Japanese Firms Develop AI-Assisted Device to Determine Fat Content of Frozen Albacore Tuna

Yomiuri Shimbun file photo
Tuna fish are displayed in a fish wholesale market.

Fujitsu Ltd. and other entities recently announced that they will release in June a device that determines the fat content of frozen albacore tuna with artificial intelligence.

The AI system in the device checks how rich in fat the tuna is, based on waveforms of ultrasonic sound waves exposed to the fish body. The companies plan to sell the device to businesses, such as seafood product processing firms, and also aim to sell them overseas.

Usually, tuna quality is assessed visually by skilled workers who cut tail fins with an electric saw and examine the fat content of the cross-section of the tail. However, due to a shortage of skilled personnel, it’s not possible to inspect all albacore tuna, which has a high volume of distribution. This has resulted in albacore tuna with high fat content sometimes being processed into canned tuna rather than used for sushi or sashimi.

The device developed by Fujitsu in collaboration with other entities, including Sonofai Inc., a startup based in Shizuoka Prefecture, can be operated by a single person without specialized technical knowledge. The inspection takes 12 seconds per fish, significantly shorter than the about 60 seconds required by skilled workers. This enables high-quality tuna to be identified quickly and distributed to the market at appropriate prices.

Fujitsu and the others plan to expand the application of the device to other types of tuna and aim to add functions that evaluate things such as taste and freshness.