Research Overview

The CLAIRE (Cross-Referencing Labels, Actions, and Interactions for Robust Explanations) project aims to integrate YOLO detections and LLM vision to analyze video frames, cross-reference object/person–person/person interactions, and generate detailed reports of the video frame's scene.

This research explores both the benefits and restraints of utilizing AI models in action recognition video analysis. The project extends to CCTV footage as well as everyday video recordings, evaluating the strengths and shortcomings of AI-driven computer vision technologies in real-world applications.