Application of Deep Learning Techniques for Detection and Classification of Human Disease
M A Navyashree, P. Nagaraju · 2023
Machine learning is used practically everywhere from technologies to healthcare. Many researchers predict that disease diagnosis through machine learning is cost efficient and time efficient. In traditional methods, an expert is needed to detect disease and sometimes, it may cause human error. So, there is a need of automated diagnosis tool for detection of disease. In this paper, a methodology for detection of malaria and breast cancer is proposed. In malaria detection, data is preprocessed, trained and tested using CNN, Resnet50 and VGG16. Accuracy of CNN, VGG16 and Resnet50 are 94.73%, 90.83% and 92.96% respectively. CNN has highest accuracy when compared to VGG16 and Resnet50. In malaria classification, YOLOv5 was used for classification of malaria which classified malaria parasite species and its stages. In Breast cancer detection, data is preprocessed, trained and tested using CNN, ResNet50 and VGG16. Accuracy of CNN, VGG16 and Resnet50 are 95.13%, 94.04% and 95.53% respectively. Resnet50 has highest accuracy when compared to VGG16 and CNN.