Computer-aided diagnosis of retinopathy is a research hotspot in the field of medical image classification. Optical coherence tomography imaging (OCT) is widely applied in the diagnosis of ocular diseases including DME and AMD. Most of the earlier research work were heavily dependent on ophthalmologists and exploited their knowledge to identify features, while some of the studies made use of very structurally complicated models, and others used small datasets of 45 images only. In this study we present the solution to classify the OCT images into 4 classes CNV, DME, DRUSEN, NORMAL using baseline 3,5 and 7 layer Convolutional Neural Networks. The novelty of this study is that it does not use any pre-trained models and yet achieves desired results just by hyper-parameter tuning and some clever observations. This paper explores the journey of our research of improvement along with substantial reasons which lay the fundamentals of convolutional neural networks. The biggest risk of overfitting in deep learning where multiple layered models are trained and tested and approaches of diminishing the overfitting effect has been discussed in detail.
Computer-aided diagnosis of retinopathy is a research hotspot in the field of medical image classification. Optical coherence tomography imaging (OCT) is widely applied in the diagnosis of ocular diseases including DME and AMD. Most of the earlier research work were heavily dependent on ophthalmologists and exploited their knowledge to identify features, while some of the studies made use of very structurally complicated models, and others used small datasets of 45 images only. In this study we present the solution to classify the OCT images into 4 classes CNV, DME, DRUSEN, NORMAL using baseline 3,5 and 7 layer Convolutional Neural Networks. The novelty of this study is that it does not use any pre-trained models and yet achieves desired results just by hyper-parameter tuning and some clever observations. This paper explores the journey of our research of improvement along with substantial reasons which lay the fundamentals of convolutional neural networks. The biggest risk of overfitting in deep learning where multiple layered models are trained and tested and approaches of diminishing the overfitting effect has been discussed in detail.