- Joined
- Oct 6, 2014
- Messages
- 1,424 (0.41/day)
System Name | octo1 |
---|---|
Processor | dual Xeon 2687W ES |
Motherboard | Supermicro |
Cooling | dual Noctua NH-D14 |
Memory | generic ECC reg |
Video Card(s) | 2 HD7950 |
Storage | generic |
Case | Rosewill Thor |
OK, the app isn't perfect, but it's a start and for 5 bucks, might be worth a try if you're at risk and don't see a doctor regularly.
It correctly identified 73 of 100 melanomas it analyzed and it's accuracy should continue to improve. Plus it also analyzes the UV index in your area and can refer you to local physicians.
It correctly identified 73 of 100 melanomas it analyzed and it's accuracy should continue to improve. Plus it also analyzes the UV index in your area and can refer you to local physicians.
SkinVision is based on a technology called "fractal geometry" — a branch of mathematics that can be used to simulate the natural growth of tissue. The SkinVision team used it to develop an algorithm that analyzes tissue growth using a picture of a mole that is uploaded to your smartphone. The algorithm can detect symptoms that are generally associated with risky moles.
"The skin is made up by cells that have replicated over and over again by following a rule," SkinVision cofounder Mircea Popa told Business Insider UK. "That leads to patterns that are measurable in a mathematical way." Malignant cells are diverted cells and they replicate in a different way, so the patterns get mixed up when you look at both the skin and the mole.
Shape and colour are each dependent on these patterns, and those are the characteristics the app targets.
.
.
.
In June, a panel of lecturers from the University of Munich in Germany praised the app, calling it "a promising tool in the pre-evaluation of moles." They found that out of all melanoma cases the app had been used on, 73 out of 100 were correctly detected.