Skin Cancer Detection and Classification using Convolutional Neural Network

Skin Cancer Detection and Classification using Convolutional Neural Network

Khushal Mehrotra¹, Pooja Sahoo²

Department of Electronics Engineering

Madhav Institute of Technology & Science, Gwalior-M.P

Abstract:

With the increasing number of cancer cases around the world, it is becoming increasingly important to monitor and identify them at an early stage. Unhealthy lifestyle choices, a lack of awareness, carelessness, and a fear of being checked up are all key contributors to the rising number of instances. Thus, development of a technical solution for it becomes extremely important, as it will be widely available and much more user friendly.

We present a Convolutional Neural Network (CNN) model that can distinguish between 9 forms of skin cancer if present or state Cancer not detected if not present. Combination of many convolutional layer and max pooling layers along with dense and dropout layers make up the model. We attempted to shape the model into a web app such that a user might implement the model by just providing a photo of the contaminated area captured from a dermatoscopic device.

KEYWORDS: CNN, Convolutional, max Pooling, dropout