Frequency Adapter with SAM for Generalized Medical Image Segmentation

TL;DR

This paper introduces FSAM, a novel frequency-based domain generalization framework using SAM and LoRA to improve medical image segmentation across diverse datasets.

Contribution

The work proposes a frequency adapter with SAM and LoRA for effective domain generalization in medical image segmentation, emphasizing high-frequency feature extraction.

Findings

FSAM outperforms existing DG methods on fundus and prostate datasets.

Frequency domain features improve model robustness to domain shifts.

Incorporating LoRA enables efficient fine-tuning of SAM for medical segmentation.

Abstract

Medical image segmentation is a critical task in computer-aided diagnosis and treatment planning. However, deep learning models often struggle to generalize across datasets due to domain shifts arising from variations in imaging protocols, scanner types, and patient populations. Traditional domain generalization (DG) methods utilize causal feature learning, adversarial consistency, and style augmentation to improve segmentation robustness. While effective, these approaches rely on explicit feature alignment, adversarial objectives, or handcrafted augmentations, which may not fully exploit the capabilities of foundation models. Recently, the Segment Anything Model (SAM) has demonstrated strong generalization capabilities in segmentation tasks. SAM-based DG methods attempt to improve medical image segmentation. However, these approaches primarily operate in the spatial domain and overlook frequency-based discrepancies that significantly affect model robustness. In this work, we propose Frequency-based Domain Generalization with SAM (FSAM), a novel framework that integrates Low-Rank Adaptation (LoRA) for efficient fine-tuning and a frequency adapter to incorporate frequency-domain representations for single-source domain generalization. FSAM enhances SAM's segmentation robustness by extracting domain-invariant high-frequency features, mitigating frequency-related domain shifts. Experimental results on fundus and prostate datasets demonstrate that FSAM outperforms existing traditional DG and SAM-based DG approaches in domain generalization. Codes and pre-trained models will be made available on GitHub.