Fish-Quality Analysis using Artificial Neural Networks and Spectroscopic Data

P. Singhal, J. Zhang, H.E. Michel, and B.R. Singh (USA)


Neural network, spectroscopy, fish quality inspection


When fish die, bacteria or the enzymes they produce invade the flesh of fish. This process produces toxic compounds in the fish and the fish becomes spoiled. Fourier Transform Infrared spectroscopy (FT-IR) allows chemical based non-destructive discrimination of intact microbial cells, and produces complex biochemical fingerprints which are reproducible and distinct for different accumulation of byproducts of spoilage. Data obtained through FT-IR spectroscopic analysis can be used to train an artificial neural network (ANN) for the development of an ANN based FT-IR Screening System for fish-quality inspection. In this paper, we present a performance analysis of various neural networks processing FT-IR spectroscopic data of spoiled and fresh codfish, and show that ANNs can generalize and discriminate with a high degree of accuracy.

Important Links:

Go Back