A South Coast university student is using artificial intelligence to develop an app to help people determine whether they have a deadly skin cancer.
Currently, the only way to find out if you have melanoma is to get a biopsy. But, UC Santa Barbara undergrad Abhishek Bhattacharya is hoping to change that. He has been working with a UC San Francisco physician using computer science to create a neural network model that uses images gathered from the web.
“This model is supposed to replicate a visual cortex of a brain. And just the way we humans recognize objects, this model recognizes objects and images. We basically trained this model to be like a doctor and diagnose skin cancer,” he said.
This model has shown a 96% accuracy rating – better than humans. However, Bhattacharya cautions that it still needs to be proven within a clinical setting.
The researchers have launched a company called SkinIQ in which they’re developing a mobile app to help both doctors and patients detect melanoma.