Seafood and aquaculture are a significant part of the Norwegian economy and the sectors will continue to grow in the future. Seafood is now Norway's second largest export. However, there are limits to the quantity of marine life that can be either caught or farmed before there are adverse environmental effects. Therefore, to increase profitability while maintaining sustainability, the fishing industry has a long-term strategy to create value through better resource use and optimization of production methods. It is estimated that up to half of all seafood is wasted. Beyond wasted nutritional value, this also causes significant energy waste and production of greenhouse gases from discarded products. Therefore, creating high-quality products with a long shelf life is important both economically and environmentally. A challenge in achieving this goal is that the underlying properties of a raw material that will lead to a low or high quality final product are often poorly understood. Advanced multiscale MRI methods will be adapted to yield information on fish tissue microstructure, chemistry and macrostructure. Machine learning algorithms will relate information from the MRI data to visual and sensory properties to determine what tissue properties yield the most desirable products for four different seafood processing methods. Other machine learning models will predict final sample properties based on original sample properties and handling methods. These results will be used to determine what physical and chemical properties of raw materials are most suited for the different processing methods and make recommendations for process optimization. Generalizability is an important aspect of the project. The project will make the measurement methods and software broad enough such that they can be applied to many types of marine products. Collaboration between Nofima, UiT/UNN, Lund University and UC Berkeley will ensure the research is on the international forefront.
Project leader: Kathryn Anderssen
Institution: NOFIMA AS