Skin Cancer Classification Using a Convolutional Neural Network: An Exploration into Deep Learning
Naga Venkata Yaswanth Lankadasu, Devendra Babu Pesarlanka, Ajay Sharma, Shamneesh Sharma, Saikat Gochhait
Abstract
Skin cancer is one the most frequent type of cancer in the world. Early detection and diagnosis are vital for effective treatment. Deep learning has been determined to be efficacious in the categorization of skin cancer. In this paper the author has presented a deep learning approach for classifying skin cancer. The algorithm was trained on approximately 10000 photos of skin cancer. In the approach author has used convolutional neural network (CNN) for skin cancer classification. The CNN model is next trained on a collection of skin data tagged as benign or malignant. In the validation of our method is done using a publicly accessible database of skin images. To train dataset, our method obtains ~92% accuracy. For the test set, the model achieves an accuracy of more than 95%. The model can accurately categorize both benign and malignant skin cancer. The predicted model is a useful method for skin cancer early detection and treatment.