bodi temperatur use wearabl technolog high accuraci predict note algorithm meet clinic valid standard algorithm incorpor individu environment data core bodi temperatur predict algorithm despit known impact individu health situat environment factor cbt robust machin learn method offer abil develop accur reliabl person cbt predict algorithm use wearabl devic includ addit data user characterist workout intens surround environ integr interoper cbt predict algorithm exist heat relat ill prevent treatment tool includ heat indic wbgt athlet manag system electron medic record prevent hri increas avail speed data access critic heat event improv clinic decis make process athlet trainer physician sport scientist employ militari offic | D018882;D018883;D000076251 | T073;T037 |
36,236,736 | adapt filter matern respir signal attenu uterin electromyogram electrohysterogram ehg uterin muscl electromyogram record abdomin surfac pregnant non pregnant woman matern respir electromyograph signal mr emg one relev interfer present ehg alvarez alv wave compon ehg indic potenti preterm term birth predict mr emg compon ehg repres issu regard alv wave applic pregnanc monitor instanc preterm birth predict subject great research interest therefor alv wave denois method design includ interfer mr emg attenu without compromis origin wave adapt filter properti make suitabl task howev select optim adapt filter paramet import task success filter oper work algorithm present automat adapt filter paramet select use synthet data filter select pool compris sixteen candid wiener recurs least squar rls household recurs least squar hrls qr decomposit recurs least squar qrd rls best perform optim paramet l filter length forget factor last three develop optim algorithm may interest applic optim filter appli real data result attenu mr emg alv wave power wiener filter power reduct quartil median quartil found respect p valu | D047928;D012815 | T046;T066 |
36,236,735 | spiral deploy optic fiber sensor distribut strain measur seven wire twist steel cabl post tension precast concret bar time monitor condit assess steel cabl provid mission critic data inform decis make ensur structur safeti post tension concret structur studi aim develop spiral deploy scheme distribut fiber optic sensor dfos monitorassess post tension forc seven wire twist steel cabl base puls pre pump brillouin optic time domain analysi dfos place spiral shape two surfac wire steel cabl glu steel cabl epoxi imag observ conduct investig entir bond condit optic fiber steel wire eight concret bar specimen cast pre embed plastic metal duct center post tension steel strand duct instrument two strain two temperatur sensor strand loadedunload monitor measur brillouin frequenc shift correl appli strain result cabl forc temperatur compens maximum minimum averag cabl forc integr measur stain data compar valid load cell maximum averag cabl forc linear relat ground truth data less error initi slack remov test setup post tension forc loss bound approxim use test setup design studi | D055100;D013232 | T073;T122 |
36,236,733 | sensor acquisit alloc real time monitor articul construct equip digit twin visibl avail equip oper dynam construct site often block various obstacl materi temporari perman facil equip worker equip monitor real time digit twin thus play crucial role accid prevent paper develop scalabl technic approach present prototyp applic framework transmit real world sensor data updat equip model insid graphic digit twin concurr visual monitor construct oper develop framework workflow extend visual construct oper occur insid dynam world simpli outfit real equip appropri sensor connect virtual counterpart implement proof concept interfac describ context real time digit twin assist excav oper prevent unintend strike underground util experi valid propos technic approach simul real time motion backho loader articul arm use orient sensor instal boom stick bucket describ experiment result character scope potenti reason spatio tempor discrep occur monitor real oper replic digit twin effect oper warn mechan base preset safeti threshold also investig describ | D017132 | T082 |
36,236,732 | wearabl sensor learn enhanc higher educ wearabl sensor tradit use measur monitor vital human sign well healthcar applic howev grow interest use deploy technolog facilit teach learn particular higher educ environ aim paper therefor systemat review rang wearabl devic use enhanc teach deliveri engin curricula higher educ moreov compar advantag disadvantag devic accord locat worn human bodi accord survey wearabl devic enhanc learn main worn head eg eyeglass wrist eg watch chest eg electrocardiogram patch fact among locat head worn devic enabl better student engag learn materi improv student attent well higher spatial visual awar identifi research question discuss research inclus exclus criteria present challeng face research implement learn technolog enhanc engin educ furthermor provid recommend use wearabl devic improv teach learn engin cours higher educ | D000076251 | T073 |
36,236,731 | autonom mutual authent protocol edg network distinct secur protocol necessari exponenti growth intellig edg devic particular autonom devic need address signific secur concern function smooth high market demand nevertheless exponenti increas connect devic made cloud network complex suffer inform process delay therefor goal work design novel server less mutual authent protocol edg network aim demonstr autonom mutual authent amongst connect smart devic within edg network solut address applic autonom car smart thing internet thing iot devic edg wireless sensor network wsn etc paper design propos use public key system octet base balanc tree transit challeng respons mechan devic uniqu id uid pseudo random number generat prng time stamp event specif session key ultim server less design requir less infrastructur avoid sever type network base communic attack eg imperson man middl mitm iot ddos etc addit system overhead elimin secret key requir result provid suffici evid protocol market competit demonstr better benchmark comparison result | D016494;D000080487 | T078;T170;T057;T097;T080 |
36,236,730 | preliminari evalu system bodi aerial sensor monitor work dog paper present system behavior environment physiolog monitor work dog use bodi aerial sensor proof concept studi present includ two train dog perform nine scent detect task uncontrol environ encompass approxim two acr dog outfit custom design wearabl har monitor heart rate activ level skin temperatur util commerci avail micro air vehicl perform aerial sens track terrain movement dog outdoor space dog free explor space work maxim speed complet scent base search retriev task throughout experi har data transfer base station via wi fi real time work also focus test perform custom electrod applic specif ergonom improv adapt filter process techniqu recov much electrocardiographi data possibl high intens motion activ abl recov use collect data observ trend heart rate general increas immedi success target local track dog aerial video footag appli state art deep learn algorithm design onlin object track qualit quantit track result promis studi present initi effort toward deploy bodi aerial sensor monitor work dog environ scent detect search rescu task order ensur welfar enabl novel dog machin interfac allow higher success rate remot autom task perform | D004562;D000085682 | T074;T015;T060 |
36,236,728 | sort center valu identif internet recycl base transfer cluster core link internet recycl process valu identif sort center great challeng due small imbalanc data set paper util transfer fuzzi c mean improv valu assess accuraci sort center transfer knowledg custom cluster ensur transfer effect inter class balanc data select method propos select balanc qualifi subset sourc domain furthermor improv rfm recenc frequenc monetari model name gfmr gap frequenc monetari repeat present attain reason attribut descript sort center consum applic field electron wast recycl show effect advantag propos method | D018505 | T068 |
36,236,727 | applic wearabl comput asr technolog underground mine support mine supervis heavi machineri chamber system use automat speech recognit industri becom popular bring benefit especi case user hand often busi environ allow use keyboard howev accuraci algorithm still big challeng articl describ attempt use asr underground mine industri improv record work heavi machineri chamber foreman particular attent paid factor case negat impact speech recognit influenc environ special mine vocabulari learn curv first foreman workflow document recogn allow select function includ applic dictionari special mine vocabulari sourc databas develop combin string match algorithm aim improv correct speech recognit text mine analysi machin learn method use creat function provid assist regist inform final prototyp applic test mine environ accuraci result present | D013060;D000076251 | T054;T073;T040 |
36,236,726 | optic oxygen sens clark electrod face face biosensor case studi last decad continu competit two method detect concentr dissolv oxygen amerometr clark electrod optic quench phosphoresc porphyrin metal complex obvious advantag disadvantag competit especi acut develop biosensor howev unbias comparison extrem difficult achiev sinc singl detect method use particular studi work microfluid system synchron detect oxygen concentr two method creat purpos direct comparison receptor element repres saccharomyc cerevisia yeast cell adsorb composit materi previous develop scientif group knowledg first work kind comparison oxygen detect method carri direct | D015374;D011166 | T059;T075;T123;T109 |
36,236,725 | eegfnir base workload classif use function brain connect machin learn high demand techniqu estim human mental workload activ product enhanc accid prevent studi focus singl physiolog sens modal use univari method analys multi channel electroencephalographi eeg data paper propos new framework reli featur hybrid eeg function near infrar spectroscopi eeg fnir support machin learn featur deal multi level mental workload classif furthermor instead well use univari power spectral densiti psd eeg record propos use bivari function brain connect fbc featur time frequenc domain three band delta hz theta hz alpha hz assist fnir oxyhemoglobin deoxyhemoglobin hbo hbr indic fbc techniqu signific improv classif perform accuraci back vs back back vs back use public dataset moreov topograph heat map visualis indic distinguish region eeg fnir show differ among back back back test result determin best region assist discrimin mental workload eeg fnir differ specif posterior area perform best posterior midlin occipit poz eeg alpha band fnir superior right frontal region af | D010108;D019265 | T059;T123;T116 |
36,236,724 | onlin detect system wheat machin harvest impur rate base deeplabv wheat one import food crop world usual harvest mechan combin harvest impur rate one import indic qualiti wheat obtain mechan harvest realiz onlin detect impur rate mechan harvest process wheat vision system base deeplabv model deep learn identifi segment wheat grain impur design studi deeplabv model construct consid four backbon mobilenetv xception resnet resnet train optim deeplabv model determin accuraci rate comprehens evalu index averag intersect ratio basi onlin detect method measur wheat impur rate mechan harvest base imag inform construct model realiz onlin detect wheat impur rate test result show resnet best recognit segment perform accuraci rate grain identif comprehens evalu index intersect ratio accuraci rate impur identif comprehens evalu index intersect ratio averag intersect ratio term speed resnet fast segment speed ms per imag therefor studi resnet select backbon network deeplabv carri identif segment mechan harvest wheat grain impur compon base manual inspect result maximum absolut error devic impur rate detect bench test largest relat error maximum absolut error devic impur rate detect field test largest relat error studi provid real time method impur rate measur wheat mechan harvest | D016571;D014908 | T170;T168;T002 |
36,236,721 | residu incept u net riu net approach comparison u shape cnn transform model build segment high resolut satellit imag build segment crucial applic extend map product urban plan nowaday still challeng due cnns inabl model global context transform high memori need studi cnn transform model generat comparison realiz alongsid propos residu incept u net riu net u net residu u net attent residu u net four cnn architectur incept incept resnet xception mobilenet implement encod u net base model last two transform base approach tran u net swin u net also use massachusett build dataset inria aerial imag label dataset use train evalu inria dataset riu net achiev highest iou score f score test accuraci respect massachusett small dataset attent residu u net achiev highest iou f score tran u net reach highest test accuraci massachusett larg dataset residu u net accomplish highest iou f score attent residu u net attain highest test accuraci result show riu net signific success inria dataset massachusett dataset residu u net attent residu u net tran u net provid success result | D007091;D016571 | T170;T066 |
36,236,720 | human human posit estim system use rssi outdoor environ method prevent collis peopl avoid traffic accid receiv signific attent measur posit non line sight nlos area direct visual recogn posit measur method use wireless communic type gps propag characterist radio signal receiv signal strength indic rssi howev convent posit estim method use rssi requir multipl receiv decreas posit estim accuraci owe presenc surround build studi propos system solv challeng use receiv posit estim method base rssi map simul particl filter moreov studi util ble peripheralcentr function capabl advertis transmitterreceiv use advertis radio wave method provid framework estim posit unspecifi transmitt effect propos system evalu studi simul experi actual environ obtain error averag distanc simul show precis propos method actual environ propos method show error averag distanc furthermor evalu accuraci propos method transmitt receiv motion consid move person outdoor nlos area result show error consequ conclud accuraci compar transmitt stationari move compar convent path loss model measur distanc wherea propos method estim posit accuraci outdoor environ addit expect use collis avoid system confirm presenc stranger nlos area | D000465;D011846 | T170;T070 |
36,236,716 | technic develop ceci social robot research present technic consider implement ceci comput electron communic interfac social robot case robot respond need achiev technolog develop emerg countri aim social impact social interact two problem social robot current market main focus research first cost afford compani univers individu emerg countri second design exclus orient function part vision inher engin creat without consid vision prefer requir end user especi social interact last reason end caus avers use type robot respons issu rais low cost prototyp propos start commerci platform research develop use open sourc code robot design present center criteria prefer end user priorit accept social interact articl detail select process hardwar capabl robot moreov program section provid introduc differ softwar packag use adapt social interact main function implement well new origin part propos final list applic current develop robot possibl applic futur research discuss | D012371 | T090 |
36,236,715 | flexibl antenna sub ghz g band comprehens review ever increas demand need high speed communic generat intens research field fifth generat g technolog sub ghz g mid band spectrum focus research due meritori eas deploy current scenario alreadi exist infrastructur g lte system g technolog find applic enorm field requir high data rate low latenc stabl radiat pattern one major sector benefit outbreak g field flexibl electron devic compact need antenna flexibl lightweight conform still excel perform characterist flexibl antenna use wireless bodi area network wban need high conform bent accord differ curvatur human bodi differ bodi part specif absorpt rate sar must permiss level antenna suit wban applic paper give comprehens review current state art flexibl antenna sub ghz g band furthermor paper give key insight materi flexibl antenna paramet consid design flexibl antenna g challeng design implement flexibl antenna g | D000076251;D059015 | T073 |
36,236,714 | low power environment imag sensor remot photogrammetri paper aim prove feasibl monitor solut model tempor monitor sandbar character speci role landscap develop solut allow studi interact river dynam veget use network low resolut low power sensor issu address concern feasibl implement photogrammetri solut use low resolut sensor well choic appropri sensor test accord differ configur imag captur storag sensor andor imag transmiss central node also detail analysi differ phase process camera initi imag captur network transmiss select appropri standbi mode reveal tini low cost board esp cam perform reconstruct propos use camera uxga resolut qualiti render energi consumpt multi node scenario base combin wi fi gsm relay propos studi show sever year autonomi system final illustr energi cost modul defin studi process identifi quantifi one one differ phase oper card better energi optim setup camera configur shoot save sd card send wi fi devic oper deploy allier river franc | D010780 | T060 |
36,236,712 | rsu base onlin intrus detect mitig vanet secur vehicular communic critic factor secur traffic manag effect secur intellig transport system requir effect time intrus detect system id paper consid fals data inject attack distribut denial servic ddos attack especi stealthi ddos attack target integr avail respect vehicular ad hoc network vanet novel machin learn techniqu intrus detect mitig base central communic roadsid unit rsu propos consid attack perform propos method evalu use traffic simul real traffic dataset comparison state art solut clear demonstr superior detect local perform propos method best case worst case achiev level fals alarm probabl | D003142;D003195 | T073;T054;T170;T057;T097;T080 |
36,236,711 | memori visual base malwar detect techniqu advanc persist threat attack campaign intrud team intrud establish long term presenc network mine sensit data becom danger combin polymorph malwar type malwar undetect also generat multipl variant type malwar network remain system main memori avoid detect research employ visual approach base comput memori detect classifi various class malwar howev preprocess step denois malwar imag consid result overfit problem prevent us perfect general model paper introduc new data engin approach compris two main stage denois dimens first aim reduc ideal remov nois malwar memori base dump file transform imag latter process clean imag compress reduc dimension avoid overfit issu lower varianc comput cost memori util built machin learn model implement new data engin approach result show perform metric accuraci precis recal f score obtain new data engin approach machin learn model outperform exist solut accuraci precis recal f score addit comput time memori usag also reduc signific | D016494;D000069550 | T066;T078;T170;T057;T097;T080 |
36,236,710 | optim decis support system base crop dynam model n fertil treatment effici handl nitrogen becom critic issu modern agricultur financi standpoint well regard reduc environment impact use excess amount nitrogen fertil manur compost use maintain rais soil chemic level without excess accumul howev best grain yield combin n fertil via studi aim develop optim decis support system indic initi fertil base nitrogen limit n limit crop growth dynam optim nitrogen fertil n fertil manag system increas crop yield maintain balanc fertil suppli crop demand studi use n limit crop growth model lintul develop optim decis support system work formul resolv two optim challeng maxim biomass growth ii maxim growth least cost paid n fertil applic furthermor two case studi develop base number field optim singl field ii optim multipl field case multipl field hypothes fertil treatment one field leak field affect nitrogen dynam differ field final numer simul carri support theori develop paper simul show propos work employ achiev goal optim nitrogen manag crop increas biomass growth certain scenario attain | D005308;D008372 | T167;T073 |
36,236,709 | thermal evalu multi antenna system propos treat bone tumor finit element analysi microwav ablat common use soft tissu tumor applic bone tumor bare analyz antenna treat bone tissu cm late design bone tumor patholog stage reach cm wide antenna cover therefor goal evalu thermal perform multi antenna array linear triangular squar configur doubl slot ds monopol mtm antenna evalu parametr studi finit element method variat distanc antenna ad bone thick bt implement array feasibl evalu swr ablat tissu volum etc linear configur ds mtm antenna show swr ad mm mm bt mm mm ad mm mm bt mm mm respect triangular show swr ad mm mm bt mm mm ad mm mm bt mm mm squar configur ds generat swr ad mm mm bt mm mm mtm swr ad mm bt mm mm ablat tissu cm cm min accord treatment time array configur maximum temperatur ablat tissu modifi bone tumor gt cm treat antenna array | D001859;D008872 | T191;T057;T070 |
36,236,708 | motion sensor knee angl recognit muscl rehabilit solut progress loss function capac due age serious problem compromis human locomot capac requir help assist reduc independ nanostim project aim develop system capabl perform treatment electrostimul patient home reduc number consult knee angl one essenti attribut context help understand patient movement treatment session articl present wearabl system recogn knee angl imu sensor hardwar chosen wearabl low cost includ esp microcontrol mpu sensor howev hardwar impair signal accuraci multitask environ expect rehabilit treatment three optim filter algorithm complex test improv signal nois complementari filter obtain best result present averag error degre improv mse furthermor interfac mobil app develop respond immedi recogn movement system test volunt real environ could success measur movement perform futur plan use recogn angl electromyographi sensor | D000076251 | T073 |
36,236,707 | method medic microscop imag sharp evalu base nsst varianc combin time frequenc domain algorithm sharp evalu microscop imag base non subsampl shearlet wave transform nsst varianc propos present studi purpos improv nois immun accuraci microscop imag autofocus first imag decompos nsst algorithm decompos sub band imag subject varianc obtain energi sub band coeffici final evalu valu obtain ratio energi high low frequenc sub band coeffici experiment result show propos algorithm deliv better nois immun perform method review studi maintain high sensit | D000465;D007091 | T170;T066 |
36,236,704 | secur onlin fingerprint authent system industri iot devic g network develop g network rapid increas use industri internet thing iiot devic control monitor process purpos biometr base user authent prevent unauthor access iiot devic therebi safeguard data secur product howev biometr authent system iiot templat protect thus risk raw biometr data store templat central databas iiot devic moreov tradit biometr authent face slow limit databas hold capac data transmiss problem address issu paper propos secur onlin fingerprint authent system iiot devic g network core propos system design cancel fingerprint templat protect origin minutia featur provid privaci secur guarante entiti user messag content transmit iiot devic cloud server via g networkscompar state art method propos authent system show competit perform six public fingerprint databas save comput cost achiev fast onlin match | D000465;D056667 | T170;T066 |
36,236,703 | compar studi frequent pattern mine trajectori data sequenti pattern mine spm major class data mine topic wide rang applic continu uncertain natur trajectori data make distinct differ typic transact data requir addit data transform prepar spm howev littl research focus compar perform spm algorithm applic context trajectori data studi select repres sequenti pattern mine algorithm evalu various paramet understand effect involv paramet perform studi result sequenti pattern runtim ram consumpt context taxi trajectori dataset drive dataset demonstr work method discret trajectori data differ spm algorithm perform trajectori databas result visual actual beij road map reflect traffic congest condit result demonstr contigu constraint base algorithm could provid concis represent output sequenc function low minsup balanc ram consumpt execut time studi use guid academ profession determin suitabl spm algorithm applic involv trajectori data | D000465;D057225 | T170;T066 |
36,236,701 | framework effici auto scale virtual network function cloud environ network function virtual nfv offer altern method design deploy manag network servic nfv decoupl network function dedic hardwar move virtual server run softwar one major strength nfv abil dynam extend reduc resourc alloc virtual network function vnf need run time need comprehens meter compon cloud store process metricssampl effici auto scale load manag vnf paper propos integr framework effici auto scale vnf use gnocchi time seri databas integr within framework store handl index time seri data object studi valid efficaci employ gnocchi auto scale vnf term aggreg data point databas size data recoveri speed memori consumpt employ methodolog perform detail empir analysi propos framework deploy fulli function cloud implement nfv architectur use sever openstack compon includ gnocchi result show signific improv legaci ceilomet configur term lower meter storag size less memori util process manag metric reduc time delay retriev monitor data evalu alarm auto scale vnf | D003201;D012984 | T090;T170;T073 |
36,236,700 | two mode underwat smart sensor object precis aquacultur base aiot technolog monitor status cultur fish essenti task precis aquacultur use smart underwat imag devic non intrus way sens monitor freeli swim fish even turbid low ambient light water paper develop two mode underwat surveil camera system consist sonar imag devic stereo camera sonar imag devic two cloud base artifici intellig ai function estim quantiti distribut length weight fish crowd fish school sonar imag noisi fish instanc overcrowd fish school often overlap machin learn technolog mask r cnn gaussian mixtur model convolut neural network semant segment network employ address difficulti analysi fish sonar imag furthermor sonar stereo rgb imag align space offer addit ai function fish annot base rgb imag propos two mode surveil camera test collect data aquacultur tank shore net cage use cloud base aiot system accuraci propos ai function base human annot fish metric data set test verifi feasibl suitabl smart camera estim remot underwat fish metric | D001185;D016571 | T090;T170;T066;T073 |
36,236,699 | sealid saimaa ring seal identif dataset wildlif camera trap crowd sourc imag materi provid novel possibl monitor endang anim speci massiv data volum call automat method solv various task relat popul monitor identif individu anim saimaa ring seal pusa hispida saimensi endang subspeci found lake saimaa finland one exist freshwat seal speci ring seal perman pelag pattern uniqu individu use identif individu larg variat pose exacerb deform natur seal togeth vari appear low contrast ring pattern rest pelag make saimaa ring seal identif task challeng provid good benchmark evalu state art identif method therefor make saimaa ring seal imag sealid dataset n public avail research purpos paper dataset describ evalu protocol identif method propos result two baselin method hotspott norppa provid sealid dataset made public avail | D046048 | T015 |
36,236,697 | design implement smart insol system measur plantar pressur temperatur intellig insol system may monitor individu foot pressur temperatur real time comfort home help captur foot problem earliest stage constant monitor foot complic essenti avoid potenti devast outcom common diseas diabet mellitus inspir goal author work propos full design wearabl insol detect plantar pressur temperatur use shelf sensor design provid detail specif temperatur pressur sensor circuit configur character sensor design consider creat small system suitabl electron procedur also detail use low power communic protocol data individu foot pressur temperatur may sent wireless central devic storag research may aid creation afford practic portabl foot monitor system patient solut use continu home monitor foot problem pressur pattern temperatur differ two feet generat map use earli detect diabet foot complic help artifici intellig | D001185;D017719 | T090;T047;T066;T073 |
36,236,694 | classif eeg use adapt svm classifi csp onlin recurs independ compon analysi effici featur extract method two class electroencephalographi eeg demonstr use common spatial pattern csp optim spatial filter howev effect artifact non stationari uncertainti pronounc csp filter use furthermor tradit csp method lack frequenc domain inform requir mani input channel therefor overcom shortcom featur extract method base onlin recurs independ compon analysi orica csp propos eeg base brain comput interfac bcis especi onlin real time bcis wide use classifi use linear discrimin analysi lda support vector machin svm previous evalu clear show svms general outperform classifi term perform case adapt support vector machin svm use classif togeth orica csp method result promis experi perform eeg data class motor imag name dataset bci competit iv | D062207;D060388 | T074;T066;T081 |
36,236,693 | comparison retinanet base singl target cascad multi target detect model administr region network map pictur critic need detect administr region network map pictur map censorship task implement target detect technolog howev map imag tend numer administr region overlay map annot symbol thus make difficult accur detect region use retinanet base target detect model integr resnet featur pyramid network fpn studi built multi target model singl target cascad model three singl target model take taiwan tibet chines mainland target exampl two model evalu classif local accuraci investig administr region detect perform result show singl target cascad model abl detect administr region higher fscore map compar multi target model respect furthermor locat box size distribut singl target cascad model look similar manual annot box size signifi propos cascad model superior multi target model studi promis provid support comput map read intellig map censorship | D007360 | T041 |
36,236,690 | multi connect g network beyond survey manag grow number user ever increas demand bandwidth futur th generat g cellular network combin differ radio access technolog cellular satellit wifi among other differ type equip pico cell femto cell small cell macro cell etc multi connect emerg paradigm aim leverag heterogen architectur achiev multi connect propos enabl ue simultan use compon carrier differ heterogen network node base station wifi access point etc could offer mani benefit term qualiti servic energi effici fair mobil spectrum interfer manag therefor survey aim present overview multi connect g network beyond comprehens review exist standard enabl technolog propos taxonomi defin classifi differ element character multi connect g futur network thereaft exist research work use multi connect improv qualiti servic energi effici fair mobil manag spectrum interfer manag analyz compar addit lesson common differ context present final open challeng multi connect g network beyond discuss | D000383 | T090 |
36,236,689 | interoper digit twin ieee standard shop floor factori floor area insid factori manufactur product execut digitalis area increas last year introduc digit twin dt industri concept dt digit represent real object entir system dt includ high divers compon differ vendor need interact effici case develop standard protocol consid need oper standard protocol caus interoper issu transduc sensor actuat use communic layer exchang inform digit contra part reason communic layer one relev aspect develop paper cover dt develop go physic visualis layer refer architectur model standard protocol focus interoper reach syntact level communic ieee iec standard semant communic layer connect transduc devic digit represent achiev semant level interoper communic layer add semant communic process allow develop interoper dt base ieee standard dt present reach syntact semant level interoper allow monitor visualis prototyp system | D012660 | T078 |
36,236,688 | height nonlinear veloc field algorithm cor station base garch model studi basic concept height nonlinear veloc field model cor station describ nois result larg deviat observ predict height arch test method heteroscedast cor height residu squar seri propos non stationari characterist cor height residu squar time seri prove cor height nonlinear veloc field reconstruct method base garch model propos first nonlinear ls period fit model establish cor height seri data garch model establish fit non stationari residu seri final signal term linear trend term garch model nois term nonlinear ls model combin reconstruct nonlinear veloc field cor height rmse nonlinear ls cycl model cor station worldwid rang mm differ veloc approxim annual semi annual amplitud sopac result mma mm mm respect compar centimet amplitud cor station height accuraci nonlinear model establish studi met requir result height nonlinear veloc field reconstruct cor station worldwid show mean squar error predict one year height movement reach mm averag predict accuraci semi annual mm compar calcul accuraci current global cor elev compon mm predict error studi mm expect goal achiev regard accuraci cor station height nonlinear veloc field model | D000465;D017711 | T170;T077 |
36,236,687 | analysi secur issu wireless bodi area network heterogen network bodi area network ban one import techniqu observ patient health real time identifi analyz diseas effect implement technolog practic benefit key issu address among issu secur high critic wban oper cooper network model multipl network homogen network purpos perform reliabl heterogen network purpos data transfer process applic point view network network hospit clinic medic expert etc patient himselfherself may move one network anoth paper bring issu relat secur wban separ network well multipl network wban work separ network ieee standard consid wban work multipl network especi heterogen network secur issu consid consid advanc artifici intellig ai paper describ ai address challeng face wban paper describ possibl approach taken address issu model secur mechan use various artifici intellig techniqu paper propos game theori stackelberg secur equilibrium gtsse model secur heterogen network wban describ experi conduct author result prove suitabl model use gtsse | D001185;D059015 | T090;T066;T073 |
36,236,684 | systemat literatur review meta analysi scalabl blockchain base electron vote system electron vote system must find solut various issu authent data privaci integr transpar verifi hand blockchain technolog offer innov solut mani problem scalabl blockchain arisen fundament barrier realiz promis technolog especi electron vote studi seek highlight solut regard scalabl blockchain base electron vote system issu link also attempt forese futur develop systemat literatur review slr use complet task lead select articl english languag januari march famous databas slr conduct identifi well known propos implement verif method various cryptograph solut previous research evalu cost time also identifi perform paramet primari advantag obstacl present differ system common approach blockchain scalabl addit outlin sever possibl research avenu develop scalabl electron vote system base blockchain technolog research help futur research propos develop solut keep mind vote requir merit demerit propos solut provid guidelin scalabl vote solut | D000081186 | T066 |
36,236,683 | soft wearabl robot develop status technic challeng recent year research begun focus flexibl lightweight design wearabl robot process mani novel concept achiev continu made shown public new problem emerg time need solv paper give overview develop status soft wearabl robot human movement assist basi clear definit perform system classif accord target assist joint attempt describ overal prototyp design level relat field addit necessari sort latest research progress key technolog structur actuat control evalu therebi analyz design idea basic characterist final discuss possibl applic field propos main challeng valuabl research direct | D012371;D000076251 | T090;T073 |
36,236,680 | hybrid carbon nanotubesgold nanoparticl composit trace nitric oxid detect wide rang humid composit function singl wall carbon nanotub swcnts gold nanoparticl au nps nm diamet drop cast print circuit board pcb substrat equip interdigit electrod make hybrid thin film addit au nps decor surfac swcnts network film act catalyst result enhanc sensit low ppb concentr detect limit composit film character scan electron microscop sem swcnts cluster load various amount au nps rang weight effect nitric oxid sensit studi optim optim composit film test air nitrogen environ well wide relat humid rang sensor also test select expos various gase nitrous oxid ammonia carbon monoxid sulfur dioxid aceton sensit found much higher test gase advantag sensor sensit low ppb level ppb estim respons time within recoveri time around min excel reproduc sensor sensor work within wide rang relat humid | D053768;D037742 | T104;T073 |
36,236,679 | first arriv pick zero phase seismic data hilbert envelop empir half window heeh method first arriv travel time pick import step mani seismic data process applic first arriv pick method search sudden jump seismic energi trace onset clear appropri minimum phase data paper propos method first arriv pick non minimum phase data base complex trace analysi hilbert integr transform generat complex seismic trace follow extract envelop first arriv identif introduc outlier detect method use wide use three sigma rule thumb common use softwar algorithm identifi outlier propos method ultim generat logic window one locat outlier zero elsewher first arriv select middl first outlier window test propos method zero phase synthet data ad random nois method detect true first arriv accur furthermor test real vibrosei data show method recogn first arriv accuraci within millisecond correspond arriv time manual pick experienc geophysicist | D000465;D012984 | T090;T170;T073 |
36,236,678 | deep metric learn use negat sampl probabl anneal multipl studi conclud select input sampl key deep metric learn triplet network select anchor posit negat pair refer triplet mine select negat consid complic task due larg number possibl goal select negat result posit triplet loss howev multipl approach semi hard negat mine hardest mine well known addit random select sinc introduct semi hard mine proven outperform negat mine techniqu howev recent year select call hardest negat shown promis result differ experi paper introduc novel negat sampl solut base dynam polici switch refer negat sampl probabl anneal aim exploit posit approach result valid experiment synthet dataset use cluster analysi method final discrimin abil train model measur real life data | D000465 | T170 |
36,236,676 | multi object locat map base deep learn visual slam simultan local map slam technolog use locat build map unknown environ construct map often suffer poor readabl interact primari secondari inform map accur grasp intellig robot interact meaning way environ must understand geometr semant properti scene surround propos method reduc absolut posit error ape improv posit perform system also construct object orient dens semant point cloud map output point cloud model object reconstruct object indoor scene fact eight categori object use detect semant map use coco weight experi object actual scene reconstruct theori experi show number point point cloud signific reduc averag posit error eight categori object technic univers munich tum dataset small absolut posit error camera also reduc introduct semant constraint posit perform system improv time algorithm segment point cloud model object environ high accuraci | D000077321;D021621 | T073;T062;T066;T060 |
36,236,675 | transform base binocular dispar predict occlus predict novel full connect layer depth estim algorithm base convolut neural network mani limit defect construct match cost volum calcul dispar use limit dispar rang authent dispar beyond predetermin rang acquir besid match process lack constraint occlus match uniqu also local featur extractor convolut neural network lack abil global context inform percept aim problem match method construct match cost volum propos dispar predict algorithm base transform specif compris swin spp modul featur extract base swin transform transform dispar match network base self attent cross attent mechan occlus predict sub network addit propos doubl skip connect fulli connect layer solv problem gradient vanish explos train process transform model thus enhanc infer accuraci propos model paper achiev epe absolut error pe percentag error greater px kitti kitti dataset respect infer time paramet low show great advantag compar algorithm various evalu metric | D016571;D015357 | T170;T081 |
36,236,674 | robust end end deep learn base approach effect reliabl btd use mr imag detect brain tumor earli stage critic clinic practic surviv rate brain tumor aris multipl shape size featur various treatment option tumor detect manual challeng time consum prone error magnet reson imag mri scan most use tumor detect due non invas properti also avoid pain biopsi mri scan one patient brain generat mani imag multipl direct make manual detect tumor difficult error prone time consum therefor consider need autonom diagnost tool detect brain tumor accur research present novel tumorresnet deep learn dl model brain detect ie binari classif tumorresnet model employ convolut layer leaki relu lrelu activ function featur map activ comput distinct deep featur final three fulli connect classif layer use classifi brain tumor mri normal tumor perform propos tumorresnet architectur evalu standard kaggl brain tumor mri dataset brain tumor detect btd contain brain tumor normal mr imag propos model achiev good accuraci btd experiment result includ cross dataset set valid superior tumorresnet model contemporari framework studi offer autom btd method aid earli diagnosi brain cancer procedur substanti impact improv treatment option patient surviv | D001932;D000077321 | T191;T066 |
36,236,673 | air firmwar updat constrain nb iot devic internet thing iot deploy provid smart solut build logist hospit mani grow billion connect devic howev tremend growth mainten support hidden burden devic deploy iot general light microcontrol low power low memori lightweight softwar softwar includ firmwar applic manag remot via wireless connect improv flexibl instal time access effect cost firmwar updat constant remov known bug improv function devic work present approach updat firmwar air ota constrain iot devic use narrowband iot nb iot wireless communic standard communic manag server devic nb iot one promis low power wide area lpwa network protocol support k devic within cell use licens spectrum work proof concept demonstr usag nb iot updat firmwar constrain devic also calcul overal power consumpt latenc differ size firmwar | D000080487 | T170 |
36,236,672 | use shear wave elastographi assess exercis induc muscl damag review shear wave elastographi method increas use assess muscl stiff clinic practic human health research recent shear wave elastographi suggest use assess exercis induc muscl damag review aim summar current knowledg util shear wave elastographi assess muscl damag general literatur support shear wave elastographi promis method assess muscl damag increas shear modulus report immedi sever day eccentr exercis studi use shear wave elastographi endur event show mix result moreov seem shear modulus increas relat declin voluntari strength loss recommend shear modulus measur multipl muscl within muscl group prefer longer muscl length studi need confirm disrupt calcium homeostasi seem primari candid under mechan explain increas shear modulus observ eccentr exercis remain investig well chang shear modulus correl direct assess amount muscl damag biopsi | D054459 | T060 |
36,236,669 | sudoscan innov simpl non invas medic devic assess sudomotor function diabet autonom neuropathi probabl undiagnos serious complic diabet main object assess preval peripher autonom neuropathi popul diabet patient analyz real life outpati unit scenario determin feasibl perform sudoscan test togeth wide use test neuropathi total patient includ studi differ score system toronto clinic neuropathi score tcns neuropathi disabl score nds neuropathi symptom score nss appli record diabet neuropathi dn sudoscan medic devic use assess sudomotor function detect diabet autonom neuropathi screen cardiac autonom neuropathi fifteen patient sudomotor dysfunct sudoscan risk score posit correl hand electrochem sweat conduct esc diastol blood pressur dbp level glycat hemoglobin well tcns nds nss perform sudoscan test togeth test dn prove feasibl approach could use daili clinic practic order screen dn well earli screen complex time consum test | D003929 | T047 |
36,236,667 | fast design optim compar analysi linear perman magnet motor magnet skew auxiliari tooth overhang structur paper present fast design optim use effect characterist analysi linear perman magnet motor lpmms techniqu improv motor perform use auxiliari tooth perman magnet pm skew overhang structur techniqu differ effect characterist lpmm depend combin result complex design optim process particular three dimension effect pm skew overhang structur take lot time analyz deal problem effect magnet field analysi method novel optim algorithm propos preferenti field reconstruct method use fast accur evalu magnet field lpmm propos magnet field analysi method chang magnet field distribut due addit auxiliari tooth predict magnet field effect pm skew overhang structur consid reduc comput burden magnet field analysi electromagnet characterist lpmms calcul quick detent forc end forc thrust forc back emf effect auxiliari tooth overhang structur optim pm skew length investig compar studi result subsequ propos optim algorithm advantag reduc time cost provid multimod optim robust evalu local peak time propos method verifi via comparison finit element analysi experiment result | D000465;D059346 | T074;T170;T073 |
36,236,663 | multimod finger puls wave sens comparison forcecardiographi photoplethysmographi sensor puls wave pws mechan wave propag ventricl whole vascular system brisk enlarg blood vessel lumen caus sudden increas local blood pressur photoplethysmographi ppg one widespread techniqu employ pw sens due abil measur blood oxygen satur sensor techniqu propos record pws includ applan tonomet piezoelectr sensor forc sensor differ kind acceleromet perform sensor analyz individu result found good agreement eg term pw morpholog physiolog paramet extract comparison led deeper comprehens strength weak ultim consider multimod approach accomplish via sensor fusion would lead robust reliabl potenti inform methodolog pw monitor howev apart various multichannel multi site system propos literatur true multimod sensor pw record propos yet acquir pw signal simultan measur site studi true multimod pw sensor present obtain integr piezoelectr forcecardiographi fcg sensor ppg sensor thus enabl simultan mechan optic measur pws site bodi novel sensor perform assess measur finger pws five healthi subject rest preliminari result studi show first time delay exist pws record simultan ppg fcg sensor despit delay puls waveform acquir ppg fcg sensor along first second deriv high normal cross correl indic excess six well establish morpholog paramet pws compar via linear regress correl bland altman analys show paramet good agreement subject preliminari result proof concept studi must confirm much larger cohort subject investig also necessari shed light physic origin observ delay optic mechan pw signal research pave way develop true multimod wearabl integr sensor potenti sensor fusion approach improv perform pw monitor various bodi site | D010092;D017156 | T059;T060 |
36,236,659 | real time non contact millimet wave radar base vital sign detect paper extract life activ spectrum base millimet mm wave radar design realiz detect target object threshold trigger modul maximum likelihood estim method select complet design averag earli warn probabl trigger function threshold trigger modul design echo signal static object echo signal interfer extract doppler frequenc shift result move target detect method select filter design static clutter interfer filter without affect phase differ detect sequenc highlight target signal improv frequenc displac thorac movement use detect data fourier transform calcul sequenc spectrum valu extract within estim rang heartbeat respir spectrum heartbeat respir signal pick propos design use modelsim quartus co simul complet simul verif function extract number logic unit occupi comput resourc verifi algorithm vital sign experi heartbeat respir detect use sport bracelet relat error heartbeat detect respir detect relat error heartbeat detect overwhelm less | D011825;D012815 | T066;T073 |
36,236,658 | comparison machin learn algorithm featur set automat vocal emot recognit speech vocal emot recognit ver natur speech often refer speech emot recognit ser remain challeng human comput appli field includ clinic diagnosi intervent social interact research human comput interact hci increas benefit effici ver algorithm sever featur set use machin learn ml algorithm discret emot classif howev consensus low level descriptor classifi optim therefor aim compar perform machin learn algorithm sever differ featur set concret seven ml algorithm compar berlin databas emot speech multilay perceptron neural network mlp j decis tree dt support vector machin sequenti minim optim smo random forest rf k nearest neighbor knn simpl logist regress log multinomi logist regress mlr fold cross valid use four opensmil featur set ie emobas gemap egemap result indic smo mlp log show better perform reach accuraci respect compar rf dt mlr knn minimum accuraci respect overal emobas featur set perform best discuss implic find applic diagnosi intervent hci | D000069550;D013060 | T054;T066;T040 |
36,236,656 | ground reaction forc base gait analysi applic gait disord assess new indic quantifi walk behavior gait assess import tool determin whether person gait disord exist gait analysi studi high error rate due heel contact event base techniqu goal overcom shortcom exist gait analysi techniqu develop object indic assess gait disord paper propos method assess gait disord via observ chang center pressur cop medial later direct ie copx gait cycl data copx use design gait cycl estim method applic patient gait disord polar gaitogram drawn use gait cycl copx data differ area insid two close curv polar gaitogram area ratio index ari slope tangenti line common two close curv propos gait analysi indic experiment studi conduct verifi two indic use differenti stroke patient healthi adult find indic potenti use propos polar gaitogram indic develop appli wearabl devic assess gait disord | D000077107;D000076251 | T060;T073 |
36,236,654 | eeg valid photobiomodul treatment dementia case studi articl perform case studi impact photobiomodul pbm brain power spectrum connect elder person self administ gerocognit exam sage score indic probabl memori think disord first design realiz prototyp near infrar nir devic pbm analys alpha band power spectrum found posit long term effect nine sixteen electrod eye open condit oe twelv sixteen electrod eye close condit ce theta band posit long term effect found nine sixteen electrod oe seven sixteen electrod ce consid theta alpha ratio tar posit long term effect found thirteen sixteen electrod oe fourteen sixteen electrod ce connect analysi use imaginari compon complex pearson correl coeffici imcpcc also perform global effici measur base connect matric threshold calcul global effici calcul long term effect higher stimul factor oe condit factor ce condit case studi suggest pbm could posit effect improv desir brain activ measur improv power spectrum connect measur theta alpha band elder peopl memori think disord | D001921;D003704 | T023;T048 |
36,236,653 | optic fiber sensor sens network overview main principl applic optic fiber sensor present sever advantag relat type sensor advantag essenti relat optic fiber properti ie small lightweight resist high temperatur pressur electromagnet passiv among other sens achiev explor properti light obtain measur paramet temperatur strain angular veloc addit optic fiber sensor use form optic fiber sens network ofsn allow manufactur creat versatil monitor solut sever applic eg period monitor along extens distanc kilomet extrem hazard environ insid structur engin cloth health monitor assist literatur avail subject focus specif field optic sens applic detail principl oper paper present broad overview provid reader literatur review describ main principl optic sens highlight versatil advantag differ real world applic optic sens moreov includ overview discuss less common architectur optic sens wireless sensor network wsns integr har benefit world | D055100 | T073 |
36,236,652 | visual slam dynam environ base object detect optic flow dynam object remov dynam indoor environ visual simultan local map vslam system oper move object consid could affect system visual odomet stabil posit estim accuraci vslam use featur point sequenc imag sourc input perform local simultan creat map environ vslam system base orb slam yolor propos paper newli propos system combin object detect model yolox appli extract featur point capabl achiev better accuraci compar vps slam ds slam static featur point sign bench use calcul camera posit dynam move object elimin use track thread specif custom person dataset includ indoor outdoor rgb pictur train station includ dynam object high densiti peopl ground truth data sequenc data video record train station x z data use valid evalu propos method result show orb slam yolor object detect achiev accuraci dynam indoor environ compar previous system vps slam | D000465;D057849 | T170;T040 |
36,236,651 | coverag path plan point interest detect use autonom drone swarm unman aerial vehicl uav drone present enhanc miniatur sensor provid inform relat environ detect chang temperatur orient altitud geograph locat electromagnet fluctuat light condit combin inform proper help produc advanc environment awar thus drone navig environ autonom wireless communic also aid creation drone swarm combin proper algorithm coordin toward area coverag various mission search rescu coverag path plan cpp field studi drone independ swarm cover area interest effici current work cpp algorithm propos swarm drone detect point interest collect inform algorithm effect evalu simul result set characterist defin describ coverag radius drone speed swarm coverag path follow result show larger swarm size mission requir less time point interest detect within area two coverag path examin parallel line spiral coverag result depict parallel line coverag time effici sinc spiral increas requir time averag case number detect point interest | D000531;D000088722 | T073;T081 |
36,236,650 | high qualiti video watermark base deep neural network video hevc compress articl present method transpar watermark high capac watermark video hhevc high effici video code compress condit maintain high qualiti encod imag aim paper present method watermark embed use neural network condit subject video lossi compress hevc codec use yuvp color model chromin channel watermark paper present method train deep neural network emb watermark compress channel present discuss method character high accuraci video embed watermark compar origin psnr peak signal nois ratio valu obtain db watermark capac bit imag resolut method enabl complet recoveri watermark singl video frame compress hevc codec within rang compress valu defin crf constant rate factor | D044962;D007090 | T062;T066;T060 |
36,236,647 | anomal network traffic detect method base elev harri hawk optim method gate recurr unit classifi recent year network traffic contain lot featur inform mani redund featur comput cost algorithm great increas paper propos anomal network traffic detect method base elev harri hawk optim method easier identifi redund featur anomal network traffic reduc comput overhead improv perform anomal traffic detect method enhanc random jump distanc function escap energi function design uniqu fit function uniqu anomal traffic detect method built use algorithm neural network anomal traffic detect method test three public network traffic dataset name unsw nb nsl kdd cicid experiment result show propos method signific reduc number featur dataset comput overhead also give better indic everi test | D046093;D016571 | T012;T170 |
36,236,645 | energi save planner model via differenti evolutionari algorithm bionic pallet robot energi save pallet robot fundament problem field industri robot howev pallet robot often suffer problem high energi consumpt lack flexibl work introduc novel differenti evolut algorithm address advers effect caus instabl initi trajectori paramet reduc energi special simplifi analyt model pallet robot first develop simplifi analyt model differenti evolutionari algorithm combin form planner goal reduc energi consumpt energi save planner optim initi paramet trajectori collect bionic demonstr system turn enabl reduct oper power consumpt pallet robot major novelti articl use differenti evolutionari algorithm save energi consumpt well boost flexibl compar tradit algorithm propos method achiev state art perform simul actual experiment result illustr optim trajectori paramet effect reduc energi consumpt pallet robot | D012371 | T090 |
36,236,644 | path form healthcar practition indoor space use mobil crowdsens numer caus wast healthcar system wast associ ineffici among propos solut address ineffici clinic layout optim optim depend oper resourc instrument place clinic order access attain particular task mobil clinician differ clinic room accomplish differ clinic task tradit optim research involv manual monitor human proctor time consum erron unproduct subject mobil pattern indoor space determin automat real time layout oper relat optim decis base pattern implement accur continu time fashion paper explor applic domain precis local requir howev determin mobil essenti real time basi given research explor mobil devic built bluetooth receiv signal strength indic rssi use determin mobil collect stationari mobil devic comput network capabl lack energi requir mobil move mobil devic determin research methodolog involv develop two new algorithm use raw rssi data creat visual movement across differ oper unit identifi stationari node compar similar approach research showcas method present paper viabl produc mobil pattern indoor space util data analysi visual | D000465;D009068 | T170;T040 |
36,236,643 | e prevent advanc support system monitor relaps prevent patient psychot disord analyz long term multimod data wearabl video captur wearabl technolog digit phenotyp foster uniqu opportun design novel intellig electron servic address various well issu patient mental disord ie schizophrenia bipolar disord thus potenti revolution psychiatri clinic practic paper present e prevent innov integr system medic support facilit effect monitor relaps prevent patient mental disord technolog offer e prevent includ long term continu record biometr behavior indic smartwatch ii video record patient interview clinician use tablet iii automat systemat storag data dedic cloud server iv abil relaps detect predict paper focus descript e prevent system methodolog develop identif featur represent correl predict psychopatholog relaps patient mental disord specif tackl problem relaps detect predict use machin deep learn techniqu collect data result promis indic predict could made lead eventu predict psychopatholog prevent relaps | D011618;D012559;D000076251 | T073;T048 |
36,236,642 | assess walk stabil base whole bodi movement deriv depth sens camera stabil walk consid crucial aspect assess gait abil current studi aim assess walk stabil appli princip compon analysi pca decompos three dimension whole bodi kinemat data healthi young adult year femal deriv depth sens camera set movement componentssynergi call princip movement pms form togeth achiev task goal effect sex focus area test three pca base variabl comput pm relat explain varianc rvar measur composit movement structur largest lyapunov expon lye measur variabl number zero cross n measur tight neuromuscular control result show sex effect appear specif pms specif pm resembl swing phase movement femal greater lye p n p valu male moreov pm repres mid stanc phase movement femal smaller rvar p greater n p valu male empir find suggest inher sex differ walk stabil consid assess train locomot | D008191 | T131;T197 |
36,236,641 | high precis vital sign monitor method use fmcw millimet wave sensor method use millimet wave radar sensor detect human vital sign name respir heart rate receiv widespread attent non contact monitor sensor compact lightweight abl sens detect various scenario howev still face serious problem noisi interfer hardwar lead low signal nois ratio snr use frequenc modul continu wave fmcw radar sensor oper ghz offic environ extract respir heart rate person accustom sit chair inde propos signal process includ novel impuls denois oper spectral estim decis method uniqu term nois reduct accuraci improv addit propos method provid high qualiti repeat respir heart rate relat error averag compar refer valu measur reliabl smart bracelet | D011825;D055986 | T073;T201 |
36,236,640 | coulomet digit control galvanostat photometr endpoint detect work coulomet develop digit control galvanostat simpl colorimet base rgb led use light emitt coupl light detector light depend resist ldr photodiod develop endpoint detector hardwar softwar adapt origin galvanostat design regard hardwar new electr signal condition filter voltag divid includ optim work system softwar develop base open sourc arduino uno microcontrol differ variabl control titrat process manag add modul excel data acquisit softwar freeli avail studi possibl variabl influenc titrat process carri system test two classic coulometr titrat iodometri thiosulf ascorb acid acidbas potassium acid phthalat standard develop system versatil differ endpoint color indic employ starch phenolphthalein investig reaction differ experiment arrang studi natur electrod pt ag type cell two separ compart singl compart light detector ldr photodiod influenc sever experiment paramet electr light integr time studi chosen obtain best perform complet system reproduc result obtain control condit case acidbas titrat presenc atmospher carbon dioxid detect whose interfer main affect stir rate titrat time | D002245;D013885 | T123;T197 |
36,236,638 | advanc electrochem nano biosensor biomed environment applic current work futur perspect modern life qualiti strong support advanc made biosensor attribut crucial viabl contribut point care poc technolog develop poc devic exploit fast trace diseas progress rapid analysi water food qualiti assess blood glucos meter home pregnanc strip covid rapid test repres common exampl success biosensor biosensor provid great specif due incorpor select bio recognit element portabl signific reduc cost electrochem biosensor platform one advantag platform offer mani merit cheap select specif rapid portabl furthermor incorpor smartphon various analyt approach order increas sensit mani properti broad interdisciplinari area research develop biosensor includ disciplin background materi scienc chemistri physic medicin microbiologybiolog engin accord state art articl histor background alongsid long journey biosens construct develop start clark oxygen electrod reach high advanc wearabl stretchabl biosens devic discuss consequ select exampl among miscellan applic nanobiosensor microbi detect cancer diagnosi toxic analysi food qualiti control assur point care health prognosi describ eventu futur perspect intellig biosensor commerci exploit real life go support machin learn artifici intellig ai state | D015374;D000086382 | T059;T047;T075;T067 |
36,236,637 | trajectori optim term energi perform industri robot manufactur industri current high demand new product automot sector requir larg invest factori automot industri character high automat larg achiev manipul robot capabl multitask work present method optim trajectori robot six degre freedom spheric wrist optim trajectori base maxim manipul minim electr energi purpos necessari take account kinemat dynam manipul order integr algorithm calcul algorithm base kalman method algorithm implement simul trajectori serial industri robot robot sealer gun locat sixth axi qualiti sealer applic depend direct orient gun optim trajectori applic sealer must guarante method also appli three differ trajectori automot sector obtain result manipul electr energi consumpt prove effici algorithm therefor method search optim solut within limit manipul maintain orient final effector use known trajectori | D012371 | T090 |
36,236,636 | detect relat quantif neodymium sillai patti carbonatit use decis tree classif hyperspectr data recent increas import rare earth element ree proport advanc modern technolog green energi defens urg research look sophist effici explor method host rock carbonatit hyperspectr remot sens long recogn great potenti identifi ree base sharp distinct absorpt featur visibl near infrar vnir shortwav infrar swir electromagnet spectral profil instanc neodymium nd one abund light rare earth element lree among distinct absorpt featur ree vnir part electromagnet spectrum center nm vnir posit absorpt featur prove independ mineralog host nd featur observ sampl low nd ppm studi neodymium index ni propos base nm absorpt featur test hyperspectr imag sillai patai carbonatit sampl identifi nd pixel deciph relat concentr nd sampl base depth absorpt featur preliminari spectral studi carbonatit sampl carri use spectroradiomet determin presenc nd sampl two absorpt featur nd center nm promin nd rich sampl absorpt featur either weak suppress featureless spectra associ miner similar absorpt featur found vnir swir imag rock sampl captur laboratori base hyperspectr camera process minimum nois fraction mnf fast fourier transform fft filter signal nois reflect data rgb fals color composit continuum remov vnir reflect band cover wavelength nm effici display spatial distribut nd rich hotspot hyperspectr imag depth nm absorpt featur correspond concentr nd pixel compar greater zone quantifi use propos ni deeper absorpt featur higher ni quantifi nd rich pixel continuum remov vnir imag differ threshold valu ni introduc decis tree classifi generat number pixel class strength propos ni coupl decis tree classifi support accuraci assess classifi imag generat kappa coeffici compar result remot sens data obtain studi previous publish studi suggest sillai patti carbonatit rich nd associ ree part sampl high nd concentr ppm | D008903;D009354 | T196;T197 |
36,236,634 | blockchain base secur iot system use devic ident manag share data secur effici identifi issu iot base smart system smart citi smart agricultur smart health etc larg number iot devic use smart system produc larg amount data iot devic general limit storag process capabl configur secur techniqu devic challeng paper propos novel devic ident manag approach blockchain base iot system provid data secur two way first lightweight time base identif protocol use hub identif valid data second data storag augment effect blockchain applic provid easi access immut data share among multipl parti initi prototyp implement show ident manag approach implement larg scale set system effect implement blockchain platform perform evalu result show prototyp fulfil system requir adequ | D000081186 | T066 |
36,236,633 | novel imag base diagnosi method use improv dcgan rotat machineri rotat machineri play import role industri system fault machineri may damag system health novel imag base diagnosi method use improv deep convolut generat adversari network dcgan propos featur recognit fault classif rotat machineri first vibrat signal data rotat machineri transform time frequenc featur imag data continu wavelet transform use fault classif neural network method adapt deep convolut neural network adcnn combin generat adversari network gan improv perform featur self learn abil input data compar differ fault diagnosi method propos method better perform imag featur classif rotat machineri | D016571;D058067 | T170;T062 |
36,236,630 | perform assess guid wave base shm system appli stiffen composit structur assess abil structur health monitor shm system varieti prerequisit contribut factor taken account within public varieti analyz activ introduc guid wave base shm system system possibl analyz perform without take account structur appli system paramet therefor interdepend perform assess display shm pyramid base structur monitor requir factor influenc qualiti capabl reliabl monitor system given put relat state art perform analysi non destruct evalu aspect similar treat similar way other locat environment condit structur depend demand novel solut use open access data set open guid wave platform detail method descript analysi path base perform assess presentedth adopt approach clear beg question decis framework threshold affect reliabl system addit find show effect propag path accord damag posit inde distanc damag direct affect system perform otherwis propag direct alter potenti detect approach despit anisotropi composit nonetheless finit waveguid make necessari look whole path singular phenomena associ reflect may appear numer investig help clarifi central wave mechan necess take sensor posit account influenc factor start find achiev issu discuss potenti futur step outlin | D015203 | T062;T081;T080 |
36,236,629 | design fabric test fulli print pressur sensor use hybrid print approach pressur sens new concept appli use differ transduct mechan manufactur techniqu includ print electron approach howev limit effort taken realis pressur sensor fulli use addit manufactur techniqu especi personalis guid prosthet biomed applic work present novel fulli print piezoresist pressur sensor realis use aerosol jet print ajp screen print aj p specif chosen print silver interconnect select laser sinter sls polyamid board customis substrat piezoresist electrod manual screen print top interconnect sens layer sensor electromechan test respons regist upon applic given signal term sensit hysteresi reproduc time drift appli ramp pressur sensor show two differ sensit region high sensit region rang mpa averag sensit mpa low sensit zone within mpa averag sensit mpa indetermin overlap region hysteresi neglig electr resist deviat observ time drift experi perform satisfi demand applic biomed field smart prosthet guid | D009757;D012834 | T109;T196;T122 |
36,236,628 | skin lesion classif imbalanc data use deep learn soft attent today rapid develop industri zone lead increas incid skin diseas pollut air accord report american cancer societi estim peopl suffer skin cancer peopl surviv context doctor provinci hospit health facil overload doctor lower level lack experi tool support doctor process diagnos skin diseas quick accur essenti along strong develop artifici intellig technolog mani solut support diagnosi skin diseas research develop paper combin one deep learn model densenet inceptionnet resnet etc soft attent unsupervis extract heat map main skin lesion furthermor person inform includ age gender also use worth note new loss function take account data imbal also propos experiment result data set ham show use inceptionresnetv soft attent new loss function give percent accuraci mean precis f score recal auc respect besid use mobilenetvlarg combin soft attent new loss function even though number paramet time less number hidden layer time less achiev accuraci time faster diagnosi inceptionresnetv | D000077321;D012871;D012878 | T191;T047;T066 |
36,236,626 | age relat differ intermuscular coher emg emg ankl joint antagonist muscl activ maxim lean intermuscular synchron one fundament aspect maintain stabl postur great import age process studi aim assess muscl synchron postur stabil asymmetri quiet stand limit stabil use wavelet analysi intermuscular synchroni antagonist semg semg surfac electromyographi coher asymmetri evalu tibiali anterior soleus muscl | D000843;D004856 | T040;T030 |
36,236,625 | optim method wide beam sonar transmit beamform imag map sonar forward look sonar fls side scan sonar sss sensor frequent use onboard autonom underwat vehicl acquir inform around vehicl desir sonar system insonifi larg area thus sonar transmit beampattern wide field view work studi problem optim wide transmiss beampattern consid convent phase array beampattern design problem array element transmit ident waveform complex weight vector adjust creat desir beampattern shape experi consid wide transmiss beampattern uniform output power paper introduc new iter convex optim method narrowband linear phase array compar exist approach convex concav convex optim iter convex method phase weight paramet allow complex disciplin convex concav program dccp compar iter convex optim method dccp standard convex optim see former method archiv optim beampattern closer desir beampattern furthermor number iter propos iter convex method achiev optim beampattern closer desir beampattern beampattern achiev optim dccp | D013016 | T070 |
36,236,624 | instanc level contrast learn weak supervis object detect weak supervis object detect wsod receiv increas attent object detect field requir imag level annot indic presenc absenc target object great reduc label cost exist method usual focus current individu imag learn object instanc represent ignor instanc correl differ imag address problem propos instanc level contrast learn icl framework mine reliabl instanc represent learn imag use contrast loss guid instanc represent learn current imag due divers instanc differ appear size shape propos instanc divers memori updat imu algorithm mine differ instanc represent store memori bank multipl represent vector per class also consid background inform enhanc foreground represent help memori bank propos memori awar instanc mine mim algorithm combin propos confid instanc similar across imag mine reliabl object instanc addit also propos memori awar propos sampl mps algorithm sampl posit propos remov negat propos balanc learn posit negat sampl conduct extens experi pascal voc voc dataset wide use wsod demonstr effect method compar baselin method bring map corloc gain pascal voc dataset map corloc gain pascal voc dataset | D000465;D007858 | T061;T170;T062;T041 |
36,236,623 | verifi delay function blockchain relat applic survey concept verifi delay function receiv attent research sinc first propos applic verifi delay also widespread blockchain research comput timestamp public random beacon resourc effici blockchain proof data replic paper introduc concept verifi delay function systemat summar type verifi delay function first descript characterist verifi delay function given weak verifi delay function increment verifi delay function decod verifi delay function trapdoor verifi delay function introduc respect construct verifi delay function general reli two secur assumpt algebra assumpt structur assumpt secur assumpt two differ verifi delay function describ base cryptographi theori second post quantum verifi delay function base super singular isogeni introduc final paper summar blockchain relat applic verifi delay function | D000081186 | T066 |
36,236,622 | identif sens rfid one key technolog iot field rfid radio frequenc identif technolog appear near year ago deploy wide earli boom develop still acceler name indic origin function identif object anim peopl applic essenti aim traceabl access control logist type use still relev today new applic context effici rfid tag rfid also evolv integr new capabl new tag known augment tag includ inform captur function explos connect object emerg internet thing iot old technolog rfid still promis futur probabl present privat profession environ field logist industri agricultur build health even space | D057233 | T057;T073 |
36,236,621 | implement morpholog filter remov spike epilept brain signal improv identif rippl epilepsi common diseas affect least popul compris number million peopl mani patient suffer drug resist version number potenti treatment method small howev sinc treatment epilepsi also proper diagnosi observ brain signal record import research area paper address problem develop reliabl techniqu remov spike sharp transient baselin brain signal use morpholog filter allow much precis identif call epilept zone resect one method epilepsi treatment use eight patient khz data set depend upon staba algorithm refer detect rippl found averag sensit fals detect rate techniqu signific respect | D004569;D004827 | T047;T060;T033;T034 |
36,236,620 | singl seed identif three medicago speci via multispectr imag combin stack ensembl learn multispectr imag msi becom new fast non destruct detect method seed identif previous research usual focus singl model msi data analysi alway employ featur increas risk effici system cost studi develop stack ensembl learn sel model success identifi singl seed sickl alfalfa medicago falcata hybrid alfalfa varia alfalfa sativa sel adopt three layer structur ie level princip compon analysi pca linear discrimin analysi lda quadrat discrimin analysi qda model dimension reduct featur extract drfe level support vector machin svm multipl logist regress mlr general linear model elast net regular glmnet extrem gradient boost xgboost basic learner level xgboost meta learner confirm valu overal accuraci kappa precis sensit specif sensit sel model signific higher basic model alon base spectral featur combin morpholog spectral featur furthermor also develop featur filter process success select optim featur one correspond content chlorophyl anthocyanin fat moistur seed sel model msi data analysi provid new way seed identif featur filter process potenti could use wide develop low cost narrow channel sensor | D000872;D029902 | T121;T109;T002 |
36,236,619 | anti bloom clock time delay integr ccds paper present investig respons time delay integr tdi charg coupl devic employ anti bloom clock use vari number tdi stage influenc charg bloom caus unus tdi stage tdi deploy select scheme shown experiment anti bloom clock mechan analyz impact bloom sensor characterist respons convers gain signal nois ratio investig comparison measur without anti bloom clock mechan present discuss detail | D059629 | T081 |
36,236,618 | field biomechan assess high low dive competit year old goalkeep wearabl sensor princip compon analysi dive save main duti footbal goalkeep biomechan investig dive techniqu conduct none sport specif environ present studi investig characterist goalkeep dive prefer ps non prefer nps side innov wearabl plus princip compon analysi pca approach nineteen competit academi goalkeep year perform seri high low dive ps nps dive perform regular footbal goal pitch full bodi kinemat collect wearabl inerti sensor mtw awinda xsen pca conduct reduc data dimension input matrix datapoint pca score extract kinemat variabl compar ps nps explain variabl gt high dive particip exhibit greater hip intern rotat less trunk later tilt p lt es gt ps nps low dive player exhibit greater ipsilater hip abduct domin lower trunk rotat p lt es gt ps nps dive nps goalkeep adopt sub optim pattern less trunk coordin limit explos ecolog test wearabl pca might help coach inspect relev dive characterist improv train effect | D004242;D012916;D000076251 | T056;T073 |
36,236,617 | multi section traffic flow predict base mlr lstm neural network aggrav road congest lead frequent traffic crash necessari reliev traffic pressur traffic flow predict well traffic flow target road section predict also close relat adjac road section therefor paper predict method base combin multipl linear regress long short term memori mlr lstm propos use incomplet traffic flow data past period time target predict section continu complet traffic flow data past period time adjac section joint predict traffic flow chang target section short time accur predict futur traffic flow chang solv base model suppos traffic flow data target road section partial miss past period time accuraci predict result current mainstream predict result base continu non miss target link flow data meanwhil small scale improv data time interv short enough case frequent mainten camera actual traffic section propos predict method feasibl wide use | D000063;D016571 | T170;T067;T037 |
36,236,616 | analysi possibl use cenospher product cement mortar use elev temperatur environ topic research includ articl evalu influenc cenospher select paramet mortar cement sampl design cem r portland cement applic differ addit amount experiment work consist compress strength bend strength test day matur heat temperatur c compress strength test half beam mm use obtain result properti mortar compar research confirm possibl produc cenospher modifi cement mortar cenospher use prepar cement mortar negat affect bend compress strength increas temperatur c increas content addit | D003252 | T073 |
36,236,614 | care robot ethic sens system older adult home mani studi explor emot mental servic robot provid older adult offer daili convers news music health inform howev ethic issu rais use sensor frail older adult monitor daili movement medic intak instanc still discuss studi develop older adult guid caregiv monitor robot dori detect recogn movement sens human pose accord two factor human center artifici intellig hcai framework design care robot servic base sens movement daili activ conduct focus group interview two group caregiv medic staff topic care robot servic patient prefrail frail elder individu live home base respons deriv focal servic area cognit support emot support physic activ support medic manag caregiv manag also found two group differ ethic judgment area digniti autonomi control privaci servic util sens care robot therefor pose recognit technolog adopt present work use joint coordin inform extract camera imag thus advantag protect human digniti person inform | D012371;D012656 | T090;T074 |
36,236,613 | hybrid detect breast abnorm base contrast agent introduc proof concept physic perspect manuscript attempt present proof concept physic perspect hybrid detect system base util contrast agent purpos indic breast tissu abnorm present concept photon count modul detector set k characterist radiat emit contrast agent wo x ray spectra use kv wal mm kv wal mm addit filtrat mm gd iodin contrast agent studi fingerprint tissu abnorm indic comput mont carlo model base previous publish valid tabul data tissu experiment measur develop purpos show present concept practic potenti howev weak accompani experiment valid studi consid two type intern tissu layer fibroustumor thick valu mm within extern layer fat tissu cm quantit number encount k photon qualit tumor fibrous ratio advantag use x ray spectra higher tube voltag kv count photon found addit quantit qualit benefit correspond domin high mm low mm tissu thick valu conclus util suitabl contrast agent fingerprint tissu abnorm acquisit combin morpholog function imag featur count k x ray could enhanc breast imag present form lead advanc prognost capabl breast abnorm | D003287;D007455 | T121;T123;T130;T196 |
36,236,611 | time seri consolid robust method tailor gestur recognit individu user robust inter session model gestur still open learn challeng sleev equip capacit strap sensor use captur two gestur data set conveni sampl eight subject two pipelin explor filt novel two stage algorithm introduc use unsupervis learn algorithm find sampl repres gestur transit discard prior train valid convent model tsc confus matrix use automat consolid common confus class label result set gestur tailor individu subject abil inter session test accuraci use time seri consolid tsc method increas baselin inter session averag retain averag possibl gestur categori pipelin use classic machin learn algorithm requir relat small amount data comput power compar deep learn solut method may also offer flexibl interfac design user suffer handicap limit manual dexter abil reliabl make gestur possibl implement edg devic low comput power | D005868;D010363 | T054;T041 |
36,236,608 | design shape optim strain gaug load cell axial forc measur test bench load cell indispens compon mani engin machineri industri autom measur sens forc torqu paper describ design analysi strain gaug load cell concept design stage shape optim base finit element method fem techniqu calibr provid ampl load capac low cost materi aluminum high accur forc measur amplifi circuit half wheatston bridg configur two strain gaug implement experiment actual load cell prototyp calibr test conduct evalu load cell characterist deriv govern equat sens unknown load depend measur output voltag measur sensit load cell approxim mvn vv maximum appli load kg find support fem result experi accept percentag error reveal overal error worst situat therefor propos load cell meet design consider axial forc measur laboratori test bench light weight g maximum axial forc capac n good sensor characterist | D000535 | T121;T196 |
36,236,607 | perform evalu mptcp simultan use g g network g cellular network come promis provid high data rate low latenc becom critic advanc technolog mobil oper current deploy g cellular network worldwid howev limit coverag high suscept g network obstacl handoff g g vice versa frequent occur especi user equip ue move handoff often caus signific delay data transmiss due packet loss retransmiss promis solut use g g network simultan solv problem yield better throughput multipath transmiss control protocol tcp effect solut problem requir signific perform evalu practic deploy studi implement mptcp testb base ns dce enabl test perform mptcp schedul congest control algorithm ccas gpp non gpp network extens simul experi scenario ue simultan util g g network found block estim blest schedul implement balanc link adapt balia cca produc highest throughput lowest delay final show receiv signal interfer nois ratio sinr congest window throughput packet loss interconnect | D017395;D011110 | T045;T123;T116 |
36,236,606 | ginkgo leaf inspir fabric micronanostructur demonstr flexibl enzym free glucos sensor flexibl enzym free glucos sensor attract widespread attent due import potenti applic clinic diagnosi flexibl wearabl devic implant devic vivo present still major problem fabric flexibl enzym free glucos sensor low detect limit high stabil high sensit low cost hinder practic applic report facil strategi fabric flexibl non enzymat glucos sensor use ginkgo leaf templat nio film pedot pss composit film deposit surfac ginkgo leaf induc micro nano hierarch structur sensit layer conduct layer respect prepar flexibl enzym free glucos sensor exhibit excel electrochem perform toward glucos oxid sensit mamm cm oper voltag v detect limit good anti interfer due simpl fabric process perform reliabl novel flexibl enzym free glucos sensor attract candid next generat wearabl implant non enzymat glucos diagnost devic | D049329;D000076251 | T073 |
36,236,605 | proactiv protect smart power grid cyberattack servic data transfer protocol comput intellig method articl discuss approach construct oper proactiv system protect smart power grid cyberattack servic data transfer protocol base combin comput intellig method identifi anomali network traffic evalu self similar detect classifi cyberattack anomali take effect protect measur use long short term memori lstm gate recurr unit gru cell fractal analysi mathemat statist neural network long short term memori use tool develop protect system issu softwar implement propos system format data set contain network packet smart grid system consid experiment result obtain use generat data set demonstr confirm high effici propos proactiv smart grid protect system detect cyberattack real near real time well predict impact cyberattack develop effici measur counter | D016571;D012984 | T090;T170;T073 |
36,236,604 | multipl fingerprint local artifici neural network fingerprint local promis indoor posit method thank advantag use preinstal infrastructur exampl wifi signal strength measur pre exist wifi router offlin phase fingerprint local method first store posit rssi measur pair dataset second predict target locat compar store fingerprint databas current measur databas size normal huge data pattern complic thus artifici neural network use model relationship fingerprint locat exist fingerprint locat howev develop predict singl locat practic mani user may requir posit servic core algorithm capabl multipl local main contribut paper paper multipl fingerprint local develop base artifici neural network analysi number target estim without loss accuraci conduct experi | D000465;D016571 | T170 |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.