Thorsten Hempel

Papers

Total Citations

H-Index

About

Thorsten Hempel is a computer vision and human-computer interaction researcher whose work bridges the gap between machine perception and real-world applications. His research focuses primarily on gaze estimation and facial expression analysis, with particular emphasis on developing robust systems for human-robot collaboration (HRC) and related domains such as autonomous driving and virtual reality. Hempel's most notable contribution lies in advancing gaze estimation techniques through deep learning. His 2024 work on fine-grained gaze estimation introduces an innovative combination of regression and classification losses within convolutional neural network (CNN) architectures, achieving improved accuracy in predicting gaze angles in unconstrained, real-world environments — a notoriously challenging problem in the field. This paper has already attracted 6 citations, signaling growing interest from the research community. Complementing this, his 2021 study on facial action recognition in aggravated HRC scenarios demonstrates a thoughtful application of affective computing to workplace safety, using automated facial expression recognition to assess human psychological responses — such as fear or irritation — when collaborating with powerful industrial robots. With 3 citations, this work contributes meaningfully to the emerging conversation around human-centered robotics design, making Hempel a researcher to watch at the intersection of computer vision and human-aware intelligent systems.