Recent advances in medical imaging enabled a large quantity of high quality data to be collected. However, most of that data goes unseen because of the data explosion problem and the time constraints. At the same time the ubiquity of smartphones and improvement in the embedded sensors create amazing opportunities for medicine.
Existing routine skin cancer checks are not only time-consuming (for both patients and doctors), but also dependent on the doctor’s skills. At Skin Analytics we are trying to do the same things in a bit more intelligent way, crafting algorithms that allow for automatic analysis of your moles.
It has nothing to do with magic, but rather with a smart combination of image processing techniques known for years with new data-specific methods. Our goal is to decipher information hidden in every single pixel of your skin photos.
The first step is to understand the problem itself i.e. how the doctors quantify the symptoms. The answer is: they have it in their eyes. And if doctors can spot the symptoms just by looking at your mole, why can’t we ask a machine to do the same from images? Clinicians, instead, could spend their precious time on cases in which technology cannot give reliable answer.
In order to implement an algorithm that analyses pictures just like human brain does, we need to teach it the principles similar to those underlying our visual system. Perhaps the most exciting area of computer vision is mimicking the way people perceive the surrounding world, and this is what we do at Skin Analytics!
Technology cannot replace doctors, but it can help them to help you.