Japan project targets faster medical image processing with AI
Fujifilm will focus on developing artificial intelligence for imaging tests and ultrasound examinations, with Hokkaido University handling design and evaluation.
- A Japanese project is developing an AI system to process medical images over five times faster than conventional graphics processing units.
- The system will use advanced chips that are also more than 10 times more energy-efficient, aiming for real-time processing with lower power consumption.
- The collaboration includes the Leading-edge Semiconductor Technology Center (LSTC), Fujifilm Corporation, and Hokkaido University.
- Fujifilm will focus on applying the technology to imaging tests and ultrasound examinations at medical institutions.
Advanced chips could enable an artificial intelligence (AI) system planned in Japan to process medical images more than five times as fast as conventional graphics processing units.
The chips would also offer over 10 times the energy efficiency.
The collaboration was announced on Tuesday (October 20) by the Leading-edge Semiconductor Technology Center (LSTC) and other partners, including Fujifilm Corporation and Hokkaido University.
Their aim is to meet medical facilities’ need for real-time processing while reducing electricity consumption.
The AI’s design and evaluation will be handled by Hokkaido University, while Fujifilm will concentrate on developing it for imaging tests and ultrasound examinations at medical institutions.
Responsibility for the project’s semiconductor designs rests with LSTC.
Rapidus Corporation, an LSTC member aiming to manufacture the advanced chips in volume, is the expected source of semiconductors with a 2-nanometre circuit line width.
Tetsuro Higashi, chairman of LSTC and Rapidus, said at Tuesday’s press conference in Tokyo: “We want to make use of cutting-edge semiconductors to develop technologies that are helpful at medical sites.”
[Copyright The Jiji Press, Ltd.]
