Mask-Enhanced Segment Anything Model for Tumor Lesion Semantic Segmentation

被引:0
|
作者
Shi, Hairong [1 ,2 ]
Han, Songhao [1 ,2 ]
Huang, Shaofei [3 ]
Liao, Yue [4 ]
Li, Guanbin [5 ]
Kong, Xiangxing [6 ]
Zhu, Hua [6 ]
Wang, Xiaomu [7 ]
Liu, Si [1 ,2 ]
机构
[1] Beihang Univ, 37 Xueyuan Rd, Beijing, Peoples R China
[2] Beihang Univ, Hangzhou Innovat Inst, 18 Chuanghui St, Hangzhou, Peoples R China
[3] Chinese Acad Sci, Inst Informat Engn, Beijing, Peoples R China
[4] Chinese Univ Hong Kong, Hong Kong, Peoples R China
[5] Sun Yat Sen Univ, Guangzhou, Peoples R China
[6] Peking Univ, Peking Univ Canc Hosp & Inst, Dept Nucl Med, Beijing, Peoples R China
[7] Nanjing Univ, Nanjing, Peoples R China
基金
北京市自然科学基金; 中国国家自然科学基金;
关键词
Tumor Lesion Segmentation; Medical Image Segmentation; Segment Anything Model; NETWORK;
D O I
10.1007/978-3-031-72111-3_38
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
Tumor lesion segmentation on CT or MRI images plays a critical role in cancer diagnosis and treatment planning. Considering the inherent differences in tumor lesion segmentation data across various medical imaging modalities and equipment, integrating medical knowledge into the Segment Anything Model (SAM) presents promising capability due to its versatility and generalization potential. Recent studies have attempted to enhance SAM with medical expertise by pre-training on large-scale medical segmentation datasets. However, challenges still exist in 3D tumor lesion segmentation owing to tumor complexity and the imbalance in foreground and background regions. Therefore, we introduce Mask-Enhanced SAM (M-SAM), an innovative architecture tailored for 3D tumor lesion segmentation. We propose a novel Mask-Enhanced Adapter (MEA) within M-SAM that enriches the semantic information of medical images with positional data from coarse segmentation masks, facilitating the generation of more precise segmentation masks. Furthermore, an iterative refinement scheme is implemented in M-SAM to refine the segmentation masks progressively, leading to improved performance. Extensive experiments on seven tumor lesion segmentation datasets indicate that our M-SAM not only achieves high segmentation accuracy but also exhibits robust generalization. The code is available at https://github.com/nanase1025/M-SAM.
引用
收藏
页码:403 / 413
页数:11
相关论文
共 50 条
  • [1] LeSAM: Adapt Segment Anything Model for Medical Lesion Segmentation
    Gu, Yunbo
    Wu, Qianyu
    Tang, Hui
    Mai, Xiaoli
    Shu, Huazhong
    Li, Baosheng
    Chen, Yang
    IEEE JOURNAL OF BIOMEDICAL AND HEALTH INFORMATICS, 2024, 28 (10) : 6031 - 6041
  • [2] SAM-Path: A Segment Anything Model for Semantic Segmentation in Digital Pathology
    Zhang, Jingwei
    Ma, Ke
    Kapse, Saarthak
    Saltz, Joel
    Vakalopoulou, Maria
    Prasanna, Prateek
    Samaras, Dimitris
    MEDICAL IMAGE COMPUTING AND COMPUTER ASSISTED INTERVENTION, MICCAI 2023 WORKSHOPS, 2023, 14393 : 161 - 170
  • [3] Enhanced Segment Anything Model for Accurate White Blood Cell Segmentation
    Zang, Yu
    Su, Yang
    Hu, Jun
    ELECTRONICS LETTERS, 2025, 61 (01)
  • [4] Parsing All Adverse Scenes: Severity-Aware Semantic Segmentation with Mask-Enhanced Cross-Domain Consistency
    Li, Fuhao
    Gong, Ziyang
    Deng, Yupeng
    Ma, Xianzheng
    Zhang, Renrui
    Ji, Zhenming
    Zhu, Xiangwei
    Zhang, Hong
    THIRTY-EIGHTH AAAI CONFERENCE ON ARTIFICIAL INTELLIGENCE, VOL 38 NO 12, 2024, : 13483 - 13491
  • [5] From SAM to CAMs: Exploring Segment Anything Model forWeakly Supervised Semantic Segmentation
    Kweon, Hyeokjun
    Yoon, Kuk-Jin
    2024 IEEE/CVF CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION (CVPR), 2024, : 19499 - 19509
  • [6] Superpixel-Guided Segment Anything Model for Liver Tumor Segmentation with Couinaud Segment Prompt
    Lyu, Fei
    Xu, Jingwen
    Zhu, Ye
    Wong, Grace Lai-Hung
    Yuen, Pong C.
    MEDICAL IMAGE COMPUTING AND COMPUTER ASSISTED INTERVENTION - MICCAI 2024, PT VIII, 2024, 15008 : 678 - 688
  • [7] PESAM: Privacy-Enhanced Segment Anything Model for Medical Image Segmentation
    Cai, Jiuyun
    Niu, Ke
    Pan, Yijie
    Tai, Wenjuan
    Han, Jiacheng
    ADVANCED INTELLIGENT COMPUTING TECHNOLOGY AND APPLICATIONS, PT II, ICIC 2024, 2024, 14863 : 94 - 105
  • [8] ProtoSAM-3D: Interactive semantic segmentation in volumetric medical imaging via a Segment Anything Model and mask-level prototypes
    Shen, Yiqing
    Dreizin, David
    Inigo, Blanca
    Unberath, Mathias
    COMPUTERIZED MEDICAL IMAGING AND GRAPHICS, 2025, 121
  • [9] Dual-Branch Fully Convolutional Segment Anything Model for Lesion Segmentation in Endoscopic Images
    He, Dongzhi
    Ma, Zeyuan
    Li, Chenxi
    Li, Yunqi
    IEEE ACCESS, 2024, 12 : 125654 - 125667
  • [10] Learning few-shot semantic segmentation with error-filtered segment anything model
    Feng, Chen-Bin
    Lai, Qi
    Liu, Kangdao
    Su, Houcheng
    Chen, Hao
    Luo, Kaixi
    Vong, Chi-Man
    VISUAL COMPUTER, 2025,