3D Human Action Representation Learning via Cross-View Consistency Pursuit [PDF]
Computer Vision and Pattern Recognition, 2021 In this work, we propose a Cross-view Contrastive Learning framework for unsupervised 3D skeleton-based action Representation (CrosSCLR), by leveraging multi-view complementary supervision signal.
Linguo Li +5 more
semanticscholar +1 more source
For SALE: State-Action Representation Learning for Deep Reinforcement Learning [PDF]
Neural Information Processing Systems, 2023 In the field of reinforcement learning (RL), representation learning is a proven tool for complex image-based tasks, but is often overlooked for environments with low-level states, such as physical control problems.
Scott Fujimoto +5 more
semanticscholar +1 more source
Masked Motion Predictors are Strong 3D Action Representation Learners [PDF]
IEEE International Conference on Computer Vision, 2023 In 3D human action recognition, limited supervised data makes it challenging to fully tap into the modeling potential of powerful networks such as transformers.
Yunyao Mao +5 more
semanticscholar +1 more source
Prompted Contrast with Masked Motion Modeling: Towards Versatile 3D Action Representation Learning [PDF]
ACM Multimedia, 2023 Self-supervised learning has proved effective for skeleton-based human action understanding, which is an important yet challenging topic. Previous works mainly rely on contrastive learning or masked motion modeling paradigm to model the skeleton ...
Jiahang Zhang, Lilang Lin, Jiaying Liu
semanticscholar +1 more source
Self-Supervised 3D Action Representation Learning With Skeleton Cloud Colorization [PDF]
IEEE Transactions on Pattern Analysis and Machine Intelligence, 2023 3D Skeleton-based human action recognition has attracted increasing attention in recent years. Most of the existing work focuses on supervised learning which requires a large number of labeled action sequences that are often expensive and time-consuming ...
Siyuan Yang +5 more
semanticscholar +1 more source
Skeleton-Contrastive 3D Action Representation Learning [PDF]
ACM Multimedia, 2021 This paper strives for self-supervised learning of a feature space suitable for skeleton-based action recognition. Our proposal is built upon learning invariances to input skeleton representations and various skeleton augmentations via a noise ...
Fida Mohammad Thoker +2 more
semanticscholar +1 more source
Self-supervised Action Representation Learning from Partial Spatio-Temporal Skeleton Sequences [PDF]
AAAI Conference on Artificial Intelligence, 2023 Self-supervised learning has demonstrated remarkable capability in representation learning for skeleton-based action recognition. Existing methods mainly focus on applying global data augmentation to generate different views of the skeleton sequence for ...
Yujie Zhou +4 more
semanticscholar +1 more source
CMD: Self-supervised 3D Action Representation Learning with Cross-modal Mutual Distillation [PDF]
European Conference on Computer Vision, 2022 In 3D action recognition, there exists rich complementary information between skeleton modalities. Nevertheless, how to model and utilize this information remains a challenging problem for self-supervised 3D action representation learning.
Yunyao Mao +4 more
semanticscholar +1 more source
Hierarchical Contrast for Unsupervised Skeleton-based Action Representation Learning [PDF]
AAAI Conference on Artificial Intelligence, 2022 This paper targets unsupervised skeleton-based action representation learning and proposes a new Hierarchical Contrast (HiCo) framework. Different from the existing contrastive-based solutions that typically represent an input skeleton sequence into ...
Jianfeng Dong +5 more
semanticscholar +1 more source
Contrastive Positive Mining for Unsupervised 3D Action Representation Learning [PDF]
European Conference on Computer Vision, 2022 . Recent contrastive based 3D action representation learning has made great progress. However, the strict positive/negative constraint is yet to be relaxed and the use of non-self positive is yet to be explored.
Haoyuan Zhang +3 more
semanticscholar +1 more source