A novel approach to recognize events in videos using limited labeled data.
― 6 min read
Armin Mustafa
A new method creates better video captions by focusing on narratives and causality.
― 5 min read
A new method enhances action detection accuracy in overlapping video scenes.
― 7 min read
A new approach improves 3D scene reconstruction from a single photo, focusing on interactions.
― 4 min read
A new model efficiently integrates audio and visual data for video classification.
― 5 min read
A new method improves shadow detection without extensive data labeling.
― 5 min read