Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

ASLseg: Adapting SAM in the Loop for Semi-supervised Liver Tumor Segmentation

Shiyun Chen, Li Lin, Pujin Cheng, Xiaoying Tang

TL;DR

This work tackles liver tumor segmentation under limited annotations by integrating the Segment Anything Model (SAM) with a three-stage semi-supervised framework (ASLseg). The method jointly leverages SSL for domain-specific knowledge, fine-tunes SAM on labeled data, and employs an adaptation network to refine SAM predictions, producing reliable pseudo-labels to expand the labeled set iteratively. Key contributions include a novel three-stage pipeline, synthetic-pseudo-label–driven adaptation training, and extensive LiTS experiments showing state-of-the-art performance with only 10% labeled data. The approach reduces misdetections and boundary errors, enabling more accurate and robust liver tumor segmentation in data-scarce clinical settings.

Abstract

Liver tumor segmentation is essential for computer-aided diagnosis, surgical planning, and prognosis evaluation. However, obtaining and maintaining a large-scale dataset with dense annotations is challenging. Semi-Supervised Learning (SSL) is a common technique to address these challenges. Recently, Segment Anything Model (SAM) has shown promising performance in some medical image segmentation tasks, but it performs poorly for liver tumor segmentation. In this paper, we propose a novel semi-supervised framework, named ASLseg, which can effectively adapt the SAM to the SSL setting and combine both domain-specific and general knowledge of liver tumors. Specifically, the segmentation model trained with a specific SSL paradigm provides the generated pseudo-labels as prompts to the fine-tuned SAM. An adaptation network is then used to refine the SAM-predictions and generate higher-quality pseudo-labels. Finally, the reliable pseudo-labels are selected to expand the labeled set for iterative training. Extensive experiments on the LiTS dataset demonstrate overwhelming performance of our ASLseg.

ASLseg: Adapting SAM in the Loop for Semi-supervised Liver Tumor Segmentation

TL;DR

This work tackles liver tumor segmentation under limited annotations by integrating the Segment Anything Model (SAM) with a three-stage semi-supervised framework (ASLseg). The method jointly leverages SSL for domain-specific knowledge, fine-tunes SAM on labeled data, and employs an adaptation network to refine SAM predictions, producing reliable pseudo-labels to expand the labeled set iteratively. Key contributions include a novel three-stage pipeline, synthetic-pseudo-label–driven adaptation training, and extensive LiTS experiments showing state-of-the-art performance with only 10% labeled data. The approach reduces misdetections and boundary errors, enabling more accurate and robust liver tumor segmentation in data-scarce clinical settings.

Abstract

Liver tumor segmentation is essential for computer-aided diagnosis, surgical planning, and prognosis evaluation. However, obtaining and maintaining a large-scale dataset with dense annotations is challenging. Semi-Supervised Learning (SSL) is a common technique to address these challenges. Recently, Segment Anything Model (SAM) has shown promising performance in some medical image segmentation tasks, but it performs poorly for liver tumor segmentation. In this paper, we propose a novel semi-supervised framework, named ASLseg, which can effectively adapt the SAM to the SSL setting and combine both domain-specific and general knowledge of liver tumors. Specifically, the segmentation model trained with a specific SSL paradigm provides the generated pseudo-labels as prompts to the fine-tuned SAM. An adaptation network is then used to refine the SAM-predictions and generate higher-quality pseudo-labels. Finally, the reliable pseudo-labels are selected to expand the labeled set for iterative training. Extensive experiments on the LiTS dataset demonstrate overwhelming performance of our ASLseg.
Paper Structure (13 sections, 5 equations, 3 figures, 2 tables)

This paper contains 13 sections, 5 equations, 3 figures, 2 tables.

Table of Contents

  1. Introduction
  2. Methods
  3. Overall Framework
  4. Semi-Supervised Learning
  5. Medical SAM Fine-Tuning and Utilization
  6. Adaptation Network
  7. Application of Liver Tumor Segmentation Task
  8. EXPERIMENTS AND RESULTS
  9. Dataset
  10. Implementation Details
  11. Comparison and Ablation Experiments
  12. CONCLUSION
  13. Acknowledgments

Figures (3)

  • Figure 1: Overview of our proposed framework. Stage I is the SSL-specific training and inference module, Stage II is the fine-tuned medical SAM inference module, Stage III is the adaptation network inference module.
  • Figure 2: The training strategy of the adaptation network.
  • Figure 3: Visualize comparison results on the LiTS dataset.