Explicit Box Detection Unifies End-to-End Multi-Person Pose Estimation

3 Feb 2023  ·  Jie Yang, Ailing Zeng, Shilong Liu, Feng Li, Ruimao Zhang, Lei Zhang ·

This paper presents a novel end-to-end framework with Explicit box Detection for multi-person Pose estimation, called ED-Pose, where it unifies the contextual learning between human-level (global) and keypoint-level (local) information. Different from previous one-stage methods, ED-Pose re-considers this task as two explicit box detection processes with a unified representation and regression supervision. First, we introduce a human detection decoder from encoded tokens to extract global features. It can provide a good initialization for the latter keypoint detection, making the training process converge fast. Second, to bring in contextual information near keypoints, we regard pose estimation as a keypoint box detection problem to learn both box positions and contents for each keypoint. A human-to-keypoint detection decoder adopts an interactive learning strategy between human and keypoint features to further enhance global and local feature aggregation. In general, ED-Pose is conceptually simple without post-processing and dense heatmap supervision. It demonstrates its effectiveness and efficiency compared with both two-stage and one-stage methods. Notably, explicit box detection boosts the pose estimation performance by 4.5 AP on COCO and 9.9 AP on CrowdPose. For the first time, as a fully end-to-end framework with a L1 regression loss, ED-Pose surpasses heatmap-based Top-down methods under the same backbone by 1.2 AP on COCO and achieves the state-of-the-art with 76.6 AP on CrowdPose without bells and whistles. Code is available at https://github.com/IDEA-Research/ED-Pose.

PDF Abstract
Task Dataset Model Metric Name Metric Value Global Rank Uses Extra
Training Data
Result Benchmark
Multi-Person Pose Estimation CrowdPose ED-Pose (Swin-L) mAP @0.5:0.95 76.6 # 4
AP Easy 83.0 # 4
AP Medium 77.3 # 4
AP Hard 68.3 # 4
2D Human Pose Estimation Human-Art ED-Pose (R50) AP 0.723 # 2
AP (gt bbox) / # 11

Methods