Injury analysis may be one of the most beneficial applications of deep learning based human pose estimation. To facilitate further research on this topic, we provide an injury specific 2D dataset for alpine skiing, covering in total 533 images. We further propose a post processing routine, that combines rotational information with a simple kinematic model. We could improve detection results in fall situations by up to 21% regarding the [email protected] metric.
@article{arxiv.2112.12193,
title = {Improved 2D Keypoint Detection in Out-of-Balance and Fall Situations -- combining input rotations and a kinematic model},
author = {Michael Zwölfer and Dieter Heinrich and Kurt Schindelwig and Bastian Wandt and Helge Rhodin and Joerg Spoerri and Werner Nachbauer},
journal= {arXiv preprint arXiv:2112.12193},
year = {2021}
}