CNN–LSTM Neural Network for Identification of Pre-Cooked Pasta Products in Different Physical States Using Infrared Spectroscopy

Abstract

Infrared (IR) spectroscopy is nondestructive, fast, and straightforward. Recently, a growing number of pasta companies have been using IR spectroscopy combined with chemometrics to quickly determine sample parameters. However, fewer models have used deep learning models to classify cooked wheat food products and even fewer have used deep learning models to classify Italian pasta. To solve these problems, an improved CNN–LSTM neural network is proposed to identify pasta in different physical states (frozen vs. thawed) using IR spectroscopy. A one-dimensional convolutional neural network (1D-CNN) and long short-term memory (LSTM) were constructed to extract the local abstraction and sequence position information from the spectra, respectively. The results showed that the accuracy of the CNN–LSTM model reached 100% after using principal component analysis (PCA) on the Italian pasta spectral data in the thawed state and 99.44% after using PCA on the Italian pasta spectral data in the frozen form, verifying that the method has high analytical accuracy and generalization. Therefore, the CNN–LSTM neural network combined with IR spectroscopy helps to identify different pasta products.

Classification of Food Additives Using UV Spectroscopy and One-Dimensional Convolutional Neural Network

“…This algorithm automatically determines the input dataset's important features without external supervision. CNN has been successfully used for texture feature detection [42], hyperspectral image classification [43], medical diagnosis [44], and facial recognition [45] and has also been associated with optical techniques for food analysis [37][38][39]. The most popular CNN architectures are AlexNet, GoogLeNet, ResNet, X-ception, Inception-V4, and others [46].…”

“…In contrast, the determination of aflatoxin B 1 in maize was accomplished using NIR (near-infrared) spectroscopy, Markov transition field (MTF) image coding, and a CNN [36]. Infrared spectroscopy and CNN were also used for the classification of Italian pasta [37], melamine and cyanuric acid detection [38], yali pear inspection [39], and polysaccharide detection [40].…”

Multimodal Gait Abnormality Recognition Using a Convolutional Neural Network–Bidirectional Long Short-Term Memory (CNN-BiLSTM) Network Based on Multi-Sensor Data Fusion

“…The CNN-LSTM network is a hybrid neural network that has achieved a wide range of applications in several fields, including emotion recognition [29], video action classification [30], pasta product classification [31], facial micro-expression recognition [32], gait recognition [33], and stock prediction [34]. Its main framework is made of a convolutional layer spliced with a BiLSTM layer, and then the attention mechanism and other networks can be added according to the data characteristics and task requirements.…”