Abstract:This paper provides a comprehensive and exhaustive study of adversarial attacks on human pose estimation models and the evaluation of their robustness. Besides highlighting the important differences between well-studied classification and human pose-estimation systems w.r.t. adversarial attacks, we also provide deep insights into the design choices of pose-estimation systems to shape future work. We benchmark the robustness of several 2D single person pose-estimation architectures trained on multiple datasets, MPII and COCO. In doing so, we also explore the problem of attacking non-classification networks including regression based networks, which has been virtually unexplored in the past. \par We find that compared to classification and semantic segmentation, human pose estimation architectures are relatively robust to adversarial attacks with the single-step attacks being surprisingly ineffective. Our study shows that the heatmap-based pose-estimation models are notably robust than their direct regression-based systems and that the systems which explicitly model anthropomorphic semantics of human body fare better than their other counterparts. Besides, targeted attacks are more difficult to obtain than un-targeted ones and some body-joints are easier to fool than the others. We present visualizations of universal perturbations to facilitate unprecedented insights into their workings on pose-estimation. Additionally, we show them to generalize well across different networks. Finally we perform a user study about perceptibility of these examples.

Human Pose Estimaties from Corrupt Silhouettes

Human Pose Estimation from Polluted Silhouettes Using Sub-manifold Voting Strategy

Based on network probabilistic graph human pose segmentation algorithm

Human Pose Estimation in Static Images Based on Region Segmentation and Monte Carlo Sampling

Robust Head-Pose Estimation Based on Partially-Latent Mixture of Linear Regressions

Robust Estimation of 3D Human Poses from a Single Image

Utilizing Uncertainty in 2D Pose Detectors for Probabilistic 3D Human Mesh Recovery

Human pose regression by combining indirect part detection and contextual information

An Optimization Based Framework for Human Pose Estimation

Robust 3D Human Pose Estimation from Single Images or Video Sequences

Probablistic Restoration with Adaptive Noise Sampling for 3D Human Pose Estimation

Subject-Specific Human Modeling for Human Pose Estimation

ManiPose: Manifold-Constrained Multi-Hypothesis 3D Human Pose Estimation

On the Robustness of Human Pose Estimation

POISE: Pose Guided Human Silhouette Extraction under Occlusions

Poseur: Direct Human Pose Regression with Transformers.

Occluded Human Pose Estimation based on Limb Joint Augmentation

Implicit 3D Human Mesh Recovery using Consistency with Pose and Shape from Unseen-view

A Two-Stage Scheme for Multi-View Human Pose Estimation

Towards Robust RGB-D Human Mesh Recovery

Hybrid model for Single-Stage Multi-Person Pose Estimation