Human Pose Estimation Using RGBD Images

Abstract

This paper presents a method to improve the accuracy of human pose estimation using RGBD images. 3D pose estimation based on mapping depth images to RGB images has a problem. It is the error in the depth direction caused by the low accuracy of the depth image and the fact that the estimated points are on the surface of the body. This paper proposes a method to improve the accuracy by using a correction algorithm based on ICP. The method consists of a measurement system that measures RGBD images and estimates human posture, and a correction algorithm using ICP.