MaskCLR: Attention-Guided Contrastive Learning for Robust Action Representation Learning

A robust action representation learning method that uses attention-guided contrastive learning to handle noisy and incomplete skeleton inputs.

Overview

MaskCLR (Attention-Guided Contrastive Learning for Robust Action Representation Learning) is a research project focused on improving robustness in skeleton-based action recognition. The method uses attention-guided masking in a contrastive framework to learn stronger and more resilient representations.

Authors

• Mohamed Abdelfattah
• Mariam Hassan
• Alexandre Alahi

Venue

CVPR 2024

One-Sentence Summary

MaskCLR improves the robustness of transformer-based action recognition methods against noisy and incomplete skeletons.

Links

• Project Page
• Paper (CVPR 2024)

BibTeX

@inproceedings{abdelfattah2024maskclr,
  title={MaskCLR: Attention-Guided Contrastive Learning for Robust Action Representation Learning},
  author={Abdelfattah, Mohamed and Hassan, Mariam and Alahi, Alexandre},
  booktitle={Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition},
  pages={18678--18687},
  year={2024}
}

Project Status: ✅ Published at CVPR 2024
Project Page: maskclr.github.io
Paper: CVPR OpenAccess PDF