GitHub

ReSurgSAM2: Referring Segment Anything in Surgical Video via Credible Long-term Tracking

Official implementation for ReSurgSAM2, an innovative model that leverages the power of the Segment Anything Model 2 (SAM2), integrating it with credible long-term tracking for real-time surgical ...
GitHub

OpenWorldSAM: Extending SAM2 for Universal Image Segmentation with Language Prompts [NeurIPS 2025]

OpenWorldSAM pushes the boundaries of SAM2 by enabling open-vocabulary segmentation with flexible language prompts. [2026-1-4]: Demo release: we’ve added simple demos to run OpenWorldSAM on images ...
IEEE

HAM-SAM2: Enhancing SAM2 for Visual Object Tracking with Adaptive Motion Modeling and Hierarchical Memory Bank

Abstract: Visual object tracking remains challenging under complex motion, long-term occlusion, and distractor interference. The Segment Anything Model 2 (SAM2) provides a strong foundation for visual ...
IEEE

UniUltra: Interactive Parameter-Efficient SAM2 for Universal Ultrasound Segmentation

Abstract: The Segment Anything Model 2 (SAM2) demonstrates remarkable universal segmentation capabilities on natural images. However, its performance on ultrasound images is significantly degraded due ...