Researchers) ITOH Toshio, Senior Researcher, CHOI Pil Gyu, Researcher, MASUDA Yoshitake, Group Leader, Electroceramics Group, Innovative Functional Materials Research Institute, National Institute of Advanced Industrial Science and Technology (AIST)
YOSHIOKA Takeya, Researcher, OGATA Yumi, Researcher, SUGAWARA Tomoaki, Senior Research Manager, Hokkaido Industrial Technology Center, Hakodate Regional Industry Promotion Organization
- Measurement by multiple semiconductor sensors
- Machine learning with simulated freshness indicator gases based on actual gas analysise
- Objectively assessing acceptability for raw consumption for expanding exports of fresh marine products
Measuring odor of fish meat and using machine learning to determine its freshness
With the registration of Japanese cuisine as a UNESCO Intangible Cultural Heritage, eating raw fish such as sushi and sashimi is gaining worldwide acceptance; and chilled fresh marine products are exported from Japan to Southeast Asia and elsewhere. Freshness of marine products is a particularly important quality factor for their prices. In Japanese markets, "connoisseurs" judge the quality of marine products based on their experience and senses, and fresh marine products are sold and offered for raw consumption based on trust with consumers. Overseas, where there are few "connoisseurs," it is difficult for local people to distinguish between marine products for raw consumption and those for cooked use; and most of the marine products are currently handled by Japanese-affiliated stores. In order to increase the export volume of Japanese marine products, objective quality assurance indicators and measurement methods are needed.
Hokkaido Industrial Technology Center applied to Ministry of Agriculture, Forestry and Fisheries (MAFF) to establish a Japanese Agricultural Standard (JAS) for test methods of K-values, the most common scientific freshness index for marine products. In March 2022, "Testing method of K-value as a freshness index for fish - High performance liquid chromatographic method" was established under JAS. However, the derivation of K-values requires several hours even if knowledge workers perform the chemical measurements for K-values in appropriate facilities. In order to quickly determine the state of freshness at distribution sites, there is a need to develop a freshness measurement device that can "visualize" freshness using new sensing technology.
Researchers at AIST, in collaboration with the Hokkaido Industrial Technology Center of Hakodate Regional Industrial Promotion Organization, have developed a sensing technology to determine the freshness of fish meat from its odor, using yellowtail as a model.
Raw fish, such as sushi and sashimi, is becoming increasingly popular around the world, and chilled fresh marine products are exported from Japan to overseas. Overseas, there are few professionals with thorough knowledge of raw fish consumption, and it is difficult to distinguish between marine products for raw consumption and those for cooked use, so most of the marine products are currently handled is by Japanese-affiliated stores. In order to expand the export volume of Japanese marine products, it is necessary to have objective quality assurance indicators and their measurement method. K-value has been proposed as a freshness index for fresh marine products. However, the derivation K-values requires to sample fish meat; and it takes a certain amount of time even if knowledge workers perform the chemical measurements to derive K-values. Therefore, there was a need to develop a new sensing technology to determine freshness of marine products easily.
AIST has developed a new sensing technology for odor determination. This is a non-destructive test that does not require to sample fish meat because it targets fish odors. AIST, in collaboration with the Hokkaido Industrial Technology Center, analyzed the odor of each freshness level of fish meat, and based on the results, prepared simulated freshness indicator gases. Using the measurement results to the indicator gases as training data, machine learning was used to determine freshness of an actual fish meat based on its odor.