It works by learning from images that it "sees" and continues to teach itself with each case to improve its performance.
To be fair, the performance of the human dermatologists improved when they were provided more information about the patients - including age, gender, and position of the lesion - plus close-up images of skin lesions of the cases.
"Most dermatologists have done less well", write the researchers in the Annals of Oncology journal where the results of the study have been issued.
Melanoma incidents are constantly increasing, with about 232,000 new diagnoses and 55,500 deaths each year worldwide. This is why this CNN is so impressive - it would be able to identify more cancers early on, thereby saving lives. Another set of 100 images was comprised of some of the most hard to diagnose lesions and was used to test both machine and real dermatologists. Once trained it was tested against 58 dermatologists from across 17 countries, who were shown benign moles and malignant melanomas. Doctors were in an "artificial setting", the test set did not include the full range of skin lesions; and there were fewer images from non-Caucasian skin types and genetic backgrounds to examine. "However, the CNN, which was still working exclusively from the dermoscopic images with no additional clinical information, continued to out-perform the physicians' diagnostic abilities".
Although machines doctors are still a thing of science-fiction, a team of medical researchers from Germany proved that artificial intelligence could outperform human diagnosticians.
The researchers said AI technology is a useful and easy way to detect skin Melanomas (cancer cells) from normal cells.
Australian experts Victoria Mar, from Melbourne's Monash University, and Peter Soyer from the University of Queensland said it was a major breakthrough in detecting skin cancers. However, before such an artificial intelligence system finds broad clinical application, some technical problems, such as the difficulty of correctly depicting certain melanomas in areas such as the fingers and toes or the skull, system to "read" the images. "Still, there is much more work to be done to implement this exciting technology safely into routine clinical care".