SkeletonVis: Interactive Visualization for Understanding Adversarial Attacks on Human Action Recognition Models

Authors

  • Haekyu Park Georgia Institute of Technology
  • Zijie J. Wang Georgia Institute of Technology
  • Nilaksh Das Georgia Institute of Technology
  • Anindya S. Paul Intel
  • Pruthvi Perumalla Georgia Institute of Technology
  • Zhiyan Zhou Georgia Institute of Technology
  • Duen Horng Chau Georgia Institute of Technology

Keywords:

Human Action Recognition, Visual Data Analytics, Interpretability Of Deep Neural Networks, Adversarial Attack, Skeleton-based Action Recognition

Abstract

Skeleton-based human action recognition technologies are increasingly used in video-based applications, such as home robotics, healthcare on the aging population, and surveillance. However, such models are vulnerable to adversarial attacks, raising serious concerns for their use in safety-critical applications. To develop an effective defense against attacks, it is essential to understand how such attacks mislead the pose detection models into making incorrect predictions. We present SkeletonVis, the first interactive system that visualizes how the attacks work on the models to enhance human understanding of attacks.

Downloads

Published

2021-05-18

How to Cite

Park, H., Wang, Z. J., Das, N., Paul, A. S., Perumalla, P., Zhou, Z., & Chau, D. H. (2021). SkeletonVis: Interactive Visualization for Understanding Adversarial Attacks on Human Action Recognition Models. Proceedings of the AAAI Conference on Artificial Intelligence, 35(18), 16094-16096. Retrieved from https://ojs.aaai.org/index.php/AAAI/article/view/18022