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.

Single-Stage Multi-Person Pose Machines

SMPR: Single-Stage Multi-Person Pose Regression

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

Rethinking the Person Localization for Single-Stage Multi-Person Pose Estimation

Unsupervised Universal Hierarchical Multi-Person 3D Pose Estimation for Natural Scenes

SMAP: Single-Shot Multi-person Absolute 3D Pose Estimation

Rethinking on Multi-Stage Networks for Human Pose Estimation

Shape and Pose Estimation for Closely Interacting Persons Using Multi-view Images.

InsPose: Instance-Aware Networks for Single-Stage Multi-Person Pose Estimation

Multi-Person 3d Pose Estimation From Monocular Image Sequences

Single-view Multi-Human Pose Estimation by Attentive Cross-Dimension Matching.

Direct Multi-view Multi-person 3D Pose Estimation

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

Multi-person 3D pose estimation from unlabelled data

Fast and Robust Multi-Person 3D Pose Estimation from Multiple Views

AdaptivePose: Human Parts As Adaptive Points

Multi-person Pose Estimation Based on Graph Grouping Optimization

Recurrent 3D Pose Sequence Machines

Skeleton Cluster Tracking for robust multi-view multi-person 3D human pose estimation

Multi-Person Full Body Pose Estimation

Full-body Motion Capture for Multiple Closely Interacting Persons.