Project combines mobile tech, AI to detect skin diseases

Japanese university, Kyocera initiative starts with research to use phone images to spot melanoma.


The maturation of artificial intelligence and mobile technology is bringing new ideas from vendors and efforts by healthcare providers to bring innovative treatments to patients.

For example, Japanese electronics manufacturer Kyocera is working with the University of Tsukuba in Japan to develop an artificial intelligence-based image recognition system that can remotely detect melanoma and other skin diseases by analyzing digital images of a patient’s skin using a smartphone.

Under the program, Kyocera is responsible for developing the AI recognition system, and the university will handle provision of a clinical image database and assessments of system accuracy and adaptability.

Also See: Smartphone selfies, app have potential to screen for jaundice

In Japan, the need for new treatments has become acute as the number of skin cancer patients doubled between 1999 through 2014.



While the Kyocera-Tsukuba alliance starts with skin cancer, the next phase of research will expand to image-based diagnostic support of any skin disease. Further, the program could branch out and support accurate diagnoses in rural and remote areas that don’t have a practicing clinician by using photos from smartphones or digital cameras.

Kyocera also envisions a cloud-based web service enabling users to create a drag-and-drop image classifier, which displays a set of images that need to be classified, which is based on deep-learning technology.

The program initially is supported with 20,000 clinical images that the university has collected during the past 20 years. The partners also plan to build a system that can identify more than 2,000 different skin diseases.

Kyocera and Tsukuba started joint research in March and expect to continue through March 2018, with the goal of having a commercialized product in 2020.

More for you

Loading data for hdm_tax_topic #better-outcomes...