Presentation information

General Session

General Session » [GS] J-2 Machine learning

[2P5-J-2] Machine learning: medicine and heathcare

Wed. Jun 5, 2019 5:20 PM - 7:00 PM Room P (Front-left room of 1F Exhibition hall)

Chair:Jun Ozawa Reviewer:Yoshikuni Sato

6:00 PM - 6:20 PM

[2P5-J-2-03] Medical image recognition of MRI brain images using deep logistic GMDH-type neural network and convolutional neural network

〇Tadashi Kondo1, Shoichiro Takao1, Sayaka Kondo2, Junji Ueno1 (1. Tokushima University, 2. Tokushima medical informatics laboratory)

Keywords:Deep neural networks

In this study, hybrid deep neural network is organized using the deep logistic Group Method of Data Handling (GMDH)-type neural network and the Convolutional Neural Network (CNN) and, is applied to the medical image recognition problem. The deep GMDH-type neural network algorithms have abilities of self-selecting the number of hidden layers, the optimum neuron architectures and useful input variables, and they can automatically organize the deep neural network architectures to minimize prediction error criterion defined as Akaike’s Information Criterion (AIC) or Prediction Sum of Squares (PSS) . This deep neural network algorithm is applied to medical image recognitions of brain regions, and the organs such as brain, the white matter and the lateral ventricle, are recognized and these regions are extracted accurately using the deep logistic GMDH-type neural networks.