This is a repository for my computer science final year project
Here is a link to the deployed web application: https://skin-lesion-analysis-app.herokuapp.com/
Two datasets that were used are:
- HAM10000 (link: https://dataverse.harvard.edu/dataset.xhtml?persistentId=doi:10.7910/DVN/DBW86T)
- 7PCE (link: https://derm.cs.sfu.ca)
Due to lack of sufficient access to dermatologists and low public awareness of skin cancer there is high-demand, especially in less developed countries, to develop computer-aided diagnostic systems that could help with early detection of skin cancers, particularly the some of the deadliest ones such as melanoma.
This project presents a system that allows to analyse and identify the type of skin lesion and the probability that it is cancerous using machine learning, specifically deep learning. The main objective of this project was to achieve successful integration of a deep learning model on a web application. This was accomplished by training and testing several deep learning algorithms to detect one of the 7 common types of skin lesions using a dataset of microscopic images. Then choosing the model that performs best and creating a web application that provides a user-friendly interface to interact with the system. The web application uses Django for the backend, REST API to communicate and React for the frontend. Main functionality of the "MySkin" web application is the ability to upload a photo of a skin lesion and have it analysed by the algorithm.