Meeting Banner
Abstract #0265

Deep-Neural-Network based image diagnosis: comparing various image preprocessing strategies to achieve higher accuracy and understanding of the decision

Yasuhiko Tachibana1, Takayuki Obata1, Jeff Kershaw1, Yoko Ikoma1, Tokuhiko Omatsu1, Riwa Kishimoto1, and Tatsuya Higashi2

1Applied MRI Research, Department of Molecular Imaging and Theranostics, National Institute of Radiological Sciences, Chiba, Japan, 2Department of Molecular Imaging and Theranostics, National Institute of Radiological Sciences, Chiba, Japan

The purpose of this study was to investigate how image preprocessing might help overcome two problems for deep-neural-network (DNN) based image diagnosis: the need for a large training database to achieve high accuracy and the difficulty humans have in understanding the internal decision process. Five DNNs were trained with a brain image series (preprocessed in five different ways), to judge the age-range of a volunteer. The performance of the DNNs was then compared statistically. The results suggested that image preprocessing may facilitate higher accuracy, and also make it easier to understand how and why a judgement was made.

This abstract and the presentation materials are available to members only; a login is required.

Join Here