Implementation of multilayer perceptron neural network on quality assessment of tomato puree in aerobic storage using electronic nose

• Main Authors: Concepcion, Ronnie S., Sybingco, Edwin, Lauguico, Sandy C., Dadios, Elmer P.
• Format: text
• Published: Animo Repository 2019
• Subjects: Olfactory sensors; Food spoilage; Neural networks (Computer science); Electrical and Computer Engineering; Electrical and Electronics
• Online Access: https://animorepository.dlsu.edu.ph/faculty_research/1912
• Institution: De La Salle University

The adulteration of food increases the number of bacteria being develop in it that is primarily affected from oxygen exposure and varying temperature, not suitable for its storage. In such case, food spoilage happens and leads to food poisoning. Tomato-based dishes stored in aerobic environment significantly varies its shelf-life. However, misclassification due to subjective human assumptions is the major problem on assessing the quality of food. To address this problem, a proposed solution is the development of an intelligent electronic nose (eNose) system that will discriminate the condition of tomato puree using artificial neural network (ANN) based only on ammonia and methane concentrations, and pH level. This system is composed of five sections: the development of electronic nose using Gizduino microcontroller and Mngan Q lai (MQ) gas sensors, olfactory data acquisition, generation of smellprint, design of ANN, and the implementation of ANN for classification of tomato puree condition. This study substantially presents analysis on computational parameters of ANN. The collection data rate was set to 2 Hz for tomato puree-emitted gas samples with varying shelf life considering outdoor aerobic storage. Multilayer perceptron neural network was implemented using feedforward backpropagation algorithm. The number of hidden layers and artificial neurons were analyzed based on performance of the system computational parameters, namely, cross-entropy (CE), learning time and regression (R) coefficient. The system classifies the tomato puree sample as not spoiled, partially spoiled, and spoiled. The smellprint of each food condition was generated and the tomato puree-spoilage determinant parameters were characterized. Through 3-layer perceptron ANN with 120 and 50 artificial neurons on the first and second hidden layers respectively, an accuracy of 93.33% was yielded for tomato puree quality deterioration classification. The developed mechanism is a potential application in domotics. © 2019 IEEE.