Kyocera to develop AI-based diagnostic tool for skin diseases


Kyocera Communication Systems (KCCS) has commenced joint research with the University of Tsukuba in Japan to develop an artificial intelligence (AI) image recognition system for the detection of skin diseases such as melanoma.

With plans for commercialisation in March 2020, the AI technology will work by analysing digital images of a patient’s skin.

The use of AI and digital images for diagnosis is expected to be more beneficial than standard procedures that are known to be based on a physician’s knowledge and experience.

Led by University of Tsukuba Faculty of Medicine Department of Dermatology professors Manabu Fujimoto and Yasuhiro Fujisawa, the development of the image-recognition system will target accurate differentiation of various types of skin malignancies.

Together, the organisations aim to deliver a system that will possess a capability to detect more than 2,000 different skin diseases.

"The organisations aim to deliver a system that will possess a capability to detect more than 2,000 different skin diseases."

While KCCS will develop the AI system, the University will contribute by evaluating the system’s accuracy and adaptability in real-world conditions.

KCCS is leveraging the University of Tsukuba Hospital’s Department of Dermatology database of more than 200 clinical images gathered over a period of 20 years.

The firm will make the system available through a cloud-based web service called Labellio to enable the creation of a simple deep learning-powered image classifier.

Apart from dermatology specialists, the system is expected to facilitate accurate diagnoses in rural and remote areas where a local clinician is not available.


Image: Potential applications of the skin disease diagnostic support system. Photo: courtesy of Kyocera Corporation.