Abstract:Multi-person pose estimation generally follows top-down and bottom-up paradigms. The top-down paradigm detects all human boxes and then performs single-person pose estimation on each ROI. The bottom-up paradigm locates identity-free keypoints and then groups them into individuals. Both of them use an extra stage to build the relationship between human instance and corresponding keypoints (e.g., human detection in a top-down manner or a grouping process in a bottom-up manner). The extra stage leads to a high computation cost and a redundant two-stage pipeline. To address the above issue, we introduce a fine-grained body representation method. Concretely, the human body is divided into several local parts and each part is represented by an adaptive point. The novel body representation is able to sufficiently encode the diverse pose information and effectively model the relationship between human instance and corresponding keypoints in a single-forward pass. With the proposed body representation, we further introduce a compact single-stage multi-person pose regression network, called AdaptivePose++, which is the extended version of AAAI-22 paper AdaptivePose. During inference, our proposed network only needs a single-step decode operation to estimate the multi-person pose without complex post-processes and refinements. Without any bells and whistles, we achieve the most competitive performance on representative 2D pose estimation benchmarks MS COCO and CrowdPose in terms of accuracy and speed. In particular, AdaptivePose++ outperforms the state-of-the-art SWAHR-W48 and CenterGroup-W48 by 3.2 AP and 1.4 AP on COCO mini-val with faster inference speed. Furthermore, the outstanding performance on 3D pose estimation datasets MuCo-3DHP and MuPoTS-3D further demonstrates its effectiveness and generalizability on 3D scenes.

HPRNet: Hierarchical Point Regression for Whole-Body Human Pose Estimation

X-HRNet: Towards Lightweight Human Pose Estimation with Spatially Unidimensional Self-Attention

Context-Guided Adaptive Network for Efficient Human Pose Estimation.

Single-Network Whole-Body Pose Estimation

Multi-Stage HRNet: Multiple Stage High-Resolution Network for Human Pose Estimation

Bottom-Up Human Pose Estimation Via Disentangled Keypoint Regression

Bottom-Up Human Pose Estimation Based on Multiple Anchor Points Regression

Bottom-Up Human Pose Estimation by Ranking Heatmap-Guided Adaptive Keypoint Estimates

Hierarchical Pose Net: Spatial Hierarchical Body Tree Driven Multi-Person Pose Estimation

DHRNet: A Dual-Path Hierarchical Relation Network for Multi-Person Pose Estimation

AdaptivePose++: A Powerful Single-Stage Network for Multi-Person Pose Regression

OpenPose : Whole-Body Pose Estimation

AdaptivePose: Human Parts As Adaptive Points

Towards High Performance One-Stage Human Pose Estimation.

SMPR: Single-Stage Multi-Person Pose Regression

EPOCH: Jointly Estimating the 3D Pose of Cameras and Humans

A Compact and Powerful Single-Stage Network for Multi-Person Pose Estimation

Hierarchical Graph Neural Network for Human Pose Estimation

Densely Connected Attentional Pyramid Residual Network for Human Pose Estimation.

Research on Multi-level Attention-based Human Pose Estimation

Human Pose Estimation Via Parse Graph of Body Structure