What if you could turn your smartphone into a pocket dermatologist? An algorithm has been developed that could soon do exactly that.
At Stanford University's Artificial Intelligence Laboratory, an algorithm has been developed that can detect skin cancers as accurately as a dermatologist: the algorithm is trained to visually recognise a malignant looking spot by running it through a database of some 130,000 images of skin lesions representing over 2000 types of skin diseases.
In an article in the science journal Nature, the algorithm was shown to match the performance of 21 board-certified dermatologists in diagnosing skin cancer. It utilises visual processing combined with deep learning, a type of artificial intelligence modeled after neural networks. Andre Esteva, co-author of the paper, explains:
“We made a very powerful machine learning algorithm that learns from data. Instead of writing into computer code exactly what to look for, you let the algorithm figure it out.”
Early Detection of Skin Cancer Saves Lives
Some three million new cases of skin cancer are detected each year globally, that is one in three of all yearly cancer diagnoses. And in some remote locations, lack of medical infrastructure, or a lack of affordable medical care means that for many, getting a diagnosis (let alone a cure) may not even be an option.
The team behind the Skin Cancer Detection Algorithm hope that the development will facilitate early detection and treatment of skin cancer, whose survival rates are around 97% when detected early, versus only 14% for late-stage detection.
At the moment the team are busy making their algorithm - which currently only exists on a computer - work through a smartphone app that everyone can use, thereby helping to bring down the number of death from skin cancer worldwide. The app will be able to facilitate diagnosis within resource-poor communities lacking medical infrastructure, as well as among people in more developed settings who, for one reason or another might be putting off their trip to the doctor.
The Skin Cancer Detection Algorithm project was one of the finalists of the INDEX Award 2017.
Here's their video for the award: