Venus: Enhancing Online Video Understanding with Edge Memory

This research introduces Venus, a novel system designed to improve online video understanding using Vision-Language Models (VLMs) by efficiently managing memory and retrieval at the edge. The system's effectiveness and potential for real-time video analysis warrant further investigation and evaluation within various application domains.