Melanoma Detection using CNN

Melanoma Detection using CNN

Shantha Kumar H C¹ Hritwik Kumar² Eshi Nagendra³ Harshitha S⁴  Deepak Kumar⁵

Department of Computer Science and Engineering

SJB Institution of Technology

Abstract – Skin is the largest organ of the body. It is exposed to a lot of harmful environmental threats. This leads to so many diseases one among then is the lethal skin cancer. Skin cancer is the abnormal growth of skin cells which turns into black coloured lesions. Melanoma is the most serious type of skin cancer which produces the melanin, this results into black lesions. These lesions can be detected by our naked eyes but for more medical efficiency the doctors use a device called dermoscope. Dermoscope is a non-invasive, in vivo technique primarily used for examination of lesions. In this project, we have developed a model which identifies melanoma automatically using the Convolutional Neural Network (CNN). This CNN model will be developed in Python using Keras and Tensorflow in the backend. The model is developed and tested with different network architectures by varying the type of layers used to train the network including Convolutional layers, Pooling layers. The model will be tested and trained on the dataset collected from the International Skin Imaging Collaboration (ISIC) challenge archives. The proposed system gives 89%  accuracy for multiclass classification model using a custom convolutional neural network model in tesnerflow.

Keywords – Deep learning, CNN, Rel-U.