文献类型：期刊文献

英文题名：Automated diabetic retinopathy grading and lesion detection based on the modified R-FCN object-detection algorithm

作者：Wang, Jialiang[1];Luo, Jianxu[1];Liu, Bin[2];Feng, Rui[3];Lu, Lina[4];Zou, Haidong[4]

机构：[1]East China Univ Sci & Technol, Coll Informat Sci & Technol, Shanghai, Peoples R China;[2]Shanghai Radio Equipment Res Inst, Shanghai, Peoples R China;[3]Fudan Univ, Coll Comp Sci & Technol, Shanghai, Peoples R China;[4]Shanghai Eye Hosp, Shanghai Eye Dis Prevent & Treatment Ctr, Shanghai, Peoples R China

年份：2020

卷号：14

期号：1

起止页码：1

外文期刊名：IET COMPUTER VISION

收录：;EI(收录号：20200708150452);WOS:【SCI-EXPANDED(收录号:WOS:000526681200001)】；

基金：This work was supported by the Scientific and Technological Innovation Action Plan of the Science and Technology Commission of Shanghai Municipality under grant number 17511107900.

语种：英文

外文关键词：feature extraction; biomedical optical imaging; image segmentation; medical image processing; object detection; eye; diseases; image classification; blood vessels; feature pyramid network; modified region proposal network; DR-grading model; Messidor data; Shanghai Eye Hospital; hospital data; modified R-FCN lesion-detection model; lesion detection; modified R-FCN object-detection algorithm; computer-aided retinal image screening system; retinal fundus images; modified object-detection method; region-based fully convolutional network

摘要：In this work, we develop a computer-aided retinal image screening system that can perform automated diabetic retinopathy (DR) grading and DR lesion detection in retinal fundus images. We propose a modified object-detection method for this task via a region-based fully convolutional network (R-FCN). A feature pyramid network and a modified region proposal network are applied to enhance the detection of small objects. The DR-grading model based on the modified R-FCN is evaluated on the Messidor data set and images provided by the Shanghai Eye Hospital. High sensitivity of 99.39% and specificity of 99.93% are obtained on the hospital data. Moreover, high sensitivity of 92.59% and specificity of 96.20% are obtained on the Messidor data set. The modified R-FCN lesion-detection model is validated on the hospital data set and achieves a 92.15% mean average precision. The proposed R-FCN can efficiently accomplish DR grading and lesion detection with high accuracy.