Skin Cancer Classification Using Image Processing and Convolutional Neural Network

In today's world, cancer is one of the most dangerous diseases, and Skin cancer is the most common among them. As pollution and global warming increase, it also increases the chance of Skin cancer occurring in people. As we know, biopsy, the skin cancer detection method using tissue, is very costly in terms of time and money, and timely skin cancer identification is necessary for proper treatment. Therefore, this paper proposes a machine-learning model to classify skin cancer based on image processing and CNN. The proposed model can classify seven different types of skin cancer. The combination of image processing and a custom CNN is 8-10% faster than just a pre-trained CNN model on a CPU-based machine. Our model accepts images as the input type and can accurately classify the mentioned diseases. Results show an accuracy of 96% in the testing phase, significantly reducing the training time. The suggested model can be applied as a biopsy clinical system to support decisions and help the Doctor examine a patient quickly for a better understanding of the problem.