MoSAM: Motion-Guided Segment Anything Model with Spatial-Temporal Memory Selection

VLM
Re-assign community