A Survey of Efficient Regression of General-Activity Human Poses from Depth Images

TL;DR

This survey reviews recent depth and RGB-D based methods for efficient regression of general-activity human poses, highlighting their advantages, limitations, and applications in various scenarios.

Contribution

It provides a comprehensive analysis of state-of-the-art depth-based human pose estimation techniques and compares their effectiveness across different contexts.

Findings

Depth sensors significantly improve 3D pose estimation accuracy.

Depth-based methods outperform RGB-only approaches in complex environments.

Different approaches have varying strengths and limitations depending on the scenario.

Abstract

This paper presents a comprehensive review on regression-based method for human pose estimation. The problem of human pose estimation has been intensively studied and enabled many application from entertainment to training. Traditional methods often rely on color image only which cannot completely ambiguity of joint 3D position, especially in the complex context. With the popularity of depth sensors, the precision of 3D estimation has significant improvement. In this paper, we give a detailed analysis of state-of-the-art on human pose estimation, including depth image based and RGB-D based approaches. The experimental results demonstrate their advantages and limitation for different scenarios.