Medical Image Classifications: Deep Learning Prospective

Machine learning and deep learning technologies are reshaping the global medical industry at a breakneck pace. Image classification is one of its rapidly expanding fields. It is incorporated into nearly all technologies aimed at achieving intelligent smart health systems. The current paper implements and applies two image classification models based on convolutional neural network (CNN) versions to various image classification datasets. The current work makes use of the significant lungs X-ray images from COVID-19 medical datasets. It analyses the models' accuracy by adjusting their parameters such as layer count and activation function in order to identify the ideal parameters for CNN that provide the highest accuracy while classifying images. It evaluated the models' performance on the desired dataset and calculated the F-score, specificity and sensitivity matrices to validate the suggested models, as well as analysing their behaviour as a function of the image type. It achieves an accuracy of 90% for lungs X-rays in the COVID-19 dataset.