Skin cancer has possibly met its deadliest, unlikeliest foe in the form of IBM, or rather, IBM’s cognitive computers.
IBM researchers, working with New York’s Memorial Sloan Kettering Cancer Center, found that by using cognitive computing-based visual analytics, they were able to spot melanoma with an accuracy of about 95 per cent in 3,000 images.
It’s a huge thing, especially when the average for current methods range from 75 to 84 percent. Furthermore, they reckon that once the technology gets commercialised, they’d be able to do scans in less than a second.
Dr. Noel Codella, the Technical Coordinator of the collaboration, tells Computerworld that “the technology can pull on massive amounts of data to help the doctor make more informed decisions." This is one area that computers will always have an edge over us – the ability to analyse and compile huge archives of data at a rate that is beyond the reach of humans.
IBM’s approach involves multiple tests, which could include looking for unusual colour distributions or texture patterns on the skin, rapid progression of lesions or deviations from normal growth, compared to the rest of the body as well as other people on file who have genetic or geographical similarities.
The success of the research thus far was also down to the company’s advanced technologies such as the IBM Multimedia and Analytics System, a visually oriented machine learning architecture, which was used in conjunction with Medical Sieve, an IBM system for analyzing medical images, and a visual recognition and search system called Intelligent Video Analytics.
The team is still working on the project in a bid to refine their processes and measurements and it should be a while before this becomes mainstream practice.