Video Gait Analysis

Project Department: Uni Research Computing (group: Center for Big Data Analysis)

About the project

Gait analysis is the systematic study of the (human) locomotion. As motion is affected by several disorders, gait analysis can significantly contribute to medical diagnosis.

An alternative to modern motion analysis laboratories, gait analysis based on simple videos are increasingly of interest. Recent developments in artificial intelligence are pushing the frontiers video processing. A method based on deep learning technologies (Goodfellow 2016) allows for human pose estimation from 2D images (Cao 2017). The corresponding software OpenPose is freely available (OpenPose).

A first investigation of the possibilities of OpenPose has been carried out by CBDA. See e.g. /media/ck_uploads/2018/09/06/openpose-example-v12.mp4

Bibliography:

* Cao et. al. (2017). Realtime Multi-Person 2D Pose Estimation using Part Affinity Fields, arXiv:1611.08050

* Goodfellow et. al. (2016). Deep Learning, The MIT Press

 

cp: 2018-12-17 20:17:35