English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Being a fundamental component in training and inference, data processing hasnot been systematically considered in human pose estimation community, to thebest of our knowledge. In this paper, we focus on this problem and find thatthe devil of human pose estimation evolution is in the biased data processing.Specifically, by investigating the standard data processing in state-of-the-artapproaches mainly including coordinate system transformation and keypointformat transformation (i.e., encoding and decoding), we find that the resultsobtained by common flipping strategy are unaligned with the original ones ininference. Moreover, there is a statistical error in some keypoint formattransformation methods. Two problems couple together, significantly degrade thepose estimation performance and thus lay a trap for the research community.This trap has given bone to many suboptimal remedies, which are alwaysunreported, confusing but influential. By causing failure in reproduction andunfair in comparison, the unreported remedies seriously impedes thetechnological development. To tackle this dilemma from the source, we proposeUnbiased Data Processing (UDP) consist of two technique aspect for the twoaforementioned problems respectively (i.e., unbiased coordinate systemtransformation and unbiased keypoint format transformation). As amodel-agnostic approach and a superior solution, UDP successfully pushes theperformance boundary of human pose estimation and offers a higher and morereliable baseline for research community. Code is public available inhttps://github.com/HuangJunJie2017/UDP-Pose

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