Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

Prompting Segment Anything Model with Domain-Adaptive Prototype for Generalizable Medical Image Segmentation

About

Deep learning based methods often suffer from performance degradation caused by domain shift. In recent years, many sophisticated network structures have been designed to tackle this problem. However, the advent of large model trained on massive data, with its exceptional segmentation capability, introduces a new perspective for solving medical segmentation problems. In this paper, we propose a novel Domain-Adaptive Prompt framework for fine-tuning the Segment Anything Model (termed as DAPSAM) to address single-source domain generalization (SDG) in segmenting medical images. DAPSAM not only utilizes a more generalization-friendly adapter to fine-tune the large model, but also introduces a self-learning prototype-based prompt generator to enhance model's generalization ability. Specifically, we first merge the important low-level features into intermediate features before feeding to each adapter, followed by an attention filter to remove redundant information. This yields more robust image embeddings. Then, we propose using a learnable memory bank to construct domain-adaptive prototypes for prompt generation, helping to achieve generalizable medical image segmentation. Extensive experimental results demonstrate that our DAPSAM achieves state-of-the-art performance on two SDG medical image segmentation tasks with different modalities. The code is available at https://github.com/wkklavis/DAPSAM.

Zhikai Wei, Wenhui Dong, Peilin Zhou, Yuliang Gu, Zhou Zhao, Yongchao Xu• 2024

Related benchmarks

TaskDatasetResultRank
Optic Cup SegmentationRIGA+ Messidor Target Domains (test)
Score (Messidor Base 1)88.24
22
Optic Disc SegmentationRIGA+ Messidor Target Domains (test)
Messidor Base 1 Score96.34
22
Prostate SegmentationMulti-site Prostate MRI (leave-one-domain-out)
Site A vs Rest Score86.34
22
Prostate SegmentationMulti-site Prostate Site A
DSC86.34
12
Prostate SegmentationMulti-site Prostate Site D
DSC85.28
12
Prostate SegmentationMulti-site Prostate Site C
DSC70.81
12
Prostate SegmentationMulti-site Prostate Site E
DSC82.91
12
Prostate SegmentationMulti-site Prostate Average
DSC81.31
12
Prostate SegmentationMulti-site Prostate Site B
DSC0.8105
12
Prostate SegmentationMulti-site Prostate Site F
DSC81.48
12
Showing 10 of 10 rows

Other info

Code

Follow for update