Person Re-identification through Hypergraphs

This project focused on the challenging area of Person Re-identification – a task that involves recognizing and tracking individuals across different cameras or video sequences. It aims to match a person’s identity in one video frame or camera view with their identity in another frame or view. This research addresses three primary challenges of person identification (i.e., illumination changes, viewpoint variations, and scale variations) to establish the correspondence among the persons observed across non-overlapping camera views.