Bengin and Malignant Disease Detection Using Deep Learning Techniques

Skin cancer is becoming more common, which is a serious global public health concern. Increasing early detection is essential to bettering patient outcomes. The goal of this project is to automatically identify skin cancer by utilizing deep learning techniques. The main objective is to develop a reliable and accurate system that can distinguish between different kinds of skin lesions and help medical professionals make well-informed diagnostic decisions. The "dataimg.csv" dataset, which includes dermatological images of both benign and malignant skin lesions, is used by the suggested model. In order to tackle the present issue of inadequate precision in the identification of skin cancer, we used multiple algorithms on the 800-image "dataimg" dataset. Data gathering, preprocessing, data cleansing, visualization, and model development are the main phases of this project. The accuracy rates of the various models used were as follows: SVM obtained 82%, CNN achieved 85%, VGG-16 reached 90%, ResNet-50 achieved 88%, and DenseNet achieved 87%. Among the models tested, VGG-16 performed the best, achieving an impressive above 90% accuracy on the "dataimg" dataset.