Attention U-Net在前列腺MRI分割中的应用  

MRI Data Segmentation of Prostate Based on Attention U-Net

作　　者：王卓雅 娄成雨 缪心闯 赖小波[1] WANG Zhuoya;LOU Chengyu;MIAO Xinchuang;LAI Xiaobo(College of Medical Technology and Information Engineering,Zhejiang Chinese Medical University,Hangzhou,China,310053)

机构地区：[1]浙江中医药大学医学技术与信息工程学院,杭州310053

出　　处：《福建电脑》2023年第10期26-29,共4页Journal of Fujian Computer

摘　　要：随着国内前列腺特异性抗原筛查的普及,每年检出的有前列腺问题的病例呈现出上升趋势。高精度的前列腺的可靠和自动分割对于医生诊断和治疗前列腺癌具有深远的意义。本文提出一种新的前列腺MRI数据分割方案,通过在U-Net引入Attention机制,以有效抑制无关区域中的激活,减少很多冗余信息。实验的结果表明,采用该方法所提取的特征对于前列腺MRI数据自动分割的精准性具有很大帮助。With the popularization of prostate-specific antigen(PSA)screening,the incidence of prostate cancer in China is showing a significant upward trend.Accurate and automated segmentation of the prostate plays a profound role in the diagnosis and treatment of prostate cancer.This article proposes a new prostate MRI data segmentation scheme by introducing an Attention mechanism in U-Net to effectively suppress activation in unrelated regions and reduce a lot of redundant information.The experimental results indicate that the features extracted using this method are of great help for the accuracy of automatic segmentation of prostate MRI data.

关 键 词：前列腺癌 磁共振图像 自动分割 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]