Edge Computing for Having an Edge on Cancer Treatment: A Mobile App for Breast Image Analysis
Eleftherios Charteros, Iordanis Koutsopoulos · 2020
Edge computing has seen tremendous advances in recent years. This progress made it possible to develop mobile applications with greater computational needs that will be able to provide useful insights to users concerning both their habits and their health, by performing computations locally on the mobile device and thus keeping all the data and protecting their privacy. The purpose of this work is to provide a proof-of-concept of an image recognition and analysis app for patients that have undergone breast cancer. The patient takes a snapshot of her breast, and the app runs a Convolutional Neural Network (CNN) model and outputs a classification of breast cosmetic status. We show that it is possible to implement computationally heavy machine learning models in edge devices and to provide real-time status monitoring to users through image analysis done on images taken from the camera of their smartphone, without the photo leaving the device. The app module may enhance a larger-scale system that uses patient-sourced image data to test the effects of surgery or radiotherapy treatment on patients.