In the present work, a nondestructive device set up for a rapid and reliable freshness assessment of rainbow trout during 10 days of storage in ice was evaluated. The device was characterized by a vector network analyzer interfaced with an open coaxial probe to be placed in contact with the fish eye. The acquisition of the reflected scattering parameter (S11), which is the ratio between the amplitude of the reflected and the incident signal, was assessed in the 50 kHz-3 GHz spectral range. S11 is composed of a real part and an imaginary part, and both parts were used to predict quality index method for freshness evaluations. Partial least squares regression predictive models of the demerit scores related to fish eye attributes (eye pupil and eye shape) and the day of storage were set up. The main results showed that both the real and imaginary parts of the S11 decrease as a function of storage time. The combination with multivariate analysis allowed to set up predictive models of the storage time and the demerit scores with R2 values up to 0.946 (root mean square error [RMSE] = 0.88 days) and 0.942 (RMSE = 3.17 demerit scores related to the fish eyes attributes), respectively (external validation). According to our results, the proposed cheap solution appears a useful tool for the freshness assessment of rainbow trout. PRACTICAL APPLICATION: This work shows that dielectric properties have the potential to discriminate stored fish according to their freshness quality. A device based on this principle can play a significant role in the postharvest processes, contributing to higher product quality and safety and supporting producers and retailers during the qualitative inspections.
Keywords: dielectric properties; principal component analysis; sensory analysis; shelf life.
© 2023 The Authors. Journal of Food Science published by Wiley Periodicals LLC on behalf of Institute of Food Technologists.