Revealing the unseen: Benchmarking video action recognition under occlusion

Open in new window