计算机科学

计算机科学 ›› 2025, Vol. 52 ›› Issue (3): 41-49.doi: 10.11896/jsjkx.240300091

• 三维视觉与元宇宙 • 上一篇    下一篇

基于边缘增强的选择性特征融合肾癌三维CT图像分割

王涛1, 白雪飞1, 王文剑2,3   

  1. 1 山西大学计算机与信息技术学院 太原 030006
    2 计算智能与中文信息处理教育部重点实验室(山西大学) 太原 030006
    3 山西警察学院网络安全保卫系 太原 030401
  • 收稿日期:2024-03-13 修回日期:2024-08-13 出版日期:2025-03-15 发布日期:2025-03-07
  • 通讯作者: 王文剑(wjwang@sxu.edu.cn)
  • 作者简介:(wt2418385260@163.com)
  • 基金资助:
    国家自然科学基金(U21A20513,62076154)

Selective Feature Fusion for 3D CT Image Segmentation of Renal Cancer Based on Edge Enhancement

WANG Tao1, BAI Xuefei1, WANG Wenjian2,3   

  1. 1 School of Computer and Information Technology,Shanxi University,Taiyuan 030006,China
    2 Key Laboratory of Computational Intelligence and Chinese Information Processing(Shanxi University),Ministry of Education,Taiyuan 030006,China
    3 Department of Network Security,Shanxi Police College,Taiyuan 030401,China
  • Received:2024-03-13 Revised:2024-08-13 Online:2025-03-15 Published:2025-03-07
  • About author:WANG Tao,born in 2000,master candidate,is a member of CCF(No.P3207G).His main research interests include machine learning and image processing.
    WANG Wenjian,born in 1968,Ph.D,professor,is an outstanding member of CCF(No.16143D).Her main research interests include machine learning,computing intelligence and image proces-sing.
  • Supported by:
    National Natural Science Foundation of China(U21A20513,62076154).

PDF (PC)

295

摘要： 针对肾癌三维CT图像存在病变区域多尺度、边缘像素稀疏、对比度低以及肿瘤形状复杂且不规则等问题,提出一种基于边缘增强的选择性特征融合肾癌三维CT图像分割网络(EE-SFF U-Net)。EE-SFF U-Net采用基于U-Net的对称编解码网络架构,编码路径中包含一个用于强化边缘信息的边缘增强模块,可有效挖掘、利用浅层特征信息以缓解边缘像素稀疏问题,同时避免小目标的漏检。此外,在网络的跳跃连接中,设计一个选择性特征融合模块,使得深浅层特征相互补充,实现不同信息的有效聚合。最后提出一个综合Generalized Dice Loss和Focal Loss的混合损失函数,利用动态权重调整策略,实现损失函数的优化训练,并降低病变区域多尺度和肿瘤形状大小不规则带来的影响。所提方法在保证病变区域整体定位准确的同时,强化对小目标特征信息的挖掘利用,从而提高分割的准确性和鲁棒性。在KiTS19公开数据集上的实验结果表明,与其他分割算法相比,该方法各项指标表现良好,分割性能有显著提升。

关键词: 肾癌三维CT分割, 边缘增强, 选择性特征融合, 3D U-Net, 深度学习

Abstract: Aiming at the problems of multi-scale lesion areas,sparse edge pixels,low contrast,as well as complex and irregular tumor shape in 3D CT images of renal cancer,this paper proposes a selective feature fusion 3D CT image segmentation network based on edge enhancement(EE-SFF U-Net).EE-SFF U-Net adopts the symmetric encoder-decoder network architecture based on U-Net,and the encoding path contains an edge enhancement module for strengthening edge information,which can effectively mine and utilize shallow feature information to alleviate the sparsity problem of edge pixels and avoid missing detection of small targets.In addition,in the skip connections of the network,a selective feature fusion module is designed to make the deep and shallow features complement each other and realize the effective aggregation of different information.Finally,a hybrid loss function with Generalized Dice Loss and Focal Loss is proposed.The dynamic weight adjustment strategy is used to realize the optimal training of the loss function,and to improve the influence of multi-scale lesions and irregular tumor shape and size.The proposed method not only ensures the accuracy of the overall localization of the lesion area,but also strengthens the mining and utilization of small target feature information,so as to improve the accuracy and robustness of segmentation.The experimental results on KiTS19 public dataset show that the proposed method performs well in various indexes and significantly improves the segmentation performance compared with other segmentation algorithms.

Key words: 3D CT segmentation of renal cancer, Edge enhancement, Selective feature fusion, 3D U-Net, Deep learning

中图分类号: 

  • TP391
[1]ZHANG Y,HUA X L,SHI H Q,et al.Systematic analyses of the role of prognostic and immunological EIF3A,a reader protein,in clear cell renal cell carcinoma[J].Cancer Cell International,2021,21(1):680.
[2]Chinese Anti-Cancer Association Urological Tumor Committee.The 2018 Annual Report of China Cancer[M].Beijing:People’s Medical Publishing House,2019.
[3]KAUR R,JUNEJA M.A Survey of Kidney Segmentation Techniques in CT Images[J].Current Medical Imaging,2016,14(2):238-250.
[4]LEE H S,HONG H,KIM J.Detection and segmentation ofsmall renal masses in contrast-enhanced CT images using texture and context feature classification[C]//2017 IEEE 14th International Symposium on Biomedical Imaging(ISBI 2017).New York:IEEE,2017:583-586.
[5]BADURA P,WIECLAWEK W,PYCINSKI B.Automatic 3DSegmentation of Renal Cysts in CT[J].Information Technologies in Medicine,2016,471:149-163.
[6]LINGURARU M G,WANG S J,SHAH F,et al.Automated noninvasive classification of renal cancer on multiphase CT[J].Medical Physics,2011,38(10):5738-5746.
[7]BAE K,PARK B,SUN H L,et al.Segmentation of individual renal cysts from MR images in patients with autosomal dominant polycystic kidney disease[J].Clinical Journal of the American Society of Nephrology,2013,8(7):1089-1097.
[8]KAUR R,JUNEJA M,MANDAL A K.A hybrid edge-basedtechnique for segmentation of renal lesions in CT images[J].Multimedia Tools and Applications,2018,77(21):1-21.
[9]IHLESIAS J E,SABUNCU M R.Multi-atlas segmentation of biomedical images:a survey[J].Medical Image Analysis,2015,24(1):205-219.
[10]ÇICEK O,ABDULKADIR A,LIENKAMP S S,et al.3D U-Net:Learning dense volumetric segmentation from sparse annotation[C]//International Conference on Medical Image Computing and Computer-Assisted Intervention.Cham:Springer,2016:424-432.
[11]RONNEBERGER O,FISCHER P,BROX T.U-Net:Convolu-tional networks for biomedical image segmentation[C]//Medical Image Computing and Computer Assisted Intervention(MICCAI).Cham:Springer,2015:234-241.
[12]HATAMIZADEH A,TANG Y C,NATH V,et al.UNETR:Transformers for 3D medical image segmentation[C]//2022 IEEE/CVF Winter Conference on Applications of Computer Vision(WACV).New York:IEEE,2022,1748-1758.
[13]ZHOU Z W,RAHMAN S M,TAJBAKHSH N,et al.UNet++:A nested U-Net architecture for medical image segmentation[C]//Deep Learning in Medical Image Analysis and Multimodal Learning for Clinical Decision Support(DLMIA ML-CDS 2018).Cham:Springer,2018:3-11.
[14]LIN Y,ZHANG D,FANG X,et al.Rethinking boundary detection deep learning models medical image segmentation[C]//Information Processing in Medical Imaging(IPMI).Cham:Sprin-ger,2023:730-742.
[15]LI J,HU J P,QIAO M,et al.An Improved Method for IntensityInhomogeneity of Brain Image Segmentation[J].Journal of Chongqing Technology and Business University(Natural Science Edition),2023,40(1):34-39.
[16]SCHLEMPER J,OKTAY O,SCHAAP M,et al.Attention gated networks:Learning to leverage salient regions in medicalimages[J].Medical Image Analysis,2019,53:197-207.
[17]MYRONENKO A,SIDDIQUEE M,YANG D et al.Automated head and neck tumor segmentation from 3D PET/CT HECKTOR 2022 challenge report[C]//Head and Neck Tumor Segmentation and Outcome Prediction.Cham:Springer,2022:31-37.
[18]MOU L,ZHAO Y,FU H.CS2-Net:Deep learning segmentation of curvilinear structures in medical imaging[J].Medical Image Analysis,2021,67:101874.
[19]HU J,SHEN L,SUN G.Squeeze-and-Excitation Networks[C]//2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition.New York:IEEE,2018:7132-7141.
[20]MILLETARI F,NAVAB N,AHMADI S A.V-net:Fully convolutional neural networks for volumetric medical image segmentation[C]//3D Vision(3DV).New York:IEEE,2016:565-571.
[21]SUDRE C H,LI W,VERCAUTEREN T et al.Generalized Dice Overlap as a Deep Learning Loss Function for Highly Unba-lanced Segmentations[C]//Deep Learning in Medical Image Analysis and Multimodal Learning for Clinical Decision Support(DLMIA ML-CDS 2017).Cham:Springer,2017:240-248.
[22]LIN T,GOYAL P,GIRSHICK R,et al.Focal Loss for DenseObject Detection[C]//2017 IEEE International Conference on Computer Vision(ICCV).New York:IEEE,2017,324:2999-3007.
[23]NICHOLAS H,FABIAN I,KLAUS H.The state of the art in kidney and kidney tumor segmentation in contrast-enhanced CT imaging:Results of the KiTS19 challenge[J].Medical Image Analysis,2021,67:101821.
[24]ISENSEE F,MAIER-HEIN K H.An attempt at beating the 3D U-Net[J].arXiv:1908.02182,2019.
[25]HSIAO C,SUN T,LIN P,et al.A deep learning-based precision volume calculation approach for kidney and tumor segmentation on computed tomography images[J].Computer Methods and Programs in Biomedicine,2022,221:106861.
[26]CARDOSO M J,LI W,BROWN R,et al.Monai:An open-source framework for deep learning in healthcare[J].arXiv:2211.02701,2022.
[27]REYAD M,SARHAN A M,ARAFA M.A modified Adam algorithm for deep neural network optimization[J].Neural Computing and Applications,2023,35(23):17095-17112.
[28]ARWA M,AHMED A,FADWA A.The Impact of Multi-Optimizers and Data Augmentation on TensorFlow Convolutional Neural Network Performance[C]//2018 IEEE Conference on Multimedia Information Processing and Retrieval(MIPR).IEEE,2018:140-145.
[29]ZHAO R R,ZHOU Z J,YAO N,et al.Some novel Dice similarity measures for picture fuzzy sets and their applications[J].Engineering Applications of Artificial Intelligence,2024,138:109385.
[30]YU J,JIANG Y,WANG Z,et al.Unitbox:An advanced object detection network[C]//Proceedings of the 24th ACM International Conference on Multimedia.2016:516-520.
[31]ZHANG W L,LI R J,DENG H T,et al.Deep convolutionalneural networks for multi-modality isointense infant brain image segmentation[J].NeuroImage,2015,108:214-224.
[32]MINAEE S,BOYKOV Y,PORIKLI F,et al.Image Segmentation Using Deep Learning:A Survey[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2022,44(7):3523-3542.
Viewed
Full text


Abstract

Cited
  Shared   
  Discussed   
No Suggested Reading articles found!
渝ICP备16013121号-4
地址：重庆市渝北区洪湖西路18号    邮编：401121  
电话：023-63500828（编辑部） 023-67039612（会议/专题） 023-67039623（财务/发票）
E-mail：jsjkx12@163.com
本系统由北京玛格泰克科技发展有限公司设计开发