Few-shot Vision-based Human Activity Recognition with MLLM-based Visual Reinforcement Learning

Open in new window