Retinal image quality classification using fine-tuned CNN

Jing Sun, Cheng Wan, Jun Cheng, Fengli Yu, Jiang Liu

Research output: Chapter in Book/Conference proceedingConference contributionpeer-review

9 Citations (Scopus)

Abstract

Retinal image quality classification makes a great difference in automated diabetic retinopathy screening systems. With the increase of application of portable fundus cameras, we can get a large number of retinal images, but there are quite a number of images in poor quality because of uneven illumination, occlusion and patients movements. Using the dataset with poor quality training networks for DR screening system will lead to the decrease of accuracy. In this paper, we first explore four CNN architectures (AlexNet, GoogLeNet, VGG-16, and ResNet-50) from ImageNet image classification task to our Retinal fundus images quality classification, then we pick top two networks out and jointly fine-tune the two networks. The total loss of the network we proposed is equal to the sum of the losses of all channels. We demonstrate the super performance of our proposed algorithm on a large retinal fundus image dataset and achieve an optimal accuracy of 97.12%, outperforming the current methods in this area.

Original languageEnglish
Title of host publicationFetal, Infant and Ophthalmic Medical Image Analysis - International Workshop, FIFI 2017 and 4th International Workshop, OMIA 2017 Held in Conjunction with MICCAI 2017, Proceedings
EditorsM. Jorge Cardoso, Tal Arbel
PublisherSpringer Verlag
Pages126-133
Number of pages8
ISBN (Print)9783319675602
DOIs
Publication statusPublished - 2017
Externally publishedYes
EventInternational Workshop on Fetal and Infant Image Analysis, FIFI 2017 and 4th International Workshop on Ophthalmic Medical Image Analysis, OMIA 2017 held in Conjunction with 20th International Conference on Medical Image Computing and Computer Assisted Intervention, MICCAI 2017 - Quebec City, Canada
Duration: 14 Sep 201714 Sep 2017

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume10554 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

ConferenceInternational Workshop on Fetal and Infant Image Analysis, FIFI 2017 and 4th International Workshop on Ophthalmic Medical Image Analysis, OMIA 2017 held in Conjunction with 20th International Conference on Medical Image Computing and Computer Assisted Intervention, MICCAI 2017
Country/TerritoryCanada
CityQuebec City
Period14/09/1714/09/17

Keywords

  • Convolutional neural networks (CNN)
  • Fine-tuning
  • No-reference image quality assessment (NR-IQA)
  • Retinal image

ASJC Scopus subject areas

  • Theoretical Computer Science
  • Computer Science (all)

Fingerprint

Dive into the research topics of 'Retinal image quality classification using fine-tuned CNN'. Together they form a unique fingerprint.

Cite this