Jae Shin Yoon Receives Top Graduate Research Award

PhD Student Jae Shin Yoon
November 13, 2019

Ph.D. student Jae Shin Yoon won the 2018-2019 Graduate Research and Discussion Seminars (GRaDS) research award for his talk “Human Behavioral Imaging”. This award is selected by Computer Science & Engineering faculty committees.

In this talk, Jae Shin shared highlights from two recent papers: “3D Semantic Trajectory Reconstruction from 3D Pixel Continuum” and “Self-Supervised Adaptation of High-Fidelity Face Models for Monocular Performance Tracking”.

He presented a framework for behavioral imaging---to computationally measure human behaviors with millimeter accuracy using a system of RGB cameras. This research is critically important because as the world’s population grows older in the coming years, their physical decline can reduce their mobility and increase the risk of falls. While many AI systems and sensors (e.g., Amazon Cloud) are poised to enter into our daily space, ubiquitously, their ability to measure and predict the behaviors of the elderly is still premature. The key challenge is that the behavioral cues indicative of a future action (e.g., fall) appear at a microscopic scale (e.g., pain facial expression and hand tremor) where the existing AIs are by large ignorant of such subtle cues. Furthermore, the application of behavioral imaging is not limited to elderly activity monitoring, but can also be an enabling factor of computational behavioral science, e.g., automatic early detection of autism spectrum disorder, telepresence (VR/AR), and smart home facilitation.

Jae Shin Yoon is a Ph.D. student at the University of Minnesota advised by Professor Hyun Soo Park. His research interests are solving the fundamental problem of reconstructing and understanding the dynamic scenes using visual data (image and video) with machine learning and computer vision knowledge.

Congratulations to Jae Shin Yoon on receiving this remarkable recognition.

The GraDS seminars are student-organized lunch talks open to CS graduate students in any area. Students speak at GRaDS to practice for conference or job talks, inform their fellow students on their research, or simply build their presentation skills. GRaDS is held biweekly on Fridays. Students gather in Keller Hall 2-212 for a catered lunch and an hour of presentation. The next presenters are Lucas Kramer (November 15) and John Harwell (December 6).

CS News
Graduate News
Graduate Student Awards