Research on Brain Glioma Segmentation Algorithm

2022 IEEE 2nd International Conference on Power, Electronics and Computer Applications (ICPECA) Pub Date : 2022-01-21 DOI:10.1109/ICPECA53709.2022.9719156

Shiqiang Zhang, Lei Shi, Xiaodong Cheng

{"title":"Research on Brain Glioma Segmentation Algorithm","authors":"Shiqiang Zhang, Lei Shi, Xiaodong Cheng","doi":"10.1109/ICPECA53709.2022.9719156","DOIUrl":null,"url":null,"abstract":"Due to the complexity of the imaging technology of medical imaging and the high heterogeneity of the surface of gliomas, image segmentation of human brain gliomas is one of the most challenging tasks in medical image analysis. This paper improves the UNet++ medical image segmentation network, in the down-sampling stage of the decoder, crosschannel fusion is carried out and deep supervision is introduced, at this time, the improved network can fuse coarsegrained semantics and fine-grained semantics at full scale. Experiments were performed on 335 images in the public BraTS brain tumor segmentation data set, using 2D and 3D comparative segmentation experiments to comprehensively evaluate the segmentation performance of the improved network, and compare the segmentation results with the results of UNet, UNet++, and UNet3 medical image segmentation networks. Among the four indicators of Dice Similarity Coefficient (DSC), 95% Hausdorff surface distance(HSD95), Sensitivity, and Positive Predictive Value (PPV), 2D contrast segmentation is achieved the mean values of the indicators are: 83.70%, 1.7, 88.40%, 84.96%; the mean values of the 3D contrast segmentation experiment are: 90.79%, 0.242, 91.23%, 91.06%. Compared with the segmentation result indicators of the other three networks, in the 2D comparison experiment, DSC increased by 1.82% on average, HSD95 decreased by 0.35 on average, Sensitivity increased by 2.13% on average, and PPV increased by 0.80% on average; in the 3D comparison experiment, DSC increased by 2.78% on average, HSD95 decreased by 0.076 on average, Sensitivity increased by 3.81% on average, and PPV increased by 0.68% on average. The experiments show that the improved algorithm makes the segmentation result of glioma and the gold standard overlap more in the region, and can better complete the segmentation of glioma. In clinical applications, it can help neurosurgeons to effectively separate brain tumors and tissues around the human brain, and achieve rapid computer diagnosis and treatment.","PeriodicalId":244448,"journal":{"name":"2022 IEEE 2nd International Conference on Power, Electronics and Computer Applications (ICPECA)","volume":"26 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"2022-01-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"1","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"2022 IEEE 2nd International Conference on Power, Electronics and Computer Applications (ICPECA)","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/ICPECA53709.2022.9719156","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}

引用次数: 1

Abstract

Due to the complexity of the imaging technology of medical imaging and the high heterogeneity of the surface of gliomas, image segmentation of human brain gliomas is one of the most challenging tasks in medical image analysis. This paper improves the UNet++ medical image segmentation network, in the down-sampling stage of the decoder, crosschannel fusion is carried out and deep supervision is introduced, at this time, the improved network can fuse coarsegrained semantics and fine-grained semantics at full scale. Experiments were performed on 335 images in the public BraTS brain tumor segmentation data set, using 2D and 3D comparative segmentation experiments to comprehensively evaluate the segmentation performance of the improved network, and compare the segmentation results with the results of UNet, UNet++, and UNet3 medical image segmentation networks. Among the four indicators of Dice Similarity Coefficient (DSC), 95% Hausdorff surface distance(HSD95), Sensitivity, and Positive Predictive Value (PPV), 2D contrast segmentation is achieved the mean values of the indicators are: 83.70%, 1.7, 88.40%, 84.96%; the mean values of the 3D contrast segmentation experiment are: 90.79%, 0.242, 91.23%, 91.06%. Compared with the segmentation result indicators of the other three networks, in the 2D comparison experiment, DSC increased by 1.82% on average, HSD95 decreased by 0.35 on average, Sensitivity increased by 2.13% on average, and PPV increased by 0.80% on average; in the 3D comparison experiment, DSC increased by 2.78% on average, HSD95 decreased by 0.076 on average, Sensitivity increased by 3.81% on average, and PPV increased by 0.68% on average. The experiments show that the improved algorithm makes the segmentation result of glioma and the gold standard overlap more in the region, and can better complete the segmentation of glioma. In clinical applications, it can help neurosurgeons to effectively separate brain tumors and tissues around the human brain, and achieve rapid computer diagnosis and treatment.

查看原文本刊更多论文

脑胶质瘤分割算法研究

由于医学影像学成像技术的复杂性和胶质瘤表面的高度异质性，人脑胶质瘤的图像分割是医学图像分析中最具挑战性的任务之一。本文对UNet++医学图像分割网络进行了改进，在解码器的下采样阶段进行了跨信道融合并引入了深度监督，此时改进的网络可以全面融合粗粒度语义和细粒度语义。对公开BraTS脑肿瘤分割数据集中的335幅图像进行实验，采用二维和三维对比分割实验，综合评价改进网络的分割性能，并将分割结果与UNet、UNet++、UNet3医学图像分割网络的分割结果进行比较。在Dice Similarity Coefficient (DSC)、95% Hausdorff surface distance(HSD95)、Sensitivity、Positive Predictive Value (PPV) 4个指标中，实现了2D对比分割，指标均值分别为:83.70%、1.7、88.40%、84.96%;三维对比分割实验的均值分别为:90.79%、0.242、91.23%、91.06%。与其他三个网络的分割结果指标相比，在二维对比实验中，DSC平均上升1.82%，HSD95平均下降0.35,Sensitivity平均上升2.13%，PPV平均上升0.80%;在3D对比实验中，DSC平均上升2.78%，HSD95平均下降0.076,Sensitivity平均上升3.81%，PPV平均上升0.68%。实验表明，改进后的算法使得胶质瘤的分割结果与金标准在区域上有更多的重叠，能够更好地完成胶质瘤的分割。在临床应用中，它可以帮助神经外科医生有效地分离脑肿瘤和人脑周围的组织，实现快速的计算机诊断和治疗。

本文章由计算机程序翻译，如有差异，请以英文原文为准。

求助全文

约1分钟内获得全文求助全文

来源期刊

2022 IEEE 2nd International Conference on Power, Electronics and Computer Applications (ICPECA)

自引率

0.00%

发文量