Markerless motion capture systems for tracking of persons in forensic biomechanics: an overview

摘要

Markerless motion capture is a pronounced topic in computer vision. In forensic science, markerless motion capture can be an important tool for identification through gait analysis. Recent studies of gait analysis in forensic science have shown that individuals can be identified when analysing the postures from a sagittal viewpoint. Although integrating all three dimensions (3D) might improve the results considerably. The purpose of this paper is to give an overview of the 3D multi-view markerless motion capture systems which could be applicable for 3D gait analysis. This paper contains presentation of the state of the art in obtaining 3D poses from multi-camera configurations and discussion of the cons and pros of the approaches.