Pose Estimation is a topic in computer vision focusing on analyzing object movement, specifically that of humans. The goal is to identify key parts of the human body, like a person's head, shoulders, elbows, and hands. By pin-pointing those key parts, we can identify a person's movements, for example when someone is making a gesture as shown in the image above.

We are using pose estimation as an educational tool to aid instructors in calculating the number of times and duration they spend facing toward and away from their students, as shown in the image above. The goal is to determine if an instructor's movements have a significant impact on the learning outcomes of their students. To do this, we have encouraged instructors here at UTRGV to upload their lecture recordings via Pose Estimation, where the recording is analyzed using AI to calculate and document their movements into an excel file for further analysis.

Collaborators

(PI) Dr. Seokmin Kang, (Co-PI) Dr. Dong-Chul Kim, Joselito Guardado, Md Ashraful Islam, Junseong Park