MotionSynth: Digital Synthesis of Human Motions
The goal of this PhD project is the AI-based synthesis of (individual) human motion data. Using the example of rehabilitation medicine, functional movements are to be synthesized, e.g. a gripping movement of Parkinson's patients with grade II, without actually recording such real people on a large scale in a time-consuming and costly procedure. The synthesis takes place in a processing pipeline to be developed for human motion data. This software and the motion data it generates are intended to support therapists and medical professionals, among others in the analysis of human motion. In addition, the pipeline supports users in the optimal selection of sensors and their positioning in real measuring stations. An abstraction to generic interfaces for input, analysis and synthesis as well as for the output of human motions is to be derived from this special application in rehabilitation medicine. This is to ensure the adaptation of the pipeline for other areas of application.
Funding: ESF
