WZH

SAM2-UNet. [VINT 2026] SAM2-UNet: Segment Anything 2 Makes Strong Encoder for Natural and Medical Image Segmentation

SAM3-UNet. Python

SAM2-UNeXT. Integrating SAM2 with DINOv2/v3 for segmentation

DINOv3-UNet. Adapting DINOv3 to downstream segmentation tasks

BMANet. [BSPC 2025] BMANet: Boundary-guided multi-level attention network for polyp segmentation in colonoscopy images

HFS-SAM2. [IEEE SPL 2025] HFS-SAM2: Segment Anything Model 2 with High-Frequency Feature Supplementation for Camouflaged Object Detection

MoE-Polyp. [MMAsia 2024] MoE-Polyp: Shifting More Attention to Small Polyp Segmentation via Mixture-of-Experts