Cataract Disease Detection Used Deep Convolution Neural Network

Cataract is one of the most frequent eyesight problems that causes vision distortion. Normally, the lens converges light to the retina in most cases and the presence of the cloud structure (cataract) causes obstructing light and not reach the lens that cause in poor visual acuity. Cataracts can develop without causing any symptoms. Cataracts are rarely painful, but they can cause visual loss in the central vision and even blindness. Therefore, Accurate and quick detection of cataracts is the greatest method to avoid painful and costly procedures and, depending on the severity of the condition, to avoid blindness. Artificial intelligence-based cataract detection methods have gained a lot of attention in the scientific community. In this paper, Deep Convolution Neural Network (DCNN) used for detection cataract automatically in fundus images. The loss and activation functions are tuned to train the network with small kernels. Also used Adam optimizer and (Kaggle, ODIR) datasets to train the model. The suggested method beats state-of-the-art cataract detection systems with an average accuracy of 99.91%, precision 99%, sensitivity 99% and f1-score 99% for Kaggle dataset, and have a average accuracy of 100%, precision 100%, sensitivity 100% for ODIR dataset, according to experimental results.