Yang Chengliang, Rangarajan Anand, Ranka Sanjay
Dept. of Computer & Information Science & Engineering, University of Florida, Gainesville, FL 32611, USA,
AMIA Annu Symp Proc. 2018 Dec 5;2018:1571-1580. eCollection 2018.
We develop three efficient approaches for generating visual explanations from 3D convolutional neural networks (3D-CNNs) for Alzheimer's disease classification. One approach conducts sensitivity analysis on hierarchical 3D image segmentation, and the other two visualize network activations on a spatial map. Visual checks and a quantitative localization benchmark indicate that all approaches identify important brain parts for Alzheimer's disease diagnosis. Comparative analysis show that the sensitivity analysis based approach has difficulty handling loosely distributed cerebral cortex, and approaches based on visualization of activations are constrained by the resolution of the convo-lutional layer. The complementarity of these methods improves the understanding of 3D-CNNs in Alzheimer's disease classification from different perspectives.
我们开发了三种有效的方法，用于从用于阿尔茨海默病分类的三维卷积神经网络（3D-CNN）生成视觉解释。一种方法是对分层三维图像分割进行敏感性分析，另外两种方法是在空间地图上可视化网络激活。视觉检查和定量定位基准表明，所有方法都能识别出用于阿尔茨海默病诊断的重要脑区。比较分析表明，基于敏感性分析的方法在处理分布松散的大脑皮层时存在困难，而基于激活可视化的方法受到卷积层分辨率的限制。这些方法的互补性从不同角度提高了对3D-CNN在阿尔茨海默病分类中的理解。
