*This is a machine-translated text of the original article.

Toshiba Digital Solutions Corporation (TDSL) announced that it would offer facial-recognition AI service “Kaometa” for broadcasters and other media.

Kaometa is AI-powered face recognition technology that identifies a person from an image. It is an on-premises version service that emphasizes real-time performance, where needs are high from media customers.

This service realizes high-precision real-time recognition for a variety of shooting environments and facial orientations.

Even from an image including many fluctuation factors such as lighting, facial orientation, and expression, it has a technology capable of extracting information for accurately identifying a person, and it enables high-precision face recognition by only registering one face image per person.

In addition, it is equipped with functions specialized for broadcasting stations, such as a person tracking function that does not interrupt recognition by tracking the detected area of the face, and recognition of a small face projected on a multi-screen that simultaneously handles multiple images.

The company intends to continue expanding its functions to accommodate broadcasting stations, such as linking with broadcasting systems to reduce the burden on program production sites and using cloud services that can be used without being aware of the system environment.

In addition to real-time recognition, it will also contribute to further utilization of broadcast video, such as person identification (metadata addition) that appears in large-capacity past video, which has been difficult to deal with until now.

Original Text: https://ampmedia.jp/2020/04/14/toshiba-digital/