Modeling the Degradation Process of Tomato Paste Color during the Thermal Process Using Artificial Neural Networks Methodology

Color is one of the important quality factors of processed tomato products such as paste and is strongly affected by the thermal process. The main purpose of this study was to evaluate the kinetics of paste color degradation during the thermal process using artificial neural networks. For this purpose, tomato paste was processed at three temperatures of 60, 70 and 80 degrees Celsius for 25 to 100 minutes and using three main color indicators including: L, a, and b, a/b ratio, Color Difference Overall (TCD), saturation index and Hugh angle were determined. The degradation kinetics of these parameters was investigated and modeled using an artificial neural network model. Index b with the highest and TCD and a/b with the lowest activation energy had the highest and lowest sensitivity to temperature changes, respectively. According to the values of MSE and R2 presented in the present study, the feed neural network with logarithmic transfer function and topology 3-8-3 (input layer with three neurons-a hidden layer with eight neuron-output layer with three neurons) with more than R² 1.38 and MSE equal to 1.2228 were selected as the optimal neural network. Application of artificial neural network in this project proved to be a useful tool to determine the tomato paste quality through non-invasive, low-cost and real time processes.

Author(s): Seyedeh Sedigheh Hashemi*, Mohamad Reza Mozafarian and Mohammad Ganjeh

Abstract | Full-Text | PDF

Share This Article