结合小波融合和深度学习的脑胶质瘤自动分割  

A Method for Segmentation of Glioma on Multimodal Magnetic Resonance Images Based on Wavelet Fusion and Deep Learning

作　　者：宫进昌 王宇 王远军[1] GONG Jin-chang;WANG Yu;WANG Yuan-jun(Institute of Medical Imaging Engineering,University of Shanghai for Science and Technology,Shanghai 200093,China)

机构地区：[1]上海理工大学医学影像工程研究所,上海200093

出　　处：《波谱学杂志》2020年第2期131-143,共13页Chinese Journal of Magnetic Resonance

基　　金：国家自然科学基金资助项目(61201067);上海市自然科学基金资助项目(18ZR1426900).

摘　　要：针对水肿区域边界模糊和瘤内结构复杂多变导致的脑胶质瘤分割不精确问题,本文提出了一种基于小波融合和3D-UNet网络的脑胶质瘤磁共振图像自动分割算法.首先,对脑胶质瘤磁共振图像的T1、T1ce、T2、Flair四种模态进行小波融合以及偏置场校正;然后,提取待分类的图像块;再利用提取的图像块训练3D-UNet网络以对图像块中的像素进行分类;最后加载损失率较小的网络模型进行分割,并采用基于连通区域的轮廓提取方法,以降低假阳性率.对57组Brats2018(Brain Tumor Segmentation 2018)磁共振图像测试集进行分割的结果显示,肿瘤的整体、核心和水肿部分的平均分割准确率(DSC)分别达到90.64%、80.74%和86.37%,这表明该算法分割脑胶质瘤准确率较高,与金标准相近.相比多模态图像融合前,该算法在减少输入网络数据量和图像冗余信息的同时,还一定程度上解决了胶质瘤边界模糊、分割不精确的问题,提高了分割的准确度和鲁棒性.An automatic algorithm based on wavelet fusion is proposed for segmenting brain glioma with blurred boundaries and complex intratumoral structures on multimodal magnetic resonance(MR) images. Firstly, the T1, T1 ce, T2 and Flair MR images of brain glioma are fused with the bias field corrected. Secondly, the image blocks to be classified are extracted, and the 3 D-UNet network is trained to classify the pixels in the image blocks. Finally, the trained network model is used for segmentation, and the contour extraction method based on connected regions is used to reduce false positives. The average segmentation accuracy(DSC) of the whole, core and edema parts of the tumors was found to be 90.64%, 80.74% and 86.37%, respectively. The results indicated that the accuracy of the algorithm proposed was similar with or higher than the gold standard method. Compared with the method without multimodal image fusion, the algorithm proposed not only reduced the amount of data and redundant information in the input network, but also improved the accuracy and robustness of segmentation.

关 键 词：脑胶质瘤 多模态磁共振图像 小波融合 深度学习 图像分割 

分 类 号：TP391[自动化与计算机技术—计算机应用技术] O482.53[自动化与计算机技术—计算机科学与技术]