Hongsong Wang

New model improves action retrieval from images using person, objects, and context.

A new framework improves model accuracy by tackling data imbalance while ensuring privacy.

New methods improve detection of rare actions in videos using innovative approaches.

Discover how text-to-motion technology is changing animated storytelling and robotics.

Learn how USDRL is changing the way we recognize human actions.

Learn how video anomaly detection identifies strange events in footage.