Exploring Pose Priors for Human Pose Estimation with Joint Angle Representations

TL;DR

This paper investigates the use of pose priors based on joint angle representations to improve the physical plausibility of human pose estimation, emphasizing the importance of joint constraints like knee angles.

Contribution

It introduces a novel approach to incorporate joint angle-based pose priors, enhancing the accuracy and validity of human pose estimation models.

Findings

Pose priors effectively prevent physically impossible poses.

Joint angle constraints improve pose estimation accuracy.

Specific joint limits, like knee angles, are crucial for realistic poses.

Abstract

Pose Priors are critical in human pose estimation, since they are able to enforce constraints that prevent estimated poses from tending to physically impossible positions. Human pose generally consists of up to 22 Joint Angles of various segments, and their respective bone lengths, but the way these various segments interact can affect the validity of a pose. Looking at the Knee-Ankle segment alone, we can observe that clearly, the Knee cannot bend forward beyond it's roughly 90 degree point, amongst various other impossible poses below.