《计算机应用》唯一官方网站 ›› 2025, Vol. 45 ›› Issue (10): 3390-3398.DOI: 10.11772/j.issn.1001-9081.2024101555

• 前沿与综合应用 • 上一篇    

轻量级微调SAM的结肠息肉分割方法SAMCP

刘娜1,2, 封筠1,2(), 霍一儒1,2, 王弘扬1,2, 杨柳1,2   

  1. 1.石家庄铁道大学 信息科学与技术学院,石家庄 050043
    2.石家庄市人工智能重点实验室(石家庄铁道大学),石家庄 050043
  • 收稿日期:2024-11-04 修回日期:2025-02-21 接受日期:2025-02-24 发布日期:2025-02-27 出版日期:2025-10-10
  • 通讯作者: 封筠
  • 作者简介:刘娜(2000—),女,河北张家口人,硕士研究生,CCF会员,主要研究方向:计算机视觉、图像分割
    封筠(1971—),女,河北石家庄人,教授,博士,CCF会员,主要研究方向:计算机视觉、机器学习 Email:fengjun@stdu.edu.cn
    霍一儒(2000—),男,河北邢台人,硕士研究生,主要研究方向:计算机视觉、图像分割
    王弘扬(2000—),男,河北张家口人,硕士研究生,主要研究方向:多模态大模型
    杨柳(1999—),女,河北唐山人,硕士研究生,主要研究方向:图像处理。
  • 基金资助:
    河北省自然科学基金资助项目(F2024210005)

SAMCP: lightweight fine-tuned SAM method for colon polyp segmentation

Na LIU1,2, Jun FENG1,2(), Yiru HUO1,2, Hongyang WANG1,2, Liu YANG1,2   

  1. 1.School of Information Science and Technology,Shijiazhuang Tiedao University,Shijiazhuang Hebei 050043,China
    2.Shijiazhuang Key Laboratory of Artificial Intelligence (Shijiazhuang Tiedao University),Shijiazhuang Hebei 050043,China
  • Received:2024-11-04 Revised:2025-02-21 Accepted:2025-02-24 Online:2025-02-27 Published:2025-10-10
  • Contact: Jun FENG
  • About author:LIU Na, born in 2000, M. S. candidate. Her research interests include computer vision, image segmentation.
    FENG Jun, born in 1971, Ph. D., professor. Her research interests include computer vision, machine learning.
    HUO Yiru, born in 2000, M. S. candidate. His research interests include computer vision, image segmentation.
    WANG Hongyang, born in 2000, M. S. candidate. His research interests include multimodal large models.
    YANG Liu, born in 1999, M. S. candidate. Her research interests include image processing.
  • Supported by:
    Natural Science Foundation of Hebei Province(F2024210005)

摘要:

在胃肠道内窥镜图像处理中,精准分割结肠息肉具有重要的临床意义。传统分割方法常因细节捕捉不足和对大规模数据的依赖,在应对复杂形态的息肉时表现不佳。尽管分割一切模型(SAM)在自然图像分割中取得显著进展,但由于自然图像与医学图像存在域差异,现有的SAM方法在结肠息肉分割任务上仍难以取得理想效果。为解决这一问题,基于SAM架构提出一种轻量级微调结肠息肉分割方法(SAMCP)。该方法引入精简适配器模块,重点关注通道维度信息,采用Dice和交并比(IoU)简化联合损失函数,并在训练时冻结原始图像编码器和提示编码器的参数,以低训练成本提升结肠息肉分割性能。在3个公开数据集上与9种先进方法的对比实验结果表明,相较于SAM方法,SAMCP在Kvasir-SEG数据集上的Dice和IoU值分别提高了56.7%和84.5%,在CVC-ClinicDB数据集上的Dice和IoU值分别提高了46.0%和86.0%,在CVC-ColonDB数据集上的Dice和IoU值分别提高了95.3%和122.2%,超过目前SAM-based类方法的最佳性能。在引入点提示的情况下,即使只使用1次点击,SAMCP仍能优于其他SAM-based方法。以上验证了SAMCP在处理复杂形状和局部细节时表现出色,可为医生提供更精确的分割指导。

关键词: 结肠息肉分割, 分割一切模型, 适配器, 损失函数, 轻量级微调

Abstract:

Precise segmentation of colon polyps in gastrointestinal endoscopy images holds significant clinical value. However, the traditional segmentation methods often struggle with capturing enough fine details and rely on large-scale data heavily, leading to poor performance when addressing complex polyp morphologies. Although Segment Anything Model (SAM) has notable progress in natural image segmentation, the ideal effect in polyp segmentation task cannot be achieved by SAM methods due to domain differences between natural and medical images. To address this issue, a lightweight fine-tuning method based on SAM architecture was proposed, named Segment Anything Model for Colon Polyps (SAMCP). In this method, a streamlined adapter module focusing on channel-dimension information was introduced, a joint loss function was simplified using Dice and Intersection over Union (IoU), and parameters of the original image encoder and prompt encoder were frozen during training to enhance polyp segmentation performance with low training cost. Experimental results on three public datasets comparing SAMCP with nine advanced methods demonstrate that SAMCP outperforms other SAM methods. Specifically, SAMCP improves the Dice and IoU values by 56.7% and 84.5%, respectively, on the Kvasir-SEG dataset, by 46.0% and 86.0%, respectively, on the CVC-ClinicDB, and by 95.3% and 122.2%, respectively, on the CVC-ColonDB dataset, surpassing the current best performance of SAM-based methods. With the introduction of point-based prompts, even with a single click, SAMCP can also outperform other SAM-based methods. The above validates that SAMCP performs well in handling complex shapes and local details, providing physicians with more precise segmentation guidance.

Key words: colon polyp segmentation, Segment Anything Model (SAM), adapter, loss function, lightweight fine-tuning

中图分类号: