About
Dr. Rehg’s research focuses on the creation of AI methods for modeling and analyzing social and cognitive behavior and its emergence in child development. His lab has pioneered egocentric computer vision, which studies the visual world via the analysis of head-worn camera images. He and his collaborators developed an egocentric approach to automatically quantifying bouts of eye contact during naturalistic face-to-face interactions. It was the first example of an AI model performing at human-level accuracy in assessing a social communication behavior. His lab is developing computational methods for understanding the behavioral underpinnings of autism, with a focus on the development of novel diagnostic and therapeutic approaches. Other research efforts analyze wearable sensor data to model risk factors and develop interventions for chronic health conditions. Dr. Rehg was the lead PI of an NSF Expedition to develop novel computational approaches to modeling social and communicative behavior via multi-modal sensing.
Awards
- Distinguished Paper Award, Proceedings of the ACM on Interactive, Mobile, Wearable, and Ubiquitous Technologies, 2018
- Method of the Year Award, Nature Methods, 2012
- Best Paper Award, International Conference on Machine Learning, 2005
- Raytheon Faculty Fellowship, Georgia Institute of Technology, 2005
- Career Award, National Science Foundation, 2001
Relevant Publications
- Li, Y., Liu, M., & Rehg, J. M. (2021). In the Eye of the Beholder: Gaze and Actions in First Person Video. IEEE Transactions on Pattern Analysis and Machine Intelligence, 45(6), 6731-6747, DOI: 10.1109/TPAMI.2021.3051319
- Chong, E., Clark-Whitney, E., Southerland, A., Stubbs, E., Miller, C., Ajodan, E. L., Silverman, M. R., Lord, C., Rozga, A., Jones, R. M., & Rehg, J. M. (2020). Detection of eye contact with deep neural networks is as accurate as human experts. Nature Communications, 11(6386), 1–30. DOI: 10.1038/s41467-020-19712-x
- Li, Y., Hou, X., Koch, C., Rehg, J. M., & Yuille, A. L. (2014). The Secrets of Salient Object Segmentation. Proceedings IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 280–287. DOI: 10.1109/CVPR.2014.43