A Lightweight Skin Cancer Detection Model Based on Convolutional Neural Network

With the development of Artificial Intelligence (AI), deep neural networks have been used in a variety of fields such as digital signal processing, object detection as well as medical use. It acts as a vital aid to people, especially when it comes to huge workloads. Skin cancer has been considered the most common cancer with universality and detectability. We aim at constructing a lightweight detection model which can be adopted on FPGA or other SCMs to function as a more convenient family doctor with high accuracy of classification. In this paper, we solved the problem of data imbalance and proposed a new model based on the structure of the fire module and CNN to achieve the goal of categorizing skin cancers. Furthermore, we compare our model with prevalent image classifiers and the result indicates that our model has a better performance in both accuracy and capacity. Finally, the trained model is lightweight but efficient and we got a high accuracy of 97.8%.