MIT EECS - Foxconn Undergraduate Research and Innovation Scholar
Machine Learning Techniques for Medical Monitoring on Portable Devices
Anantha P. Chandrakasan
Skin conditions require consistent monitoring over time in order to evaluate their progress and the efficacy of treatments, but in practice, obtaining accurate, frequent observations has been difficult. This project aims to create a portable multispectral imaging module and software application that patients and doctors can use to reliably track vitiligo. My work focuses on the software application, particularly the image processing algorithm, which will use computational photography and machine learning techniques to extract important features of skin lesions. We also hope to extend the algorithm to other conditions beyond vitiligo and provide diagnostic predictions. Ultimately, we seek to expand current imaging techniques, optimize them for portable hardware, and improve the medical monitoring process.
I worked with Prof. Glen Urban at Sloans Center for Digital Business studying the effectiveness of various types of media for advertising, developing the technology for delivering the experimental treatments and analyzing the results. I interned on Googles cloud computing team, building an app to manage and distribute software to virtual machines.