The Devil is in the Details: Delving into Unbiased Data Processing for Human Pose Estimation

Abstract

Being a fundamental component in training and inference, data processing has not been systematically considered in human pose estimation community, to the best of our knowledge. In this paper, we focus on this problem and find that the devil of human pose estimation evolution is in the biased data processing. Specifically, by investigating the standard data processing in state-of-the-art approaches mainly including coordinate system transformation and keypoint format transformation (i.e., encoding and decoding), we find that the results obtained by common flipping strategy are unaligned with the original ones in inference. Moreover, there is a statistical error in some keypoint format transformation methods. Two problems couple together, significantly degrade the pose estimation performance and thus lay a trap for the research community. This trap has given bone to many suboptimal remedies, which are always unreported, confusing but influential. By causing failure in reproduction and unfair in comparison, the unreported remedies seriously impedes the technological development. To tackle this dilemma from the source, we propose Unbiased Data Processing (UDP) consist of two technique aspect for the two aforementioned problems respectively (i.e., unbiased coordinate system transformation and unbiased keypoint format transformation). As a model-agnostic approach and a superior solution, UDP successfully pushes the performance boundary of human pose estimation and offers a higher and more reliable baseline for research community. Code is public available in https://github.com/HuangJunJie2017/UDP-Pose

project:https://github.com/HuangJunJie2017/UDP-Pose

References in corpus (7)

• Joint Training of a Convolutional Network and a Graphical Model for Human Pose Estimation
• Multi-task Deep Learning for Real-Time 3D Human Pose Estimation and Action Recognition
• Rethinking on Multi-Stage Networks for Human Pose Estimation
• Cascade Feature Aggregation for Human Pose Estimation
• Single-Stage Multi-Person Pose Machines
• Exploiting Offset-guided Network for Pose Estimation and Tracking
• A Context-and-Spatial Aware Network for Multi-Person Pose Estimation

Cited by in corpus (9)

• TFPose: Direct Human Pose Estimation with Transformers
• Bottom-Up Human Pose Estimation Via Disentangled Keypoint Regression
• SHaRPose: Sparse High-Resolution Representation for Human Pose Estimation
• AID: Pushing the Performance Boundary of Human Pose Estimation with Information Dropping Augmentation
• Graph-PCNN: Two Stage Human Pose Estimation with Graph Pose Refinement
• Deep Dual Consecutive Network for Human Pose Estimation
• Pose Recognition with Cascade Transformers
• Train Your Data Processor: Distribution-Aware and Error-Compensation Coordinate Decoding for Human Pose Estimation
• The Influence of Faulty Labels in Data Sets on Human Pose Estimation