A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2020; you can also visit the original URL. The file type is application/pdf.
Filters
Shuffle and Learn: Unsupervised Learning using Temporal Order Verification [article]
2016 arXiv pre-print
Preserved Fulltext
Other Versions
When used as pre-training for action recognition, our method gives significant gains over learning without external data on benchmark datasets like UCF101 and HMDB51. ... We formulate our method as an unsupervised sequential verification task, i.e., we determine whether a sequence of frames from a video is in the correct temporal order. ... Acknowledgments: The authors thank Pushmeet Kohli, Ross Girshick, Abhinav Shrivastava and Saurabh Gupta for helpful discussions. Ed Walter for his timely help with the systems. ...
arXiv:1603.08561v2 fatcat:ydbpme3cdfhlvalrbrkzs4gjyu
SeCo: Exploring Sequence Supervision for Unsupervised Representation Learning [article]
2021 arXiv pre-print
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2020; you can also visit the original URL. The file type is application/pdf.
Note that this fulltext copy is not of the "primary" version of this work. The version it corresponds to is:
2020 pre-print arXiv:2008.00975v1 fatcat:pc3yzk4dandwhb67xof6kwtcxa
The inherent supervision existing in such sequential structure offers a fertile ground for building unsupervised learning models. ... More remarkably, SeCo demonstrates considerable improvements over recent unsupervised pre-training techniques, and leads the accuracy by 2.96 ImageNet pre-training in action recognition task on UCF101 ... Experiments We empirically verify the merit of SeCo for unsupervised representation learning in three downstream tasks: action recognition, untrimmed activity recognition and object tracking. ...
arXiv:2008.00975v2 fatcat:eb3hnhbuqfhjvhquax6oro4kei
SeCo: Exploring Sequence Supervision for Unsupervised Representation Learning
2021 PROCEEDINGS OF THE THIRTIETH AAAI CONFERENCE ON ARTIFICIAL INTELLIGENCE AND THE TWENTY-EIGHTH INNOVATIVE APPLICATIONS OF ARTIFICIAL INTELLIGENCE CONFERENCE
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2022; you can also visit the original URL. The file type is application/pdf.
The inherent supervision existing in such sequential structure offers a fertile ground for building unsupervised learning models. ... action recognition task on UCF101 and HMDB51, respectively. ... and HMDB51 for action recognition. ...
doi:10.1609/aaai.v35i12.17274 fatcat:fn4gwveoqfdf7hmxylfjla5qh4
Anomaly Detection Based on Deep Learning Using Video for Prevention of Industrial Accidents [article]
2020 arXiv pre-print
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2020; you can also visit the original URL. The file type is application/pdf.
This paper proposes an anomaly detection method for the prevention of industrial accidents using machine learning technology. ... In this paper, we first introduce a framework for unsupervised anomaly detection using deep learning. Next, we introduce verification experiments. ... We acquired 4,758 sequential skeleton maps in the same way as described above. B. Training We performed LSTM-VAE learning using 4,589 sequential skeleton maps including only normal actions. ...
arXiv:2005.13734v1 fatcat:ms3cmwlofbgn3fiwkzgqdm7diq
Multimodal Biometrics For Identity Documents And Smart Cards: European Challenge
2007 Zenodo
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2017; you can also visit the original URL. The file type is application/pdf.
EUROPEAN OBJECTIVES AND BENEFITS The main objective of the European COST 2101 Action is to investigate novel technologies for unsupervised multimodal biometric authentication systems using a new generation ... for an application, by combining biometric measurements obtained either in parallel or sequentially. ...
doi:10.5281/zenodo.40238 fatcat:fbx2gf47ozak5cqxoe3yr6wblu
Unsupervised Deep Learning of Mid-Level Video Representation for Action Recognition
2018 PROCEEDINGS OF THE THIRTIETH AAAI CONFERENCE ON ARTIFICIAL INTELLIGENCE AND THE TWENTY-EIGHTH INNOVATIVE APPLICATIONS OF ARTIFICIAL INTELLIGENCE CONFERENCE
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2022; you can also visit the original URL. The file type is application/pdf.
In this paper, we propose an unsupervised deep learning method which employs unlabeled local spatial-temporal volumes extracted from action videos to learn midlevel video representation for action recognition ... Current deep learning methods for action recognition rely heavily on large scale labeled video datasets. ... Zitnick, and Hebert (2016) proposed to use an unsupervised sequential verification method to train a deep neural network for action recognition. ...
doi:10.1609/aaai.v32i1.12300 fatcat:cxeqndlr7zd33m5xljvydy4yx4
Unsupervised Learning of Long-Term Motion Dynamics for Videos
2017 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR)
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2018; you can also visit the original URL. The file type is application/pdf.
We present an unsupervised representation learning approach that compactly encodes the motion dependencies in videos. ... We argue that in order for the decoder to reconstruct these sequences, the encoder must learn a robust video representation that captures long-term motion dependencies and spatial-temporal relations. ... Finally, we thank all members of the Stanford Vision Lab and Stanford Computational Vision and Geometry Lab for their useful comments and discussions. ...
doi:10.1109/cvpr.2017.751 dblp:conf/cvpr/LuoPHAF17 fatcat:mnk4ejnkcreg5edgogrmoq6nuy
Unsupervised Learning of Long-Term Motion Dynamics for Videos [article]
2017 arXiv pre-print
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2020; you can also visit the original URL. The file type is application/pdf.
Note that this fulltext copy is not of the "primary" version of this work. The version it corresponds to is:
2017 pre-print arXiv:1701.01821v2 fatcat:bnjudcwwbfch7buivvejediql4Other Versions
We present an unsupervised representation learning approach that compactly encodes the motion dependencies in videos. ... We argue that in order for the decoder to reconstruct these sequences, the encoder must learn a robust video representation that captures long-term motion dependencies and spatial-temporal relations. ... [15] presented an approach to learn visual representation with an unsupervised sequential verification task, and showed performance gain for supervised tasks like activity recognition and pose estimation ...
arXiv:1701.01821v3 fatcat:othxk4rzqfh4rpoxbq7ljkdfsu
Time-Contrastive Networks: Self-Supervised Learning from Multi-view Observation
2017 2017 IEEE Conference on Computer Vision and Pattern Recognition Workshops (CVPRW)
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2020; you can also visit the original URL. The file type is application/pdf.
Unsupervised Object Interactions We compare our multi-view TCN model against the Shuffle & Learn[1] approach, using the exact same architecture and only changing the loss and the last layer. ... We also show what is, to the best of our knowledge, the first self-supervised results for end-to-end imitation learning of human motions with a real robot (Table 2) . ... Unsupervised learning using sequential verification for action recognition. CoRR, abs/1603.08561, 2016. ...
doi:10.1109/cvprw.2017.69 dblp:conf/cvpr/SermanetLHL17 fatcat:zd7cicgfsjhzlig3rc2npeo4nq
Self-Learning with Stochastic Triplet Loss
2020 2020 International Joint Conference on Neural Networks (IJCNN)
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2022; you can also visit the original URL. The file type is application/pdf.
Deep learning has offered significant performance improvements on several pattern recognition problems. ... Hence, this paper proposes an adaptation of the triplet loss for self-learning with entirely unlabeled data, where there is uncertainty in the generated triplets. ... The authors wish to acknowledge the creators and administrators of the UofTDB database (University of Toronto, Canada) and the YouTube Faces database (Tel Aviv University, Israel), which were essential for ...
doi:10.1109/ijcnn48605.2020.9206799 dblp:conf/ijcnn/PintoC20 fatcat:mpl4uiqkqrdf5km4hi7kd5gzkm
A SURVEY ON DEEP LEARNING TECHNIQUES FOR BIG DATA IN BIOMETRICS
2018 International Journal of Advanced Research in Computer Science
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2018; you can also visit the original URL. The file type is application/pdf.
Hence deep learning algorithms which are specialized in analysing such large volumes of unsupervised data can be utilized. ... Deep learning is appropriate for exploiting large volumes of data and for analysing raw data from multiple sources and in different styles. ... An unsupervised algorithm may be used to simultaneously learn more than one properties and the results from unsupervised learning can be further used for supervised learning [19] . ...
doi:10.26483/ijarcs.v9i1.5136 fatcat:3oniwqwp4fbzzkicxozhn3ms5y
Learning Facial Representations from the Cycle-consistency of Face [article]
2021 arXiv pre-print
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2021; you can also visit the original URL. The file type is application/pdf.
At training time, our model learns to disentangle two distinct facial representations to be useful for performing cycle-consistent face reconstruction. ... At test time, we use the linear protocol scheme for evaluating facial representations on various tasks, including facial expression recognition and head pose regression. ... We are grateful to the National Center for High-performance Computing for providing computing services and facilities. ...
arXiv:2108.03427v1 fatcat:7eajqmwwofe7pb4spkuqv7yiu4
Discriminative Invariant Kernel Features: A Bells-and-Whistles-Free Approach to Unsupervised Face Recognition and Pose Estimation
2016 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR)
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2016; you can also visit the original URL. The file type is application/pdf.
As a special case, a single common framework can be used to generate subject-specific pose-invariant features for face recognition and vice-versa for pose estimation. ... We additionally benchmark on CMU MPIE and outperform previous work in almost all cases on off-angle face matching while we are on par with the previous state-of-the-art on the LFW unsupervised and image-restricted ... Algorithm 1 formally describes DIKF, where G in our case models pose-variation for face verification. Sequential invariance to multiple transforms. ...
doi:10.1109/cvpr.2016.603 dblp:conf/cvpr/PalJS16 fatcat:p7azcyu4ezbong6dxjj7yyljk4
Front Matter: Volume 12084
2022 Fourteenth International Conference on Machine Vision (ICMV 2021)
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2022; you can also visit the original URL. The file type is application/pdf.
The publisher is not responsible for the validity of the information or for any outcomes resulting from reliance thereon. ... and clip augmentation for spatio-temporal action detection SESSION 3 MACHINE LEARNING 0V Robust deep unsupervised learning framework to discover unseen plankton species [12084-8] 0W Exploring loss functions ... approach based on fast multi-scale hybrid wavelet network for supporting diagnosis of neuromuscular disorders [12084-26] 10 Two-stream deep representation for human action recognition [12084-29] 11 Method ...
doi:10.1117/12.2625908 fatcat:zrgauhqj7ng65flfcmbhqx2ony
3D Features for Human Action Recognition with Semi-supervised Learning
2019 IET Image Processing
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2022; you can also visit the original URL. The file type is application/pdf.
These local features are used to build the trees for the random forest technique. During tree building, a semi-supervised learning is proposed for better splitting of data points at each node. ... For recognition of an action, mutual information is estimated for all the extracted interest points to each of the trained class by passing them through the random forest. ... [21] have used spatial-optical data organisation along with sequential learning to recognise different actions. ...
doi:10.1049/iet-ipr.2018.6045 fatcat:vqpek22slvaulmm5yqcguyzlwq
« Previous Showing results 1 — 15 out of 2,919 results