Tokyo, December 16, 2021 - NEC Corporation (NEC; TSE: 6701) announced today the development of an AI technology with Ono Pharmaceutical Co., Ltd. that detects particles in glass bottles. This technology uses AI to learn the difference in movements between bubbles and particles in liquid for injections within glass bottles. The technology is even able to detect micro-particles that have been difficult to identify until now with high accuracy. Both companies aim to commercialize the technology in the 2022 fiscal year.

In the process of inspecting liquids for injections that are contained in glass bottles, such as anti-cancer drugs and vaccines, it is common for bottles suspected of containing particles to be automatically removed by inspection equipment, and for a secondary visual inspection to take place by inspectors. Initially, an image-based inspection is performed on all glass bottles. However, normal bottles are often recognized as defective and eliminated. This is due to bubbles being recognized as particles. In addition, advanced skill and know-how are required for inspectors to perform secondary visual inspections, inspection quality often varies, and securing inspectors at the time of production is a challenge.

NEC has partnered with Ono Pharmaceutical to reduce the number of visual inspections by applying this new AI technology to automated inspections, aiming to realize rapid processes that feature uniform inspection quality and help to support the expansion of production.

Technology Features
1. Detect particles with high accuracy based on inspection procedures by inspectors
Visual inspection by inspectors involves tilting glass bottles and stirring liquid in order to identify the differences in the movement between particles and bubbles. This technology uses a high-speed camera to capture the inside state of the bottle. Then, it is analyzed as slow-motion video in order to accurately understand the movement of particles and bubbles, and to recognize the slight differences. This makes it possible to detect particles with high accuracy.

2. Various types of micro-particles can be detected by focusing on motion
In recent years, AI that is capable of recognizing objects within images by using machine learning technologies, such as deep learning, is attracting attention. However, even if a system learns to see particles in an image, it is not possible for the system to work for non-learned particles. Also, it is impractical to enlarge the image so that small particles and bubbles can be distinguished. This new technology distinguishes them by focusing on the movement in which the specific gravity and shape features of particles appear. This allows the technology to work for various kinds of tiny particles.

Going forward, NEC will continue to collaborate with Ono Pharmaceutical and to proceed with verification activities, aiming to bring inspection equipment using this technology into practical use.

***

About NEC Corporation
NEC Corporation has established itself as a leader in the integration of IT and network technologies while promoting the brand statement of “Orchestrating a brighter world.” NEC enables businesses and communities to adapt to rapid changes taking place in both society and the market as it provides for the social values of safety, security, fairness and efficiency to promote a more sustainable world where everyone has the chance to reach their full potential. For more information, visit NEC at https://www.nec.com.