Automatic differentiation of Glaucoma visual field from non-glaucoma visual filed using deep convolutional neural network

Abstract

BackgroundTo develop a deep neural network able to differentiate glaucoma from non-glaucoma visual fields based on visual filed (VF) test results, we collected VF tests from 3 different ophthalmic centers in mainland China.MethodsVisual fields obtained by both Humphrey 30–2 and 24–2 tests were collected. Reliability criteria were established as fixation losses less than 2/13, false positive and false negative rates of less than 15%.ResultsWe split a total of 4012 PD images from 1352 patients into two sets, 3712 for training and another 300 for validation. There is no significant difference between left to right ratio (P = 0.6211), while age (P = 0.0022), VFI (P = 0.0001), MD (P = 0.0039) and PSD (P = 0.0001) exhibited obvious statistical differences. On the validation set of 300 VFs, CNN achieves the accuracy of 0.876, while the specificity and sensitivity are 0.826 and 0.932, respectively. For ophthalmologists, the average accuracies are 0.607, 0.585 and 0.626 for resident ophthalmologists, attending ophthalmologists and glaucoma experts, respectively. AGIS and GSS2 achieved accuracy of 0.459 and 0.523 respectively. Three traditional machine learning algorithms, namely support vector machine (SVM), random forest (RF), and k-nearest neighbor (k-NN) were also implemented and evaluated in the experiments, which achieved accuracy of 0.670, 0.644, and 0.591 respectively.ConclusionsOur algorithm based on CNN has achieved higher accuracy compared to human ophthalmologists and traditional rules (AGIS and GSS2) in differentiation of glaucoma and non-glaucoma VFs.Electronic supplementary materialThe online version of this article (10.1186/s12880-018-0273-5) contains supplementary material, which is available to authorized users.

Deep learning in ophthalmology: The technical and clinical considerations

“…79 Using an entirely different strategy, Li et al trained a CNN to learn the Pattern Deviation probability plots of normal and glaucomatous eyes and was able to detect glaucoma with 93.2% sensitivity and 82.6 sensitivity. 80 Yousefi et al used an alternative Gaussian mixture and expectation maximization method to decompose VFs along different axes to detect VF progression. 23 This approach was as good or superior to current algorithms, including Glaucoma Progression Analysis, Visual field Index and Mean Deviation slope, in detecting VF progression.…”

Development and clinical deployment of a smartphone-based visual field deep learning system for glaucoma detection

“…To date, many groups have applied machine learning and deep learning methods in VF interpretation to detect glaucoma and predict glaucoma progression [7][8][9][10][11][12] . Most algorithms, however, are trained using single glaucoma parameter such as MD or PDP.…”

Automatic Detection of Diabetic Eye Disease Through Deep Learning Using Fundus Images: A Survey

“…Orlando et al [41] [89], with AUC of 95.7% and on Inception-ResNet with the same classifier yielded AUC of 86% on RIM-ONE-r3 [89]. Fei et al [43] used the VGG network to classify glaucoma and non-glaucoma visual fields based on the results of the visual field (VF) study and, for this test, they obtained VF samples from three different ophthalmic centres in mainland China. They obtained an Acc of 87.6%, while the specificity was 82.6% and sensitivity was 93.2%, respectively.…”