How to build the best medical image segmentation algorithm using foundation models: a comprehensive empirical study with Segment Anything Model
It provides systematic guidelines for fine-tuning foundation models in medical image segmentation, addressing a gap for researchers and practitioners in radiology, though it is incremental as it builds on existing SAM capabilities.
This work conducted a comprehensive empirical study to determine optimal fine-tuning strategies for the Segment Anything Model (SAM) in medical image segmentation, evaluating 18 combinations across 17 datasets and finding that fine-tuning SAM slightly outperforms previous methods, with parameter-efficient learning in both encoder and decoder yielding the best results.
Automated segmentation is a fundamental medical image analysis task, which enjoys significant advances due to the advent of deep learning. While foundation models have been useful in natural language processing and some vision tasks for some time, the foundation model developed with image segmentation in mind - Segment Anything Model (SAM) - has been developed only recently and has shown similar promise. However, there are still no systematic analyses or "best-practice" guidelines for optimal fine-tuning of SAM for medical image segmentation. This work summarizes existing fine-tuning strategies with various backbone architectures, model components, and fine-tuning algorithms across 18 combinations, and evaluates them on 17 datasets covering all common radiology modalities. Our study reveals that (1) fine-tuning SAM leads to slightly better performance than previous segmentation methods, (2) fine-tuning strategies that use parameter-efficient learning in both the encoder and decoder are superior to other strategies, (3) network architecture has a small impact on final performance, (4) further training SAM with self-supervised learning can improve final model performance. We also demonstrate the ineffectiveness of some methods popular in the literature and further expand our experiments into few-shot and prompt-based settings. Lastly, we released our code and MRI-specific fine-tuned weights, which consistently obtained superior performance over the original SAM, at https://github.com/mazurowski-lab/finetune-SAM.