Detection of Diabetic Retinopathy Using Bichannel Convolutional Neural Network

Shu-I Pao; Hong-Zin Lin; Ke-Hung Chien; Ming-Cheng Tai; Jiann-Torng Chen; Gen-Min Lin

doi:10.1155/2020/9139713

Detection of Diabetic Retinopathy Using Bichannel Convolutional Neural Network

J Ophthalmol. 2020 Jun 19:2020:9139713. doi: 10.1155/2020/9139713. eCollection 2020.

Authors

Shu-I Pao¹, Hong-Zin Lin^{2

3}, Ke-Hung Chien^{1

4}, Ming-Cheng Tai^{1

4}, Jiann-Torng Chen¹, Gen-Min Lin^{4

5

6}

Affiliations

¹ Department of Ophthalmology, Tri-Service General Hospital and National Defense Medical Center, Taipei 114, Taiwan.
² Department of Ophthalmology, Buddhist Tzu Chi General Hospital, Hualien 970, Taiwan.
³ Institute of Medical Sciences, Tzu Chi University, Hualien 970, Taiwan.
⁴ Department of Medicine, Hualien Armed Forces General Hospital, Hualien 971, Taiwan.
⁵ Department of Medicine, Tri-Service General Hospital and National Defense Medical Center, Taipei 114, Taiwan.
⁶ Department of Preventive Medicine, Northwestern University, Chicago, IL 60611, USA.

Abstract

Deep learning of fundus photograph has emerged as a practical and cost-effective technique for automatic screening and diagnosis of severer diabetic retinopathy (DR). The entropy image of luminance of fundus photograph has been demonstrated to increase the detection performance for referable DR using a convolutional neural network- (CNN-) based system. In this paper, the entropy image computed by using the green component of fundus photograph is proposed. In addition, image enhancement by unsharp masking (UM) is utilized for preprocessing before calculating the entropy images. The bichannel CNN incorporating the features of both the entropy images of the gray level and the green component preprocessed by UM is also proposed to improve the detection performance of referable DR by deep learning.