Multi-scale nested UNet with transformer for colorectal polyp segmentation

被引:6
作者
Wang, Zenan [1 ]
Liu, Zhen [1 ]
Yu, Jianfeng [1 ]
Gao, Yingxin [1 ]
Liu, Ming [2 ]
机构
[1] Capital Med Univ, Beijing Chaoyang Hosp, Dept Gastroenterol, Clin Med Coll 3, Beijing, Peoples R China
[2] Hunan Key Lab Nonferrous Resources & Geol Hazard E, Changsha, Peoples R China
来源
JOURNAL OF APPLIED CLINICAL MEDICAL PHYSICS | 2024年 / 25卷 / 06期
关键词
colorectal polyp; deep learning; polyp segmentation; transformer; MISS RATE; COLONOSCOPY;
D O I
10.1002/acm2.14351
中图分类号
R8 [特种医学]; R445 [影像诊断学];
学科分类号
1002 ; 100207 ; 1009 ;
摘要
BackgroundPolyp detection and localization are essential tasks for colonoscopy. U-shape network based convolutional neural networks have achieved remarkable segmentation performance for biomedical images, but lack of long-range dependencies modeling limits their receptive fields.PurposeOur goal was to develop and test a novel architecture for polyp segmentation, which takes advantage of learning local information with long-range dependencies modeling.MethodsA novel architecture combining with multi-scale nested UNet structure integrated transformer for polyp segmentation was developed. The proposed network takes advantage of both CNN and transformer to extract distinct feature information. The transformer layer is embedded between the encoder and decoder of a U-shape net to learn explicit global context and long-range semantic information. To address the challenging of variant polyp sizes, a MSFF unit was proposed to fuse features with multiple resolution.ResultsFour public datasets and one in-house dataset were used to train and test the model performance. Ablation study was also conducted to verify each component of the model. For dataset Kvasir-SEG and CVC-ClinicDB, the proposed model achieved mean dice score of 0.942 and 0.950 respectively, which were more accurate than the other methods. To show the generalization of different methods, we processed two cross dataset validations, the proposed model achieved the highest mean dice score. The results demonstrate that the proposed network has powerful learning and generalization capability, significantly improving segmentation accuracy and outperforming state-of-the-art methods.ConclusionsThe proposed model produced more accurate polyp segmentation than current methods on four different public and one in-house datasets. Its capability of polyps segmentation in different sizes shows the potential clinical application
引用
收藏
页数:10
相关论文
共 50 条
  • [41] Multi-Scale Orthogonal Model CNN-Transformer for Medical Image Segmentation
    Zhou, Wuyi
    Zeng, Xianhua
    Zhou, Mingkun
    INTERNATIONAL JOURNAL OF PATTERN RECOGNITION AND ARTIFICIAL INTELLIGENCE, 2023, 37 (10)
  • [42] MULTI-SCALE CONVOLUTION-TRANSFORMER FUSION NETWORK FOR ENDOSCOPIC IMAGE SEGMENTATION
    Zou, Baosheng
    Zhou, Zongguang
    Han, Ying
    Li, Kang
    Wang, Guotai
    2023 IEEE 20TH INTERNATIONAL SYMPOSIUM ON BIOMEDICAL IMAGING, ISBI, 2023,
  • [43] MCSF-Net: a multi-scale channel spatial fusion network for real-time polyp segmentation
    Liu, Weikang
    Li, Zhigang
    Xia, Jiaao
    Li, Chunyang
    PHYSICS IN MEDICINE AND BIOLOGY, 2023, 68 (17)
  • [44] MULTI-SCALE SWIN TRANSFORMER ENABLED AUTOMATIC DETECTION AND SEGMENTATION OF LUNG METASTASES USING CT IMAGES
    Masood, Anum
    Naseem, Usman
    Razzak, Imran
    2023 IEEE 20TH INTERNATIONAL SYMPOSIUM ON BIOMEDICAL IMAGING, ISBI, 2023,
  • [45] MDA-Unet: A Multi-Scale Dilated Attention U-Net for Medical Image Segmentation
    Amer, Alyaa
    Lambrou, Tryphon
    Ye, Xujiong
    APPLIED SCIENCES-BASEL, 2022, 12 (07):
  • [46] Visual Mamba UNet fusion multi-scale attention and detail infusion for unsound corn kernels segmentation
    Zhao, Kuibin
    Zhang, Qinghui
    Wan, Chenxia
    Pan, Quan
    Qin, Yao
    SCIENTIFIC REPORTS, 2025, 15 (01):
  • [47] DEMF-Net: A dual encoder multi-scale feature fusion network for polyp segmentation
    Cao, Xiaorui
    Yu, He
    Yan, Kang
    Cui, Rong
    Guo, Jinming
    Li, Xuan
    Xing, Xiaoxue
    Huang, Tao
    BIOMEDICAL SIGNAL PROCESSING AND CONTROL, 2024, 96
  • [48] CafeNet: A Novel Multi-Scale Context Aggregation and Multi-Level Foreground Enhancement Network for Polyp Segmentation
    Ji, Zhanlin
    Li, Xiaoyu
    Wang, Zhiwu
    Zhang, Haiyang
    Yuan, Na
    Zhang, Xueji
    Ganchev, Ivan
    INTERNATIONAL JOURNAL OF IMAGING SYSTEMS AND TECHNOLOGY, 2024, 34 (05)
  • [49] BCNet: integrating UNet and transformer for blood cell segmentation
    Jiang, Yue
    Wang, Shuaidan
    Yao, Minjie
    Xiao, Qing
    Li, Yinghui
    Bai, Hua
    Zhang, Zhuo
    SIGNAL IMAGE AND VIDEO PROCESSING, 2025, 19 (01)
  • [50] Joint Transformer and Multi-scale CNN for DCE-MRI Breast Cancer Segmentation
    Chuanbo Qin
    Yujie Wu
    Junying Zeng
    Lianfang Tian
    Yikui Zhai
    Fang Li
    Xiaozhi Zhang
    Soft Computing, 2022, 26 : 8317 - 8334