32iQ - A low-cost cavity detection application
2023, May 07
32IQ is a product of my Capstone project for UC Berkeley M.S. Data Science program. The project was completed in collaboration with Jonathan Whitely, Connor McCormick, and Alexa Coughlin.
The project involved creating a computer vision model to detect cavities, and an easy-to-use web interface. Dentists simply upload a dental X-ray and the model returns whether a cavity is detected, along with confidence intervals. The model achieved near SOTA accuracy (~95%), and could be made available to dentists for a fraction of the cost of current market competitors.
Click here for a link to a recorded demo. The website is still available although the analyze API is no longer available.
Below is the final presentation summarizing our work.