Adversarial Attack on Skeleton-Based Human Action Recognition
TL;DR: In this paper , the constrained iterative attack for skeleton actions (CIASA) is proposed to perturb joint locations in an action sequence such that the resulting adversarial sequence preserves the temporal coherence, spatial integrity, and the anthropomorphic plausibility of the skeletons.
read more
Abstract: Deep learning models achieve impressive performance for skeleton-based human action recognition. Graph convolutional networks (GCNs) are particularly suitable for this task due to the graph-structured nature of skeleton data. However, the robustness of these models to adversarial attacks remains largely unexplored due to their complex spatiotemporal nature that must represent sparse and discrete skeleton joints. This work presents the first adversarial attack on skeleton-based action recognition with GCNs. The proposed targeted attack, termed constrained iterative attack for skeleton actions (CIASA), perturbs joint locations in an action sequence such that the resulting adversarial sequence preserves the temporal coherence, spatial integrity, and the anthropomorphic plausibility of the skeletons. CIASA achieves this feat by satisfying multiple physical constraints and employing spatial skeleton realignments for the perturbed skeletons along with regularization of the adversarial skeletons with generative networks. We also explore the possibility of semantically imperceptible localized attacks with CIASA and succeed in fooling the state-of-the-art skeleton action recognition models with high confidence. CIASA perturbations show high transferability in black-box settings. We also show that the perturbed skeleton sequences are able to induce adversarial behavior in the RGB videos created with computer graphics. A comprehensive evaluation with NTU and Kinetics data sets ascertains the effectiveness of CIASA for graph-based skeleton action recognition and reveals the imminent threat to the spatiotemporal deep learning tasks in general.
read more
Chat with Paper
AI Agents for this Paper
Find similar papers on Google Scholar, PubMed and Arxiv
Write a critical review of this paper
Analyze citations of this paper to find unaddressed research gaps
Citations
SenseFi: A library and benchmark on deep-learning-empowered WiFi human sensing
TL;DR: The SenseFi benchmark as discussed by the authors is a comprehensive benchmark with an open-source library for deep learning in WiFi sensing research that offers researchers a convenient tool to validate learning-based WiFi-sensing methods on multiple datasets and platforms.
56
Adversarial Bone Length Attack on Action Recognition
TL;DR: Zhang et al. as mentioned in this paper restrict the perturbations to the lengths of the skeleton's bones, which allows an adversary to manipulate only approximately 30 effective dimensions, and demonstrate that the proposed attack successfully deceived models with sometimes greater than 90% success rate.
6
Uncovering Hidden Vulnerabilities in Convolutional Neural Networks through Graph-based Adversarial Robustness Evaluation
TL;DR: In this article , the authors proposed a new adversarial robustness evaluation framework based on Graph of Patterns (GoPs) models and graph distance algorithms, which provides a fine-grained analysis of AR from three perspectives, providing targeted insight into the vulnerability of CNNs.
3
•Posted Content
Adversarial Bone Length Attack on Action Recognition.
TL;DR: Zhang et al. as mentioned in this paper restrict the perturbations to the lengths of the skeleton's bones and demonstrate that the proposed attack successfully deceived models with sometimes greater than 90% success rate.
1
Motion Recognition System in VR Training System for Train Attendants
12 May 2023
TL;DR: In this paper , the authors used the MediaPipe module as the skeleton key point recognition framework on the basis of virtual reality technology, and designed a gesture recognition solution for train conductors by using skeleton information extraction technology and deep learning related methods.
References
Realtime Multi-person 2D Pose Estimation Using Part Affinity Fields
Zhe Cao,Tomas Simon,Shih-En Wei,Yaser Sheikh +3 more
- 21 Jul 2017
TL;DR: Part Affinity Fields (PAFs) as discussed by the authors uses a nonparametric representation to learn to associate body parts with individuals in the image and achieves state-of-the-art performance on the MPII Multi-Person benchmark.
6.2K
Least Squares Generative Adversarial Networks
Xudong Mao,Qing Li,Haoran Xie,Raymond Y. K. Lau,Zhen Wang,Stephen Paul Smolley +5 more
- 01 Oct 2017
TL;DR: The Least Squares Generative Adversarial Network (LSGAN) as discussed by the authors adopts the least square loss function for the discriminator to solve the vanishing gradient problem in GANs.
5.1K
The Limitations of Deep Learning in Adversarial Settings
Nicolas Papernot,Patrick McDaniel,Somesh Jha,Matt Fredrikson,Z. Berkay Celik,Ananthram Swami +5 more
- 21 Mar 2016
TL;DR: This work formalizes the space of adversaries against deep neural networks (DNNs) and introduces a novel class of algorithms to craft adversarial samples based on a precise understanding of the mapping between inputs and outputs of DNNs.
5K
Practical Black-Box Attacks against Machine Learning
Nicolas Papernot,Patrick McDaniel,Ian Goodfellow,Somesh Jha,Z. Berkay Celik,Ananthram Swami +5 more
- 02 Apr 2017
TL;DR: This work introduces the first practical demonstration of an attacker controlling a remotely hosted DNN with no such knowledge, and finds that this black-box attack strategy is capable of evading defense strategies previously found to make adversarial example crafting harder.
3.6K
One Pixel Attack for Fooling Deep Neural Networks
TL;DR: This paper proposes a novel method for generating one-pixel adversarial perturbations based on differential evolution (DE), which requires less adversarial information (a black-box attack) and can fool more types of networks due to the inherent features of DE.
2.6K