text
stringlengths
27
153
label
stringlengths
27
153
id
int64
0
40k
natural is of description the process of a language image. an generating Image captioning
Image captioning is the process of generating a natural language description of an image.
37,100
can recognition and load systems. as issue challenging be important for Traffic categorized such an
Traffic load recognition can be categorized as an important and challenging issue for such systems.
37,101
KITTI the depth predictions for deep Our state-of-the-art monocular excel on approaches benchmark.
Our deep predictions excel state-of-the-art approaches for monocular depth on the KITTI benchmark.
37,102
tracking inter-dependence minimum of and The Bayesian classification risk. is the utilized detection, then with
The inter-dependence of detection, tracking and classification is then utilized with the minimum Bayesian risk.
37,103
are simulations. numerical superiority using The verified effectiveness of the and proposed approach
The effectiveness and superiority of the proposed approach are verified using numerical simulations.
37,104
the diseases of In reparation. mechanisms impair endothelial addition,chronic
In addition,chronic diseases impair the mechanisms of endothelial reparation.
37,105
conclusion,sepsis,aging dysfunction. chronic features endothelial and induce of In similar diseases
In conclusion,sepsis,aging and chronic diseases induce similar features of endothelial dysfunction.
37,106
an to approach using convolutional We analysis scene networks. semantic deep present
We present an approach to semantic scene analysis using deep convolutional networks.
37,107
to method approaches, contrast surface our volumetric on In operates geometry. directly
In contrast to volumetric approaches, our method operates directly on surface geometry.
37,108
data. point Crucially, other to noisy is construction the real-world clouds and applicable unstructured
Crucially, the construction is applicable to unstructured point clouds and other noisy real-world data.
37,109
algorithm The of types alternates updates. between Metropolis-Hastings two
The Metropolis-Hastings algorithm alternates between two types of updates.
37,110
temporal variations these occur spatial gradually observe and dynamics. to we However, relative that
However, we observe that these variations occur gradually relative to spatial and temporal dynamics.
37,111
deformation the Combined method our with viewpoints. coefficient arbitrary and from texture regression, neural-network-based synthesizes
Combined with neural-network-based coefficient regression, our method synthesizes the texture and deformation from arbitrary viewpoints.
37,112
processes identified The for specific and are inference implementation. reasoning future
The specific inference and reasoning processes are identified for future implementation.
37,113
segmentation. deep algorithm for to semantic learning use a is One instance-aware possibility
One possibility is to use a deep learning algorithm for instance-aware semantic segmentation.
37,114
a data. Such high achieve large methods amount accuracies annotated of but training require
Such methods achieve high accuracies but require a large amount of annotated training data.
37,115
only the We for that each is training class assume image, known. object
We assume that for each training image, only the object class is known.
37,116
background. The system object the corresponding the automatically segments from
The system automatically segments the corresponding object from the background.
37,117
datasets annotated help features. discriminative visual These richly network learn the
These richly annotated datasets help the network learn discriminative visual features.
37,118
introduced to The ensures corresponding is utilization training strategy discriminator. of
The corresponding training strategy is introduced to ensures utilization of discriminator.
37,119
visualization We conduct to experiment fine-grained motivation. Dogs on dataset also Stanford verify our
We also conduct visualization experiment on fine-grained Stanford Dogs dataset to verify our motivation.
37,120
as as programming requires Today's in proficient individuals are software who languages possible. industry many
Today's software industry requires individuals who are proficient in as many programming languages as possible.
37,121
SQL process the However, be challenging. out turns to learning of
However, the process of learning SQL turns out to be challenging.
37,122
being a descriptor only transformed MCC into is a texture originally minutia descriptor.
MCC originally being a minutia only descriptor is transformed into a texture descriptor.
37,123
texture converted minutiae codes minutia codes (MTCC). are The to cylinder cylinder
The minutia cylinder codes are converted to minutiae texture cylinder codes (MTCC).
37,124
setting information. Specifically, facial conditional cycle-consistency the additional into adapt we by landmarks leveraging
Specifically, we adapt cycle-consistency into the conditional setting by leveraging additional facial landmarks information.
37,125
images annotations resolution pathological cancer of and lack of areas. of high (gigapixel)
high resolution (gigapixel) of pathological images and lack of annotations of cancer areas.
37,126
mislabel patches remove set. proposed can The the RAL in the training
The proposed RAL can remove the mislabel patches in the training set.
37,127
also the of classification DenseNet images. is proposed (ADN), for pathological atrous
atrous DenseNet (ADN), is also proposed for the classification of pathological images.
37,128
evaluated ADN proposed and RAL on two datasets, been i.e. The have pathological
The proposed RAL and ADN have been evaluated on two pathological datasets, i.e.
37,129
the performance of The proposed + experimental framework, i.e. ADN demonstrate results excellent the RAL
The experimental results demonstrate the excellent performance of the proposed ADN + RAL framework, i.e.
37,130
from a is reconstruction connectivity for detection Microscopy data. (EM) task Electron critical Synaptic neural
Synaptic connectivity detection is a critical task for neural reconstruction from Electron Microscopy (EM) data.
37,131
well-justified and clean, relies on is training The conceptually and prediction expressive resulting model procedures.
The resulting model is conceptually clean, expressive and relies on well-justified training and prediction procedures.
37,132
to The relationship be explained. a existence such of need remote
The existence of such a remote relationship need to be explained.
37,133
practical overlooked point, This which obvious by literature, implications. has has been the
This point, which has been overlooked by the literature, has obvious practical implications.
37,134
generate Video visual by frames dynamic future learning realistic patterns. prediction aims to
Video prediction aims to generate realistic future frames by learning dynamic visual patterns.
37,135
propose this an Conditional to address Appearance-Motion challenge. We GAN
We propose an Appearance-Motion Conditional GAN to address this challenge.
37,136
dynamic In a adds new categories to dimension inventory practice, with challenge. another this fine-grained
In practice, a dynamic inventory with new fine-grained categories adds another dimension to this challenge.
37,137
arbitrary suitable simulating boundaries. our This makes flows for effectively turbulent with geometrical method
This effectively makes our method suitable for simulating turbulent flows with arbitrary geometrical boundaries.
37,138
Learning in a complex science. number of applications large systems data features the of dynamics
Learning the dynamics of complex systems features a large number of applications in data science.
37,139
graphs with unveil provide patterns that user behavioral They also interpretable and changes. the
They also provide the user with interpretable graphs that unveil behavioral patterns and changes.
37,140
to The individual accommodate changes edges. is notion at local breakpoint of proposed
The notion of local breakpoint is proposed to accommodate changes at individual edges.
37,141
in tracking can time. scheme that real track novel rigid present We object objects a
We present a novel object tracking scheme that can track rigid objects in real time.
37,142
uses subpixel-precise objects image approach The track with to accuracy. edges high
The approach uses subpixel-precise image edges to track objects with high accuracy.
37,143
always parameter However, available. small not models may be
However, small parameter models may not be always available.
37,144
by recent satellite the captured been a moving In years, satellite platform. have videos
In recent years, the satellite videos have been captured by a moving satellite platform.
37,145
noise In the detecting on particular, modeling. we a novel algorithm local offer based
In particular, we offer a novel detecting algorithm based on the local noise modeling.
37,146
exponential noise an potential by distribution. We from the targets probability vehicle differentiate patterns
We differentiate the potential vehicle targets from noise patterns by an exponential probability distribution.
37,147
in proposed background detection Thus, foreground for been enormous decades. object methods recent have subtraction
Thus, enormous background subtraction methods for foreground object detection have been proposed in recent decades.
37,148
combination solution other we is strategies. show And more than that our efficient
And we show that our solution is more efficient than other combination strategies.
37,149
During this polymerases, are ribosomes, proteome all the RNA time, doubled. and
During this time, ribosomes, RNA polymerases, and the proteome are all doubled.
37,150
develops inferential factor This dimensions. theory an state-varying paper of for models large
This paper develops an inferential theory for state-varying factor models of large dimensions.
37,151
are process. recurrent models, functions some loadings Unlike state general factor constant of
Unlike constant factor models, loadings are general functions of some recurrent state process.
37,152
methods Our analysis. estimator principal component nonparametric combines with
Our estimator combines nonparametric methods with principal component analysis.
37,153
conditions. to system This significantly some changing overcome the suddenly helps
This significantly helps the system to overcome some suddenly changing conditions.
37,154
scene proposed. spotting for trainable model An text is end-to-end network neural
An end-to-end trainable neural network model for scene text spotting is proposed.
37,155
both drastically GPU for training and usage computation prediction. times a reduces of Optional the
Optional usage of a GPU drastically reduces the computation times for both training and prediction.
37,156
dial-a-ride problem. paper multi-atomic annealing a static to called This solve (MATA) novel proposes algorithm
This paper proposes a novel algorithm called multi-atomic annealing (MATA) to solve static dial-a-ride problem.
37,157
evolved in We theory. explore has the and involved are it how problems what
We explore how it has evolved and what problems are involved in the theory.
37,158
explained terms of consciousness information. integrated IIT in can is be hypothesis that a
IIT is a hypothesis that consciousness can be explained in terms of integrated information.
37,159
In this urgently a is needed. context, survey IIT for critical
In this context, a critical survey for IIT is urgently needed.
37,160
this we IIT of and concepts fundamental To first related issues. introduce end,
To this end, we first introduce fundamental concepts of IIT and related issues.
37,161
transitions Thereafter, IIT major point related we issues. has intra-model through and out discuss been
Thereafter, we discuss major transitions IIT has been through and point out related intra-model issues.
37,162
in Finally, some issues the extra-model are in last section, IIT's presented. involved theoretical, principles
Finally, in the last section, some theoretical, extra-model issues involved in IIT's principles are presented.
37,163
crisp in such thereby often are Typically, crisp semantics answers. the of resulting queries
Typically, the semantics of such queries are often crisp thereby resulting in crisp answers.
37,164
in address to have past strategies been Several proposed this.
Several strategies have been proposed in past to address this.
37,165
taxonomy Entities generalized. have don't they end a being up and
Entities don't have a taxonomy and they end up being generalized.
37,166
the numbers of estimating crowd Crowd is people task in images. counting
Crowd counting is the task of estimating people numbers in crowd images.
37,167
hard. Density the very in is regression small area person general on
Density regression on the small person area is in general very hard.
37,168
distinct a fusion streams deep TSLFN two network. The followed by includes
The TSLFN includes two distinct deep streams followed by a fusion network.
37,169
datasets. on comprehensive benchmark conduct We experiments four
We conduct comprehensive experiments on four benchmark datasets.
37,170
path feature is in moving important an design. architecture Human
Human moving path is an important feature in architecture design.
37,171
path, elements the architects studying (e.g. By to basic the know arrange where
By studying the path, architects know where to arrange the basic elements (e.g.
37,172
simulation. a system SimArch, path moving presents multi-agent for human This paper
This paper presents SimArch, a multi-agent system for human moving path simulation.
37,173
using behavior Process. model built Decision It by a Markov a involves
It involves a behavior model built by using a Markov Decision Process.
37,174
characteristics human It kinds different by transition also assigning different models of probabilities.
It also models different kinds of human characteristics by assigning different transition probabilities.
37,175
agents. quickly algorithm sub-optimal path of plans A modified weighted the the search A*
A modified weighted A* search algorithm quickly plans the sub-optimal path of the agents.
37,176
traditional rebinning to proposed method. is a compared method The
The proposed method is compared to a traditional rebinning method.
37,177
directly works on clouds. algorithm point Our
Our algorithm works directly on point clouds.
37,178
to regions input An cloud vector. encoded with point is feature missing a first
An input point cloud with missing regions is first encoded to a feature vector.
37,179
and obtained analysed. plots can be different recurrence From variables RQA
From recurrence plots different RQA variables can be obtained and analysed.
37,180
such BER Viterbi-decoder, techniques and inter-leaving. Time are Different used improvement as, Frequency
Different BER improvement techniques are used such as, Viterbi-decoder, Time and Frequency inter-leaving.
37,181
Effect also observed. modulation of different on BER techniques performance is
Effect of different modulation techniques on BER performance is also observed.
37,182
natural metric a image an is for of important images. Learning patches analyzing tool
Learning a metric of natural image patches is an important tool for analyzing images.
37,183
stage. of joint behavior present design the helps the study to predict in The configuration
The present study helps to predict the behavior of joint configuration in the design stage.
37,184
interest attracted powerful of because captioning image in mechanisms performance. have its Attention considerable
Attention mechanisms have attracted considerable interest in image captioning because of its powerful performance.
37,185
propose accessing We mathematical areas (LGAs). a model for among government chemotherapy local
We propose a mathematical model for accessing chemotherapy among local government areas (LGAs).
37,186
the features noise the main relevant Removing without image goal. the of affecting is
Removing the noise without affecting relevant features of the image is the main goal.
37,187
approaches become of solve this over to Sparse decomposition problem. dictionaries used most the redundant
Sparse decomposition over redundant dictionaries become of the most used approaches to solve this problem.
37,188
parts-based NMF naturally sparse, codes favor representations.
NMF codes naturally favor sparse, parts-based representations.
37,189
sparse redundant dictionary as representation, linear a combination represented a of atoms. In signals
In sparse representation, signals represented as a linear combination of a redundant dictionary atoms.
37,190
The one. gives better from denoised proposed an N-NMF image reconstruction of a
The proposed N-NMF gives a better reconstruction of an image from denoised one.
37,191
matrix non-negative dictionary sparse pursuit, denoising, representation, matching Keywords: learning, factorization. Image
Keywords: Image denoising, sparse representation, dictionary learning, matching pursuit, non-negative matrix factorization.
37,192
increasingly vision developed. networks convolutional robotic are systems neural Machine applications for being (CNNs) using
Machine vision systems using convolutional neural networks (CNNs) for robotic applications are increasingly being developed.
37,193
meaning not public affects economic just welfare, from drawn such The health. statements
The meaning drawn from such statements affects economic welfare, not just public health.
37,194
quantum-inspired the on the method shrinks proposed based function. complex threshold wavelet adaptive Then, coefficients
Then, the proposed method shrinks complex wavelet coefficients based on the quantum-inspired adaptive threshold function.
37,195
characterization semantics closure. of develop the bi-preference for We defining a skeptical a semantics
We develop a bi-preference semantics semantics for defining a characterization of the skeptical closure.
37,196
this adversarial propose based problem, adaptation unsupervised To on domain alleviate networks. we framework an
To alleviate this problem, we propose an unsupervised domain adaptation framework based on adversarial networks.
37,197
predictions introduce discriminator ground from truth that masks. we Hence, a distinguishes segmentation
Hence, we introduce a discriminator that distinguishes segmentation predictions from ground truth masks.
37,198
several a performing challenging performers activities. such with the this dataset In we introduce paper
In this paper we introduce such a challenging dataset with several performers performing the activities.
37,199