text
stringlengths
27
153
label
stringlengths
27
153
id
int64
0
40k
is input challenging single-view essential from images problem. yet an Generating a multi-view
Generating multi-view images from a single-view input is an essential yet challenging problem.
36,900
and has graphics, in vision, It broad applications robotics.
It has broad applications in vision, graphics, and robotics.
36,901
to propose problem. this We CR-GAN address
We propose CR-GAN to address this problem.
36,902
aggregate features then these approach. We a through bag-of-word
We then aggregate these features through a bag-of-word approach.
36,903
then previewed. draped and parameters can edited recovered updated the further The be garment
The recovered parameters can then be further edited and the updated draped garment previewed.
36,904
further of quantitative and proposed effectiveness evaluations and the robustness the qualitative work. Both demonstrate
Both quantitative and qualitative evaluations further demonstrate the effectiveness and robustness of the proposed work.
36,905
merely of not are the similar. shapes slowest the nearly that identical, eigenvectors find We
We find that the shapes of the slowest eigenvectors are nearly identical, not merely similar.
36,906
provided. updated review literature also is brief A
A brief updated literature review is also provided.
36,907
show high AIT for The SNR. outperforms WM that results
The results show that WM outperforms AIT for high SNR.
36,908
a estimation MMA accurate. feasible DoA In and single with is conclusion,
In conclusion, DoA estimation with a single MMA is feasible and accurate.
36,909
results modular to some Our on approaches lights shed knowledge compilation.
Our results shed some lights on modular approaches to knowledge compilation.
36,910
results apply epistemic we Furthermore, planning. the our in multi-agent
Furthermore, we apply our results in the multi-agent epistemic planning.
36,911
In samples annotated via addition, a selected and low-confidence few AL. with predictions are
In addition, a few samples with low-confidence predictions are selected and annotated via AL.
36,912
We to coefficients. rise that gives a assume common distribution regression the regime-specific
We assume that a common distribution gives rise to the regime-specific regression coefficients.
36,913
These flexible across coefficients a shrinkage which manner. in to enable differ regimes priors assess
These shrinkage priors enable to assess which coefficients differ across regimes in a flexible manner.
36,914
may thousands false cases. such in alarms of Methods yield
Methods may yield thousands of false alarms in such cases.
36,915
We CT pulmonary algorithm of scans. the a present novel for registration
We present a novel algorithm for the registration of pulmonary CT scans.
36,916
the Both fields driven data gradient term. are normalized by steps same
Both steps are driven by the same normalized gradient fields data term.
36,917
an paper proposes for This efficient digit Devanagari recognition handwritten script. system
This paper proposes an efficient digit recognition system for handwritten Devanagari script.
36,918
provide Observation multitude sensing heterogeneous Earth of Nowadays, a remote data. systems
Nowadays, Earth Observation systems provide a multitude of heterogeneous remote sensing data.
36,919
labor-intensive phenotypic information is collection Traditional and time-consuming. of
Traditional collection of phenotypic information is labor-intensive and time-consuming.
36,920
is of imagery phenotyping. popular for becoming Use
Use of imagery is becoming popular for phenotyping.
36,921
then from rectified the mosaic A generated is imagery. orthophoto
A rectified orthophoto mosaic is then generated from the imagery.
36,922
the plant. each statistical individual to find sorghum model present location We a of
We present a statistical model to find the location of each individual sorghum plant.
36,923
we recognition devices. the with way interact natural Gesture currently enables extension a of
Gesture recognition enables a natural extension of the way we currently interact with devices.
36,924
end, propose classifier a gesture a modeling probabilistic approach. hierarchical based we To this on
To this end, we propose a gesture classifier based on a hierarchical probabilistic modeling approach.
36,925
strictly increasing symmetries, symmetries sizes. compute generalizes many more for BV can block VV and
BV strictly generalizes VV symmetries, and can compute many more symmetries for increasing block sizes.
36,926
in regime driven encoding. population fluctuation the ultrafast realize can Cortical neurons
Cortical neurons in the fluctuation driven regime can realize ultrafast population encoding.
36,927
well mechanisms, The underlying however, biophysical not understood. are
The underlying biophysical mechanisms, however, are not well understood.
36,928
dissociation explanation provide for ultrafast sharp and of action response. the an potential We onset
We provide an explanation for the dissociation of sharp action potential onset and ultrafast response.
36,929
to datasets. up work noisily-annotated learning recognition Much visual recent to aims scale on massive,
Much recent work on visual recognition aims to scale up learning to massive, noisily-annotated datasets.
36,930
include representations. the domain-specific fashion in advances development numerical of Recent computational
Recent advances in computational fashion include the development of domain-specific numerical representations.
36,931
that from purpose, firstly extracted grid are the optical-flow, cuboid virtual-inertia video. log-covariance, For feature,
For that purpose, firstly grid optical-flow, virtual-inertia feature, log-covariance, cuboid are extracted from the video.
36,932
The datasets. was proposed of evaluated egocentric a on framework number
The proposed framework was evaluated on a number of egocentric datasets.
36,933
that outperforms the multi-modal The showed existing features using MKL results with methods.
The results showed that using multi-modal features with MKL outperforms the existing methods.
36,934
on benchmarks Extensive show of our approach. four the effectiveness challenging experiments
Extensive experiments on four challenging benchmarks show the effectiveness of our approach.
36,935
limit severely variations appearance-based the performance tracking. Personal of gaze
Personal variations severely limit the performance of appearance-based gaze tracking.
36,936
standard model network is difficult. neural using Adapting methods variations these to adaptation
Adapting to these variations using standard neural network model adaptation methods is difficult.
36,937
these by tackle Adaptive the GaZe (SPAZE). introducing Estimator We problems SPatial
We tackle these problems by introducing the SPatial Adaptive GaZe Estimator (SPAZE).
36,938
Calibrating reduces problem. SPAZE new for optimization a person a solving small to
Calibrating SPAZE for a new person reduces to solving a small optimization problem.
36,939
examining gaze model-based expected We to low-dimensionality that tracking. approaches by show is this
We show that this low-dimensionality is expected by examining model-based approaches to gaze tracking.
36,940
show possible. accurate that We head pose-free tracking gaze is also
We also show that accurate head pose-free gaze tracking is possible.
36,941
applications. instrumental variety becoming Intelligence of is a Artificial in
Artificial Intelligence is becoming instrumental in a variety of applications.
36,942
Context recognition. accurate for visual important is
Context is important for accurate visual recognition.
36,943
state. graphical Faster object infer to which model a aims R-CNN) with
Faster R-CNN) with a graphical model which aims to infer object state.
36,944
connectivity a A network scale-free is density if is proportional power-law function to distribution. its
A network is scale-free if its connectivity density function is proportional to a power-law distribution.
36,945
challenge. tracing path scattering using participating highly media brute a is force Rendering
Rendering highly scattering participating media using brute force path tracing is a challenge.
36,946
differential The solving simple to linear reduces the diffusion problem partial a equation. approximation
The diffusion approximation reduces the problem to solving a simple linear partial differential equation.
36,947
method problems. discuss prior to our Finally, how compares we standard work for various
Finally, we discuss how our method compares to prior work for various standard problems.
36,948
each two screening projections of breast. for Standard mammography breast X-ray involves the cancer acquisition
Standard breast cancer screening involves the acquisition of two mammography X-ray projections for each breast.
36,949
the views tumor supports a and task Typically, of detection localization. challenging comparison of both
Typically, a comparison of both views supports the challenging task of tumor detection and localization.
36,950
Siamese for network mammography. in dual-view introduce lesion learning, We matching patch-based a deep
We introduce a deep learning, patch-based Siamese network for lesion matching in dual-view mammography.
36,951
the matching evaluated of contribution the a pipeline. detection in we Finally, full network
Finally, we evaluated the contribution of the matching network in a full detection pipeline.
36,952
our approach. improved of using promise results Experimental the accuracy detection demonstrate
Experimental results demonstrate the promise of improved detection accuracy using our approach.
36,953
social form daily of the our in the life. basis network
form the basis of the social network in our daily life.
36,954
these cues studies. largely additional by are the previous However, overlooked
However, these additional cues are largely overlooked by the previous studies.
36,955
We neural networks. classification deep image propose new a with framework for
We propose a new framework for image classification with deep neural networks.
36,956
The to intermediate the framework outputs computational network. introduces graph a of
The framework introduces intermediate outputs to the computational graph of a network.
36,957
publicly benchmarks. been have up-to-date several available Experiments on conducted
Experiments have been conducted on several publicly available up-to-date benchmarks.
36,958
module PointSIFT integrated into improve architecture various the can to PointNet-based be ability. representation
PointSIFT module can be integrated into various PointNet-based architecture to improve the representation ability.
36,959
datasets. our Extensive outperforms framework show experiments PointSIFT-based method benchmark on standard state-of-the-art
Extensive experiments show our PointSIFT-based framework outperforms state-of-the-art method on standard benchmark datasets.
36,960
code accompanied and trained model by The be this paper. will published
The code and trained model will be published accompanied by this paper.
36,961
and training of biases methods exploit known Most the capture learning machine are data. to
Most machine learning methods are known to capture and exploit biases of the training data.
36,962
for harmful. While others beneficial learning, are some biases are
While some biases are beneficial for learning, others are harmful.
36,963
training present to Specifically, in captioning models tend data. image exaggerate biases
Specifically, image captioning models tend to exaggerate biases present in training data.
36,964
caption of specific gender words (e.g. investigate generation We
We investigate generation of gender specific caption words (e.g.
36,965
on the context. or appearance the man, person's image woman) based
man, woman) based on the person's appearance or the image context.
36,966
have Species high also abundant. which more are influence a selection on generally
Species which have a high influence on selection are also generally more abundant.
36,967
behavior from one How friendly can adversarial detect and raw data?
How can one detect friendly and adversarial behavior from raw data?
36,968
active for century study field of Neuroethology been an has more now. than a
Neuroethology has been an active field of study for more than a century now.
36,969
is embedding ignored. underlying space the However, feature
However, the underlying feature embedding space is ignored.
36,970
A also sampling balanced introduced. is strategy
A balanced sampling strategy is also introduced.
36,971
will be code of released. The trained model our and weights
The code and trained weights of our model will be released.
36,972
our understand behind lagged ability to has data always ability it. Our to collect
Our ability to understand data has always lagged behind our ability to collect it.
36,973
of the areas. over system of variant is new urban variety The evaluated a
The new variant of the system is evaluated over a variety of urban areas.
36,974
Computer and aided this make can processing image faster accurate. more
Computer aided image processing can make this faster and more accurate.
36,975
under by I begin key ideas discussion. the paper summarizing of
I begin by summarizing key ideas of the paper under discussion.
36,976
has for compression many been investigated decades. Image
Image compression has been investigated for many decades.
36,977
by CAEs compact than efficiency results, features. achieve coding experimental extracting to According better JPEG
According to experimental results, CAEs achieve better coding efficiency than JPEG by extracting compact features.
36,978
and ratio quality subjective high compression reconstruction. on GANs advantages large potential show
GANs show potential advantages on large compression ratio and high subjective quality reconstruction.
36,979
to among performance (RD) the best them, comparable achieves BPG. is rate-distortion Super-resolution which
Super-resolution achieves the best rate-distortion (RD) performance among them, which is comparable to BPG.
36,980
the between and elements visual is of a video. a natural There correlation auditive
There is a natural correlation between the visual and auditive elements of a video.
36,981
transform constructed on the fractional In the (FRWT). FrScatNet is approach, our based Methods: wavelet
Methods: In our approach, the FrScatNet is constructed based on the fractional wavelet transform (FRWT).
36,982
and scattering coefficients modulus iteratively operators. fractional FRWTs using The computed are
The fractional scattering coefficients are iteratively computed using FRWTs and modulus operators.
36,983
used signal fractional constructed feature The scattering usually for vectors are classification. coefficients by
The feature vectors constructed by fractional scattering coefficients are usually used for signal classification.
36,984
and we classify background into Then those target or patches malignant) groups. (benign
Then we classify those patches into target (benign or malignant) and background groups.
36,985
to Conclusion: and FrScatNet accurate The is robust achieve shown results.
Conclusion: The FrScatNet is shown to achieve accurate and robust results.
36,986
work. in discriminative by this obtained the and are scattering stable FrScatNet More coefficients fractional
More stable and discriminative fractional scattering coefficients are obtained by the FrScatNet in this work.
36,987
are observed signals by The corrupted noise. usually
The observed signals are usually corrupted by noise.
36,988
that heavy processing different the observed signal In be may tailed. applications errors it is
In different signal processing applications it is observed that the errors may be heavy tailed.
36,989
It both are has estimators strongly been consistent. that the observed
It has been observed that both the estimators are strongly consistent.
36,990
the estimators. asymptotic as have squares least distribution least The the squares same estimators approximate
The approximate least squares estimators have the same asymptotic distribution as the least squares estimators.
36,991
proposed simulations some see how numerical work. the to perform estimators We
We perform some numerical simulations to see how the proposed estimators work.
36,992
mesh. a such volumetric requires discretization, shapes, conforming this as a For tetrahedral boundary solid
For solid shapes, this requires a volumetric discretization, such as a boundary conforming tetrahedral mesh.
36,993
Experiments approach of benchmarks the on show our methods. advantages over several state-of-art
Experiments on several benchmarks show the advantages of our approach over state-of-art methods.
36,994
our results proposed baselines. benchmark method datasets outperforms state-of-the-art four on demonstrate that Experimental the
Experimental results on four benchmark datasets demonstrate that our proposed method outperforms the state-of-the-art baselines.
36,995
expensive. time-consuming both and Manual is annotation
Manual annotation is both time-consuming and expensive.
36,996
There three the in dataset. challenges are
There are three challenges in the dataset.
36,997
sparse naturalistic a behaviors First, setting. are driving driving in predefined
First, predefined driving behaviors are sparse in a naturalistic driving setting.
36,998
of distribution is Second, behaviors driving long-tail. the
Second, the distribution of driving behaviors is long-tail.
36,999