The system employs deep learning, a programming based on complex algorithms capable of analyzing images and relating certain patterns that are repeated in the development of malignant cells that lead to melanoma.
Japanese technology company Kyocera and the University of Tsukuba, Japan, developed a system to detect skin cancer in its initial phase, from the analysis of photographs of patients using Artificial Intelligence (AI).
The method, still in the experimental phase, reached a precision in its diagnoses of 90% in 4 thousand analyzed snapshots of patients’ skin, according to the Japanese newspaper “Nikkei”.
The system designed by Kyocera employs deep learning, a programming based on complex algorithms capable of analyzing images and relating certain patterns that are repeated in the development of malignant cells that lead to melanoma.
Researchers at the University of Tsukuba plan to create more variants of the matrix program applicable to the diagnosis of other diseases, including contagious diseases.
Kyocera and the university are expected to start distributing a prototype of the system by next year, at hospitals in the country, with the idea of marketing it in 2019, says the Nikkei.
The medical applications of AI could help cope with the increased demand for services in Japan as a result of the accelerated aging of the population, a trend that has led to a shortage of health professionals.
Other Japanese technology companies, such as Hitachi, NEC, and Toshiba, are also developing medical diagnostic systems based on imaging and artificial intelligence.
The government intends to promote these applications of AI, whose potential was highlighted in a report by the Ministry of Health in June and is expected to offer economic benefits to companies that invest in the sector.