A01-KB105 Extraction and Visualization of Cellular Morphological Features and Their Pathological Changes Using Deep Neural Network

From Multidisciplinary Computional Anatomy
Jump to: navigation, search

Member

  • Primary Investigator
    Yoichi Miyawaki (The University of Electro-Communications, Graduate School of Informatics and Engineering, Professor)
  • Co-Investigator
    Kazuto Masamoto (The University of Electro-Communications, Graduate School of Informatics and Engineering, Professor)


Overview

Glial cells play important roles in supporting brain functions with respect to metabolism. Glial cells are known to change their morphology by pathological conditions. However, it remains unclear what morphological features change specifically. In this study, we apply deep convolutional neural network, which shows accurate performance in object recognition, to glial cell images measured by a microscopy, and analyze their morphological changes along progress of pathological states. Using this approach, we try to elucidate relationship between morphological features and pathological states, and will further apply it to other cellular and tissue images to help clinical diagnosis.

Project Design

A01-KB105-E.png