As many jobs are disappearing to automation, the latest profession to also start seeing the future may be dermatology. Stanford University researchers have developed a deep convolutional neural network, an artificial intelligence technique for building a knowledge set, to learn how to spot suspect cancer lesions.
Today this process is manual and prone to errors of subjectivity. Dermatologists simply look through a dermatoscope and judge based on their education and experience. The Stanford system was given 130,000 images of skin lesions simply labeled with previously established diagnoses that included more than 2,000 diseases. The system processed these images and learned on its own what to look for.
To test whether it was a good student, the system was pitted against 21 board-certified dermatologists that were asked to tell apart keratinocyte carcinomas, a common type of cancer, from benign seborrheic keratoses and malignant melanomas, the deadliest cancer, from benign nevi. And here’s the lowdown on the results according to the study abstract in Nature: “The CNN [convolutional neural network] achieves performance on par with all tested experts across both tasks, demonstrating an artificial intelligence capable of classifying skin cancer with a level of competence comparable to dermatologists.”
It seems it should be easy to now create a smartphone app that uses the system to do the very thing at the point of care and at low cost.
Study in Nature: Dermatologist-level classification of skin cancer with deep neural networks…
Via: Stanford…