Skin Tone Analysis: Creating a Model for Detecting Skin Tones Under Various Conditions and a Visual Interface for Interacting With it
Problem Statement
Image and video-based tools like PPG are relied upon in technologies that detect heart diseases, including atrial fibrillation. However, PPG is limited in that it has varying accuracy when used on individuals with darker skin tones. This can lead to disproportionately poor health and missed opportunities for treatment for people in that category.
Our Goal
We want to design a tool that can effectively analyze and categorize skin tones of users under the various lighting conditions experienced by users of VPG Medical's application. Ideally, it should be:
Accurate
Fast
Simple
Internal use of this tool at VPG would allow for engineers to more effectively handle the problems with PPG.