sam

A technique is shared for using the Segment Anything Model (SAM) to segment images without prompting by exploiting similarities in microscopy images, enabling zero-shot mask propagation similar to video.

360 AI Research Institute's MoSA (Motion-Grounded Segment Anything), accepted at ECCV 2026, trains the Segment Anything Model to segment objects using motion cues from unlabeled video, generating millions of pseudo-labels without manual annotation.

This paper introduces SAM-MT, an extension of SAM2 for real-time interactive multi-target video segmentation, achieving high FPS independent of target count.

Group Prompting introduces a training-free framework for cell instance segmentation that requires only one click per cell type, using the Segment Anything Model's feature space to recursively expand prompts, achieving competitive performance without training.

Porting SAM 2.1 models to Apple silicon with MLX, achieving 1.25x inference speed increase on the small model, with quantized versions planned.

Researchers at the University of Pennsylvania are using AI models like DINO and SAM to automate and modernize medical triage in emergency response.