An Expert System to Predict Eye Disorder Using Deep Convolutional Neural Network

Glaucoma according to the W.H.O is one of major causes blindness worldwide. Due its complexity and silent nature early detection this disease makes it hard detect. There have been several techniques over years for classification which shown significant improvement past decade or two. Some many models are SVM (support vector machine), KNN (K- Nearest Neighbors), Decision tree, Logistic Regression ANN (Artificial Neural Network) back propagation. For paper we would consider different procedure method glaucoma using MATLAB Deep Convolutional Network (DCNN). The DCNN based expert system basically works like human brain with input, neurons, hidden layers output. project Fundus image both healthy collected good lighting condition so that all features can be identify. then passed through processing such as Grayscale, B&W, Complement, Robert, Resize power Transform. fundus a texture feature extraction algorithm know gotten Contrast, Correlation, energy, Homogeneity, Entropy, Mean, Standard deviation, Variance, skewness Kurtosis. After data arrangement on spreadsheet serves means record. Lastly, deep convolutional neural network written layer, 16 input neuron 2 output either not. split into train test dataset 70% training 15% validation testing. Accuracy was 92.78% execution time 5.33s only depending number iteration epochs.