Snack Texture Estimation System Using a Simple Equipment and Neural Network Model

Abstract

Texture evaluation is manually performed in general, and such analytical tasks can get cumbersome. In this regard, a neural network model is employed in this study. This paper describes a system that can estimate the food texture of snacks. The system comprises a simple equipment unit and an artificial neural network model. The equipment simultaneously examines the load and sound when a snack is pressed. The neural network model analyzes the load change and sound signals and then outputs a numerical value within the range (0,1) to express the level of textures such as “crunchiness” and “crispness”. Experimental results validate the model’s capacity to output moderate texture values of the snacks. In addition, we applied the convolutional neural network (CNN) model to classify snacks and the capability of the CNN model for texture estimation is discussed.

Measurement of vertical and horizontal vibrations of a probe for acoustic evaluation of food texture

“…Zdunek, Konopacka, and Jesionkowska (2010) recorded the crushing sounds of apple and counted acoustic emission numbers for apple crispness. Recently, the neural network model has also been adopted for evaluating food texture using load and sound data (Kato et al, 2019; Kato & Wada, 2019). Even though the combination of acoustical recording with mechanical tests afforded more reliable data, several difficulties were reported in the collection of reproducible acoustical data (Bourne, 2002; Duizer, 2004; Luyten, Pluter, & Van ilet, 2004; Roudout, Dacremont, Pamies, Colas, & le Meste, 2002; Saeleaw & Schleining, 2011).…”

Classification of crispness of food materials by deep neural networks

“…These networks, including back propagation neural networks (BPNNs), feedforward neural networks (FNNs), and multi-layer perceptrons (MLPs), have been used to analyze acoustic signals generated during mechanical tests on food samples. The frequency range of these signals varies, but often falls within 0-20 kHz (Chen & Ding, 2021;Iliassafov & Shimoni, 2007;Kato et al, 2018Kato et al, , 2019aKato et al, , 2019bLiu, Cai, et al, 2021;Liu & Tan, 1999;Liu, Wu, et al, 2021;Przybył et al, 2020;Sanahuja et al, 2018;Srisawas & Jindal, 2003;Świetlicka et al, 2015).…”

“…Training an ANN to classify crispy foods is the first step. Studies in this area approximated the texture function using force data from texturometers (Kato et al, 2018; Kato et al, 2019a; Tunick et al, 2013). They had challenges regarding the equipment's noise, a small change in the sound caused deviations in the results (Andreani et al, 2020; de Moraes et al, 2022).…”

“…For instance, Chen and Ding (2021) used a BPNN model to analyze the crispness of vegetables like potatoes and carrots, while Iliassafov and Shimoni (2007) used a similar model to predict the sensory crispness of coated turkey breasts. Kato et al (2018, 2019a, 2019b) employed a simple BPNN model to quantify the texture of snacks such as rice crackers and potato chips. Liu and Tan (1999) used a FNN to evaluate the crispness of chex mix products, and Przybył et al (2020) utilized MLP to analyze the quality of dried strawberries.…”

A Magnetic Food Texture Sensor and Comparison of the Measurement Data of Chicken Nuggets

“…If there is sufficient measurement data, artificial intelligence-utilizing techniques may solve this problem. Neural networks have been used to estimate food texture in terms of crispness and crunchiness based on force and acoustic signals [ 18 ]. The other issue is that there is no standard sensor for simultaneously measuring force and vibration.…”