Reproduce “OFL-SAM2: Prompt SAM2 with Online Few-shot Learner for Efficient Medical Image Segmentation” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Meng Lan, Lefei Zhang, Xiaomeng Li
OFL-SAM2 adds a learned target-specific feature stream and adaptive fusion to frozen SAM2 for prompt-free segmentation. It trains the mapping network on limited annotated data and updates it online when prediction confidence exceeds 0.8, combining target features with SAM2 memory features. On Synapse-CT with three training volumes, it achieves 82.52% mean Dice, 6.14 percentage points above FATE-SAM2; this result is limited to the evaluated datasets and protocols.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.