Self-supervised learning in CV在计算机视觉(CV)领域,目前的方法主要依赖大量的标注样本来学习丰富的视觉表征,从而在各项CV任务中取得较好的表现。然而在许多情况下,大规模的人工标注并不容易获得。因此,我们希望可以利用无监督方法去学习那些不带标注的样本。自监督学习,是无监督学习的一种,即无需额外的人工标签,仅利用数据自身的信息作为监督(自己监督自己)。利用来自数据自身的监督信息,设计一个pretext task,训练网络去完成该pretext task,从而促使网络学习到数据特征。
international conference on robotics and automation, no. 2 (2021): 1312-1319
We presented BADGR, an end-to-end learning-based mobile robot navigation system that can be trained entirely with selfsupervised, off-policy data gathered in real-world environments, without any simulation or human supervision, and can improve as it gathers more data
Cited by18BibtexViews405DOI
0
0
NIPS 2020, (2020)
Similar approaches are common in NLP, we demonstrate that this approach can be a surprisingly strong baseline for semi-supervised learning in computer vision, outperforming the state-of-the-art by a large margin
Cited by130BibtexViews1721
0
0
There exist several comprehensive reviews related to Pre-trained Language Models, Generative Adversarial Networks, Autoencoder and contrastive learning for visual representation
Cited by32BibtexViews5254
0
0
NIPS 2020, (2020)
We demonstrate that these self-supervised representations learn occlusion invariance by employing an aggressive cropping strategy which heavily relies on an object-centric dataset bias
Cited by20BibtexViews722
0
0
Triantafyllos Afouras, Andrew Owens,Joon Son Chung,Andrew Zisserman
european conference on computer vision, pp.208-224, (2020)
We have proposed a unified model that learns from raw video to detect and track speakers
Cited by12BibtexViews42
0
0
We studied self-supervised learning via an information-theoretical perspective
Cited by7BibtexViews752
0
0
Wallace Bram,Hariharan Bharath
european conference on computer vision, pp.717-734, (2020)
We found that even with traditional early stopping, validation accuracy could oscillate as the pretext overfit to the training data, potentially resulting in a poor model as the final result
Cited by6BibtexViews6
0
0
Ashish Jaiswal, Ashwin Ramesh Babu, Mohammad Zaki Zadeh, Debapriya Banerjee,Fillia Makedon
Technologies, no. 1 (2020): 2
The works based on contrastive learning have shown promising results on several downstream tasks such as image/video classification, object detection, and other natural language processing tasks
Cited by0BibtexViews20DOI
0
0
Alliegro Antonio, Boscaini Davide,Tommasi Tatiana
In this work we investigated how to deal with 3D labeled and unlabeled data possibly coming from different domains and with data annotation scarcity
Cited by0BibtexViews11
0
0
CVPR, (2019)
As part of our study, we drastically boost the performance of previously proposed techniques and outperform previously published state-of-the-art results by a large margin
Cited by188BibtexViews305
0
0
arXiv: Computer Vision and Pattern Recognition, (2019): 1-1
This paper provides an extensive review of deep learning-based self-supervised general visual feature learning methods from images or videos
Cited by148BibtexViews369DOI
0
0
Dan Hendrycks, Mantas Mazeika, Saurav Kadavath,Dawn Song
ADVANCES IN NEURAL INFORMATION PROCESSING SYSTEMS 32 (NIPS 2019), (2019): 15637-15648
We found large improvements in robustness to adversarial examples, label corruption, and common input corruptions
Cited by141BibtexViews234
0
0
Michelle A. Lee,Yuke Zhu, Krishnan Srinivasan, Parth Shah,Silvio Savarese,Li Fei-Fei,Animesh Garg,Jeannette Bohg
international conference on robotics and automation, (2019)
The primary goal of our experiments is to examine the effectiveness of the multimodal representations in contactrich manipulation tasks
Cited by122BibtexViews447DOI
0
0
Xiaohua Zhai, Avital Oliver, Alexander Kolesnikov,Lucas Beyer
International Conference on Computer Vision, (2019): 1476-1485
We further showed that S4L methods are complementary to existing semisupervision techniques, and Mix Of All Models, our proposed combination of those, leads to state-of-the-art performance
Cited by114BibtexViews283
0
0
CVPR, (2019): 4571-4580
We have presented a self-supervised approach to learning accurate optical flow estimation
Cited by98BibtexViews124
0
0
CVPR, pp.6706-6716, (2019)
We studied Pretext-Invariant Representation Learning for learning representations that are invariant to image transformations applied in self-supervised pretext tasks
Cited by87BibtexViews130DOI
0
0
computer vision and pattern recognition, (2019)
Data in H36M was captured with markers, and have high accuracy and consistency in 20] for two-dimensional annotations across subject and scenes; on the other hand, the annotations in MPII were done by humans and some of the keypoints are localized differently
Cited by66BibtexViews20
0
0
ADVANCES IN NEURAL INFORMATION PROCESSING SYSTEMS 32 (NIPS 2019), (2019): 8812-8824
As described in §4.2, Partial Registration Network currently requires inference-time fine-tuning on real scans to learn useful data-dependent representations; this makes PRNet slow during inference
Cited by55BibtexViews156
0
0
Medical Image Analysis, (2019): 101539
We proposed a novel self-supervised learning strategy based on context restoration
Cited by30BibtexViews22DOI
0
0