串行处理卷积神经网络的海马子区分割  

Cascaded convolutional neural network based hippocampus subfields segmentation

作　　者：时永刚[1] 郝华胤 刘志文[1] 

机构地区：[1]北京理工大学信息与电子学院,北京100081

出　　处：《中国图象图形学报》2018年第1期74-83,共10页Journal of Image and Graphics

基　　金：国家自然科学基金项目(60971133;61271112)~~

摘　　要：目的海马体积很小,对比度极低,传统标记融合方法选用手工设计的特征模型,难以提取出适应性好、判别性强的特征。近年来,深度学习方法取得了极大成功,基于深度网络的方法已应用于医学图像分割中,但海马结构复杂,子区较多且体积差别较大,特别是CA2和CA3子区体积极小,常见的深度网络无法准确分割海马子区。为了解决这些问题,提出一种结合多尺度输入和串行处理神经网络的海马子区分割方法。方法针对海马中体积差距较大的子区,设计两种不同的网络,结合多种尺度图像块信息,为小子区建立类别数量均衡的训练集,避免网络被极端化训练,最后,采用串行标记的方式对海马子区进行分割。结果在Tail,SUB和PHG子区上的准确率达到了0.865,0.81,0.773,较现有的多图谱子区分割方法有较大提高,并且将体积较小子区CA2,CA3上的准确率分别提高了6%和9%。结论该算法将基于卷积神经网络的分类方法引入到标记融合阶段,根据海马子区特殊的灰度及结构特点,设计两种针对性网络,实验证明,该算法能提取出适应性好、判别性强的特征,提高了分割准确率。Objective Hippocampus is a structure in the brain in memory consolidation and can be divided into nine sub- fields. Hippoeampus atrophy has been mostly studied in various neurological diseases, such as Alzheimer＇ s disease and mild cognitive impairment. Accurate hippocampus subfield segmentation in magnetic resonance （MR） images plays a cruci- al role in the diagnosis, prevention, and treatment of neurological diseases. However, the segmentation is a challenging task due to small size, relatively low contrast, complex shape, and indistinct boundaries of hippocampus subfields. Numer- ous scholars have been engaged in hippoeampus subfield segmentation. Multi-atlas-based methods can obtain accurate seg- mentation results by fusing propagated labels of multiple atlases in a target image space. However, the performance of multi-atlas significantly relies on the effectiveness of the label fusion method. Deep learning algorithms have emerged as promising machine-learning tools in general imaging and eomputer vision domains, such as medical image segmentation. However, Cornu Ammonis （CA） 2 and CA3 are limited by MR resolution and are thus significantly smaller than other sub-fields in hippocampus MR images. Most deep learning algorithms with identical network models and uniform patches present poor segmentation accuracy regardless of the considerable differences in the volume of different subfields. Method This study proposes a combined multi-scale patch and cascaded convolutional neural network （CNN） -based classification algo- rithm for segmenting the hippocampus into nine subfields to address the aforementioned deficiencies. In comparison with traditional label fusion method, the proposed method does not rely on explicit features but learns to extract important fea- tures for classification. Two different CNNs are designed considering the significant volume differences among different sub- fields. Network 1, which considers large patches as inputs, is trained to segment large subfields accurately. Networ

关 键 词：海马子区分割 多尺度 卷积神经网络 串行分割 多图谱 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]