Author : Jagrithi Singh 1
Date of Publication :2nd July 2020
Abstract: The retina is the sensory membrane that lines the inner surface of the back of an eyeball. There is a wide variety of problems, conditions, and diseases like Macular degeneration, Diabetic retinopathy, etc that affects the retina and thus affect the human vision. Chances of retinal damage increase as a person get aged or have retinopathic diabetes. Retinal damage can be cured if detected in the early state but if not cured on time it may lead to blurry vision and also can eventually lead to blindness. The probability of retinal damage is increasing as the number of diabetic patients is increasing therefore there is a need for an effective and efficient method with high accuracy to detect retinal damage. Chances of retinal Damage through classification of optical coherence tomography (OCT) images can be achieved with high accuracy using classical convolution neural networks (CNN), a commonly used deep learning network for computer-aided diagnosis. In this study, we attempt to improve classification accuracy by using CNN based custom model. From an OCT dataset, we produced a training dataset of 83,484 images and a test and validation dataset of 1000 images. The dataset was further classified into images of choroidal neovascularization (CNV), diabetic macular edema (DME), drusen, and normal. We have built a custom model using Convolution Neural Network(CNN), Relu activation function, MaxPooling, and Dropout Layers. Classification of OCT images using our method achieved a high accuracy of 99.69% while using fewer numbers of parameters when compared with other models in the literature. We further compared the custom model with already existing models by training them in a similar manner.
- Kermany, Daniel; Zhang, Kang; Goldbaum, Michael (2018), “Labeled Optical Coherence Tomography (OCT) and Chest X-Ray Images for Classification”, Mendeley Data, v2 http://dx.doi.org/10.17632/rscbjbr9sj.2
- Ahn JM, Kim S, Ahn K-S, Cho S-H, Kim US. Accuracy of machine learning for differentiation between optic neuropathies and pseudopapilledema. BMC Ophthalmol. 2019;19:178. https://doi.org/10.1186/s12886-019-1184-0.
- Kermany DS, Goldbaum M, Cai W, Lewis MA. Identifying Medical Diagnoses and Treatable Diseases by Image-Based Deep Learning Resource Identifying Medical Diagnoses and Treatable Diseases by ImageBased Deep Learning. Cell. 2018;172:1122–31.e9. https://doi.org/10.1016/j.cell.2018.02.010.
- Swanson EA, Fujimoto JG. The ecosystem that powered the translation of OCT from fundamental research to clinical and commercial impact [invited]. Biomed Opt Express. 2017;8:1638. https://doi.org/10.1364/BOE.8.001638.
- Kornblau IS, El-Annan JF. Adverse reactions to fluorescein angiography: a comprehensive review of the literature. Surv Ophthalmol. 2019;64:679–6793. https://doi.org/10.1016/j.survophthal.2019.02.004.
- Musa F, Muen WJ, Hancock R, Clark D. Adverse effects of fluorescein angiography in hypertensive and elderly patients. Acta Ophthalmol Scand. 2006;84:740–2.
- Lim LS, Mitchell P, Seddon JM, Holz FG, Wong TY. Age-related macular degeneration. Lancet. 2012;379:1728–38. https://doi.org/10.1016/S0140- 6736(12)60282-7.
- Tsuji, T., Hirose, Y., Fujimori, K. et al. Classification of optical coherence tomography images using a capsule network. BMC Ophthalmol 20, 114 (2020). https://doi.org/10.1186/s12886-020-01382-4