Medical SAM Adapter: Adapting Segment Anything Model for Medical Image Segmentation

TL;DR

This paper introduces the Medical SAM Adapter, a lightweight method to adapt the Segment Anything Model for medical image segmentation by incorporating domain-specific knowledge and novel adaptation techniques, significantly improving performance.

Contribution

The paper presents Med-SA, a novel adaptation approach that enhances SAM for medical images without extensive fine-tuning, using SD-Trans and HyP-Adpt modules.

Findings

Med-SA outperforms state-of-the-art medical segmentation methods.

Only 2% of parameters are updated in Med-SA.

Effective adaptation for 3D medical images achieved.

Abstract

The Segment Anything Model (SAM) has recently gained popularity in the field of image segmentation due to its impressive capabilities in various segmentation tasks and its prompt-based interface. However, recent studies and individual experiments have shown that SAM underperforms in medical image segmentation, since the lack of the medical specific knowledge. This raises the question of how to enhance SAM's segmentation capability for medical images. In this paper, instead of fine-tuning the SAM model, we propose the Medical SAM Adapter (Med-SA), which incorporates domain-specific medical knowledge into the segmentation model using a light yet effective adaptation technique. In Med-SA, we propose Space-Depth Transpose (SD-Trans) to adapt 2D SAM to 3D medical images and Hyper-Prompting Adapter (HyP-Adpt) to achieve prompt-conditioned adaptation. We conduct comprehensive evaluation experiments on 17 medical image segmentation tasks across various image modalities. Med-SA outperforms several state-of-the-art (SOTA) medical image segmentation methods, while updating only 2\% of the parameters. Our code is released at https://github.com/KidsWithTokens/Medical-SAM-Adapter.