Automatic diagnosis of imbalanced ophthalmic images using a cost-sensitive deep convolutional neural network

Jiewei Jiang,Xiaoyan Li,Dongni Wang,Zhenzhen Liu,Wangting Li,Lin Liu,Zhuoling Lin,Shuai Wang,Jiangtao Cui,Xiyang Liu,Jing Li,Xiaohang Wu,Kai Zhang,Ming Zhu ,Liming Wang,Erping Long,Qian Cao ,Jingjing Chen,Jinghui Wang

doi:10.1186/s12938-017-0420-1

Jiewei Jiang, Xiaoyan Li + Show 17 more

Open Access

https://doi.org/10.1186/s12938-017-0420-1

Copy DOI

Abstract

BackgroundOcular images play an essential role in ophthalmological diagnoses. Having an imbalanced dataset is an inevitable issue in automated ocular diseases diagnosis; the scarcity of positive samples always tends to result in the misdiagnosis of severe patients during the classification task. Exploring an effective computer-aided diagnostic method to deal with imbalanced ophthalmological dataset is crucial.MethodsIn this paper, we develop an effective cost-sensitive deep residual convolutional neural network (CS-ResCNN) classifier to diagnose ophthalmic diseases using retro-illumination images. First, the regions of interest (crystalline lens) are automatically identified via twice-applied Canny detection and Hough transformation. Then, the localized zones are fed into the CS-ResCNN to extract high-level features for subsequent use in automatic diagnosis. Second, the impacts of cost factors on the CS-ResCNN are further analyzed using a grid-search procedure to verify that our proposed system is robust and efficient.ResultsQualitative analyses and quantitative experimental results demonstrate that our proposed method outperforms other conventional approaches and offers exceptional mean accuracy (92.24%), specificity (93.19%), sensitivity (89.66%) and AUC (97.11%) results. Moreover, the sensitivity of the CS-ResCNN is enhanced by over 13.6% compared to the native CNN method.ConclusionOur study provides a practical strategy for addressing imbalanced ophthalmological datasets and has the potential to be applied to other medical images. The developed and deployed CS-ResCNN could serve as computer-aided diagnosis software for ophthalmologists in clinical application.

Highlights

Ocular images play an essential role in ophthalmological diagnoses
Imbalanced datasets are inevitable in a variety of medical data analysis situations [6, 8,9,10,11], which causes the existing classifiers to exhibit a high false negative rate (FNR) or false positive rate (FPR)
Conclusions and future work In this paper, we proposed a feasible and automatic approach based on our CS-residual CNN network (ResCNN) model to effectively address the problem of misclassifications resulting from imbalanced ophthalmic images datasets

Summary

Introduction

Having an imbalanced dataset is an inevitable issue in automated ocular diseases diagnosis; the scarcity of positive samples always tends to result in the misdiagnosis of severe patients during the classification task. Ophthalmic imaging technologies play an important role in diagnosing eye diseases [2,3,4]. Imbalanced datasets are inevitable in a variety of medical data analysis situations [6, 8,9,10,11], which causes the existing classifiers to exhibit a high false negative rate (FNR) or false positive rate (FPR). It is imperative to explore a feasible and efficient strategy to address the problem of imbalanced ophthalmic image datasets to achieve higher-performance of computeraided diagnostic systems

Methods

Results

Conclusion