Inspiration

In 2021, more than 54,000 adults in the US were diagnosed with oral cancer. Of those diagnosed, an estimated 10,000 people have died, killing roughly 1 person per hour. Everyday, there will be approximately 132 new individuals diagnosed with oral cancer. Worldwide the problem is much greater, with more than 640,000 new cases projected to occur each year. The death rate is particularly high not because it is hard to discover or diagnose, but due to the cancer being routinely discovered late in its development. Warning signs of pre-cancerous growths and early oral cancers include ulcers, sores, and red or white patches in the mouth which do not heal. They are usually painless at this stage, so most people either ignore them or self-medicate using gels. As a result, most patients are diagnosed at the later stages of oral cancer whereby it is more difficult to treat.

What it does

DentalAnywhere is an image recognition system that is able to identify oral conditions such as ulcers, inflammation, fluorosis, etc., thus providing a reference to users on their oral health.

How we built it

VGG-16 model, a pre-trained model will be deployed for transfer learning due to the limited training dataset.

What's next for DentalAnywhere

More datasets could be acquired to improve the accuracy.

Built With

Share this project:

Updates