Use of colorimetric data and artificial neural networks for the determination of freshness in fish

Abstract

The determination of fish freshness is conducted by different methods, of which, in general, reagents with a high degree of toxicity or dangerousness are used, besides requiring the use of large volumes of drinking water and electricity. Thus, the objective was to develop an alternative analytical method, fast, easy to perform and environmentally friendly for the determination of freshness in fish, based on Total Volatile Basic Nitrogen (TVB-N) and CIELab and RGB colorimetric data associated with chemometrics and the Artificial Neural Networks (ANN) technique. Through the evaluation of the figures of merit, it was possible to verify promising results for the use of the developed alternative in future predictions of freshness in fish, demonstrating its suitability for a more robust quality control. In view of the above, the modeling of colorimetric data by ANN models is in line with the requirements of the 4.0 food industry, since it is a fast method and because it is a sustainable alternative not only environmentally, but also economically, since it encourages the application of green and low-cost tools.

Keywords

• Total volatile basic nitrogen
• Colorimetry
• Smartphone
• Machine learning
• Industry 4.0