Diagnose colon disease by feature selection based on artificial neural network and group teaching optimization algorithm

Abstract

There are various images available today for doctors to diagnose various diseases. The variety of medical imaging techniques increases the likelihood of early disease detection, allowing physicians to begin the treatment process more quickly. This article presents an intelligent approach to diagnosing colorectal cancer. The proposed method uses the group teaching optimization algorithm for feature selection to select the vital features of the image for plant disease. This feature uses to learn the multilayer artificial neural network to classify images into two categories, normal and malignant. Experiments show that the mean index of accuracy, sensitivity, and F1-Score in the proposed method on the color data set for the diagnosis of colon cancer is 92.72 %, 93.14 %, and 94.26 %, respectively. The proposed method's accuracy, precision, sensitivity, plant disease, and F1-Score in classifying Kvasir data set images are 96.42 %, 88.62 %, 92.69 %, and 89.54 %, respectively. Experiments show that the proposed method is more accurate in classifying plant disease images than 3Layer CNN, TFL, Random Forest, and CNN DropBlock.