This repository contains the implementation code for the paper titled Addressing Adversarial Vulnerabilities in 3D Action Recognition with Hybrid ST-GCNs, associated with the DOI: [https://doi.org/10.1109/IPRIA68579.2025.11263650].
For citation, please use the following reference: S. Shafaati and J. Mohammadzadeh, “Addressing Adversarial Vulnerabilities in 3D Action Recognition with Hybrid ST-GCNs,” 2025 7th International Conference on Pattern Recognition and Image Analysis (IPRIA), Lahijan, Iran, Islamic Republic of, 2025, pp. 1–5, doi: 10.1109/IPRIA68579.2025.11263650. keywords: Training; Image recognition; Three-dimensional displays; Accuracy; Surveillance; Robustness; Pattern recognition; Spatiotemporal phenomena; Noise measurement; Robots; adversarial robustness; 3D action recognition; spatiotemporal Graph Convolutional Network; human motion analysis; robust action recognition.
I hope you find it useful.😎