Nondestructive Freshness Discriminating of Shrimp Using Visible/Near-Infrared Hyperspectral Imaging Technique and Deep Learning Algorithm

Non-destructive and in-situ detection of shrimp freshness using mid-infrared fiber-optic evanescent wave spectroscopy

There is strong interest in non-destructive and rapid determination of food freshness in food research. In this study, mid-infrared (MIR) fiber-optic evanescent wave (FOEW) spectroscopy was applied to monitor shrimp freshness through the evaluation of protein, chitin, and calcite contents in conjunction with a Partial Least Squares Discriminant Analysis (PLS-DA) model. Shrimp shells were wiped with a micro fiber-optic probe to obtain a FOEW spectrum which quickly and nondestructively allowed evaluation of the shrimp freshness. Peaks for proteins, chitin, and calcite, which are closely related to shrimp freshness, were detected and quantified. Compared with the standard indicator for evaluating shrimp freshness (total volatile basic nitrogen), the PLS-DA model gave recognition rates for shrimp freshness using calibration and validation sets of the FOEW data of 87.27%, 90.28%, respectively. Our results show that FOEW spectroscopy is a feasible method for non-destructive and in-site detection of shrimp freshness.

Machine learning models to predict micronutrient profile in food after processing

The information on nutritional profile of cooked foods is important to both food manufacturers and consumers, and a major challenge to obtaining precise information is the inherent variation in composition across biological samples of any given raw ingredient. The ideal solution would address precision and generability, but the current solutions are limited in their capabilities; analytical methods are too costly to scale, retention-factor based methods are scalable but approximate, and kinetic models are bespoke to a food and nutrient. We provide an alternate solution that predicts the micronutrient profile in cooked food from the raw food composition, and for multiple foods. The prediction model is trained on an existing food composition dataset and has a 31% lower error on average (across all foods, processes and nutrients) than predictions obtained using the baseline method of retention-factors. Our results argue that data scaling and transformation prior to training the models is important to mitigate any yield bias. This study demonstrates the potential of machine learning methods over current solutions, and additionally provides guidance for the future generation of food composition data, specifically for sampling approach, data quality checks, and data representation standards.

Multiclass Skin Lesion Classification Using a Novel Lightweight Deep Learning Framework for Smart Healthcare

Skin lesion classification has recently attracted significant attention. Regularly, physicians take much time to analyze the skin lesions because of the high similarity between these skin lesions. An automated classification system using deep learning can assist physicians in detecting the skin lesion type and enhance the patient’s health. The skin lesion classification has become a hot research area with the evolution of deep learning architecture. In this study, we propose a novel method using a new segmentation approach and wide-ShuffleNet for skin lesion classification. First, we calculate the entropy-based weighting and first-order cumulative moment (EW-FCM) of the skin image. These values are used to separate the lesion from the background. Then, we input the segmentation result into a new deep learning structure wide-ShuffleNet and determine the skin lesion type. We evaluated the proposed method on two large datasets: HAM10000 and ISIC2019. Based on our numerical results, EW-FCM and wide-ShuffleNet achieve more accuracy than state-of-the-art approaches. Additionally, the proposed method is superior lightweight and suitable with a small system like a mobile healthcare system.

Using artificial intelligence to automate meat cut identification from the semimembranosus muscle on beef boning lines

The identification of different meat cuts for labeling and quality control on production lines is still largely a manual process. As a result, it is a labor-intensive exercise with the potential for not only error but also bacterial cross-contamination. Artificial intelligence is used in many disciplines to identify objects within images, but these approaches usually require a considerable volume of images for training and validation. The objective of this study was to identify five different meat cuts from images and weights collected by a trained operator within the working environment of a commercial Irish beef plant. Individual cut images and weights from 7,987 meats cuts extracted from semimembranosus muscles (i.e., Topside muscle), post editing, were available. A variety of classical neural networks and a novel Ensemble machine learning approaches were then tasked with identifying each individual meat cut; performance of the approaches was dictated by accuracy (the percentage of correct predictions), precision (the ratio of correctly predicted objects relative to the number of objects identified as positive), and recall (also known as true positive rate or sensitivity). A novel Ensemble approach outperformed a selection of the classical neural networks including convolutional neural network and residual network. The accuracy, precision, and recall for the novel Ensemble method were 99.13%, 99.00%, and 98.00%, respectively, while that of the next best method were 98.00%, 98.00%, and 95.00%, respectively. The Ensemble approach, which requires relatively few gold-standard measures, can readily be deployed under normal abattoir conditions; the strategy could also be evaluated in the cuts from other primals or indeed other species.

Handheld Devices for Food Authentication and Their Applications: A Review

This review summarises miniaturised technologies, commercially available devices, and device applications for food authentication or measurement of features that could potentially be used for authentication. We first focus on the handheld technologies and their generic characteristics: (1) technology types available, (2) their design and mode of operation, and (3) data handling and output systems. Subsequently, applications are reviewed according to commodity type for products of animal and plant origin. The 150 applications of commercial, handheld devices involve a large variety of technologies, such as various types of spectroscopy, imaging, and sensor arrays. The majority of applications, ~60%, aim at food products of plant origin. The technologies are not specifically aimed at certain commodities or product features, and no single technology can be applied for authentication of all commodities. Nevertheless, many useful applications have been developed for many food commodities. However, the use of these applications in practice is still in its infancy. This is largely because for each single application, new spectral databases need to be built and maintained. Therefore, apart from developing applications, a focus on sharing and re-use of data and calibration transfers is pivotal to remove this bottleneck and to increase the implementation of these technologies in practice.

Artificial intelligence, big data, and blockchain in food safety

Food safety plays an essential role in our daily lives, and it becomes serious with the development of worldwide trade. To tackle the food safety issues, many advanced technologies have been developed to monitor the process of the food industry (FI) to ensure food safety, including the process of food production, processing, transporting, storage, and retailing. These technologies are often referred to as artificial intelligence (AI), big data, and blockchain, which have been widely applied in many research areas. In this review, we introduce these technologies and their applications in the food safety domain. Firstly, basic concepts of these technologies are presented. Then, applications for food safety from a data perspective based on these technologies are analyzed. Finally, future challenges of the applications of AI, big data, and blockchain are discussed.

Performance Analysis of Deep Learning CNN Models for Variety Classification in Hazelnut

In evaluating agricultural products, knowing the specific product varieties is important for the producer, the industrialist, and the consumer. Human labor is widely used in the classification of varieties. It is generally performed by visual examination of each sample by experts, which is very laborious and time-consuming with poor sensitivity. There is a need in commercial hazelnut production for a rapid, non-destructive and reliable variety classification in order to obtain quality nuts from the orchard to the consumer. In this study, a convolutional neural network, which is one of the deep learning methods, was preferred due to its success in computer vision. A total of 17 widely grown hazelnut varieties were classified. The proposed model was evaluated by comparing with pre-trained models. Accuracy, precision, recall, and F1-Score evaluation metrics were used to determine the performance of classifiers. It was found that the proposed model showed a better performance than pre-trained models in terms of performance evaluation criteria. The proposed model was found to produce 98.63% accuracy in the test set, including 510 images. This result has shown that the proposed model can be used practically in the classification of hazelnut varieties.

Evaluation of the total volatile basic nitrogen (TVB-N) content in fish fillets using hyperspectral imaging coupled with deep learning neural network and meta-analysis

Recently, hyperspectral-imaging (HSI), as a rapid and non-destructive technique, has generated much interest due to its unique potential to monitor food quality and safety. The specific aim of the study is to investigate the potential of the HSI (430-1010 nm) coupled with Linear Deep Neural Network (LDNN) to predict the TVB-N content of rainbow trout fillet during 12 days storage at 4 ± 2 °C. After the acquisition of hyperspectral images, the TVB-N content of fish fillets was obtained by a conventional method (micro-Kjeldahl distillation). To simplify the calibration models, nine optimal wavelengths were selected by the successive projections algorithm. A seven layers LDNN was designed to estimate the TVB-N content of samples. The LDNN model showed acceptable performance for prediction of TVB-N content of fish fillet (R²p = 0.853; RSMEP = 3.159 and RDP = 3.001). The performance of LDNN model was comparable with the results of previous works. Although, the results of the meta-analysis did not show any significant difference between various chemometric models. However, the least-squares support vector machine algorithm showed better prediction results as compared to the other models (RMSEP: 2.63 and R²_p = 0.897). Further studies are required to improve the prediction power of the deep learning model for prediction of rainbow-trout fish quality.

Rapid and non-destructive spectroscopic method for classifying beef freshness using a deep spectral network fused with myoglobin information

A simple, novel, rapid, and non-destructive spectroscopic method that employs the deep spectral network for beef-freshness classification was developed. The deep-learning-based model classified beef freshness by learning myoglobin information and reflectance spectra over different freshness states. The reflectance spectra (480-920 nm) were measured from 78 beef samples for 17 days, and the datasets were sorted into three freshness classes based on their pH values. Myoglobin information showed statistically significant differences depending on the freshness; consequently, it was utilized as a crucial parameter for classification. The model exhibited improved performance when the reflectance spectra were combined with the myoglobin information. The accuracy of the proposed model improved to 91.9%, whereas that of the single-spectra model was 83.6%. Further, a high value for the area under the receiver operating characteristic curve (0.958) was recorded. This study provides a basis for future studies on the investigation of myoglobin information associated with meat freshness.

Hyperspectral Imaging (HSI) Technology for the Non-Destructive Freshness Assessment of Pearl Gentian Grouper under Different Storage Conditions

This study used visible/near-infrared hyperspectral imaging (HSI) technology combined with chemometric methods to assess the freshness of pearl gentian grouper. The partial least square discrimination analysis (PLS-DA) and competitive adaptive reweighted sampling-PLS-DA (CARS-PLS-DA) models were used to classify fresh, refrigerated, and frozen–thawed fish. The PLS-DA model achieved better classification of fresh, refrigerated, and frozen–thawed fish with the accuracy of 100%, 96.43%, and 96.43%, respectively. Further, the PLS regression (PLSR) and CARS-PLS regression (CARS-PLSR) models were used to predict the storage time of fish under different storage conditions, and the prediction accuracy was assessed using the prediction correlation coefficients (R_p²), root mean squared error of prediction (RMSEP), and residual predictive deviation (RPD). For the prediction of storage time, the CARS-PLS model presented the better result of room temperature (R_p² = 0.948, RMSEP = 0.255, RPD = 4.380) and refrigeration (R_p² = 0.9319, RMSEP = 1.188, RPD = 3.857), while the better prediction of freeze was by obtained by the PLSR model (R_p² = 0.9250, RMSEP = 2.910, RPD = 3.469). Finally, the visualization of storage time based on the PLSR model under different storage conditions were realized. This study confirmed the potential of HSI as a rapid and non-invasive technique to identify fish freshness.

Advances in infrared spectroscopy combined with artificial neural network for the authentication and traceability of food

The authentication and traceability of food attract more attention due to the increasing consumer awareness regarding nutrition and health, being a new hotspot of food science. Infrared spectroscopy (IRS) combined with shallow neural network has been widely proven to be an effective food analysis technology. As an advanced deep learning technology, deep neural network has also been explored to analyze and solve food-related IRS problems in recent years. The present review begins with brief introductions to IRS and artificial neural network (ANN), including shallow neural network and deep neural network. More notably, it emphasizes the comprehensive overview of the advances of the technology combined IRS with ANN for the authentication and traceability of food, based on relevant literature from 2014 to early 2020. In detail, the types of IRS and ANN, modeling processes, experimental results, and model comparisons in related studies are described to set forth the usage and performance of the combined technology for food analysis. The combined technology shows excellent ability to authenticate food quality and safety, involving chemical components, freshness, microorganisms, damages, toxic substances, and adulteration. As well, it shows excellent performance in the traceability of food variety and origin. The advantages, current limitations, and future trends of the combined technology are further discussed to provide a thoughtful viewpoint on the challenges and expectations of online applications for the authentication and traceability of food.

Using near-infrared hyperspectral imaging with multiple decision tree methods to delineate black tea quality

The evaluation of tea quality tended to be subjective and empirical by human panel tests currently. A convenient analytical approach without human involvement was developed for the quality assessment of tea with great significance. In this study, near-infrared hyperspectral imaging (HSI) combined with multiple decision tree methods was utilized as an objective analysis tool for delineating black tea quality and rank. Data fusion that integrated texture features based on gray-level co-occurrence matrix (GLCM) and short-wave near-infrared spectral features were as the target characteristic information for modeling. Three different types of supervised decision tree algorithms (fine tree, medium tree, and coarse tree) were proposed for the comparison of the modeling effect. The results indicated that the performance of models was enhanced by the multiple perception feature fusion. The fine tree model based on data fusion obtained the best predictive performance, and the correct classification rate (CCR) of evaluating black tea quality was 93.13% in the prediction process. This work demonstrated that HSI coupled with intelligence algorithms as a rapid and effective strategy could be successfully applied to accurately identify the rank quality of black tea.

NIR Hyperspectral Imaging Technology Combined with Multivariate Methods to Identify Shrimp Freshness

In this study, a hyperspectral imaging system of 866.4–1701.0 nm, combined with a variety of spectral processing methods were adopted to identify shrimp freshness. To gain the optimal model combination, three preprocessing methods (Savitzky-Golay first derivative (SG1), multivariate scatter correction (MSC), and standard normal variate (SNV)), three characteristic wavelength extraction algorithms (random frog algorithm (RFA), uninformative variables elimination (UVE), and competitive adaptive reweighted sampling (CARS)), and four discriminant models (partial least squares discrimination analysis (PLS-DA), least squares support vector machine (LSSVM), random forest (RF), and extreme learning machine (ELM)) were employed for experimental study. First of all, due to the full wavelength modeling analysis, three preprocessing methods were utilized to preprocess the original spectral data. The analysis showed that the spectral data processed by the SNV method had the best performance among the four discriminant models. Secondly, due to the characteristic wavelength modeling analysis, three characteristic wavelength extraction algorithms were utilized to extract the characteristic wavelength of the SNV-processed spectral data. It was found that the CARS algorithm achieved the best performance among the three characteristic wavelength extraction algorithms, and the combining adoption of the ELM model and different characteristic wavelength extraction algorithms obtained the best results. Therefore, the model based on SNV-CARS-ELM obtained the best performance and was elected as the optimal model. Lastly, for accurately and explicitly displaying the refrigeration days of shrimps, the original hyperspectral images of shrimps were substituted into the SNV-CARS-ELM model, thus obtaining the general classification accuracy of 97.92%, and the object-wise method was used to visualize the classification results. As a result, the method proposed in this study can effectively detect the freshness of shrimps.

Textural and quality changes of hairtail fillets (Trichiurus haumela) related with water distribution during simulated cold chain logistics

This study investigated the total viable counts, sensory and physicochemical qualities (total volatile basic nitrogen, K-value, thiobarbituric acid reactive substances value, water holding capacity and texture parameters), as well as water distribution of fresh hairtail fish during simulated cold chain. The results showed that total volatile basic nitrogen, thiobarbituric acid reactive substances and K-values increased with the increasing time, while sensory scores, water holding capacity, hardness and elasticity decreased. The transverse relaxation time T2 data detected by low-field nuclear magnetic resonance also showed that T22 (trapped water) gradually decreased with the increasing time, while T23 (free water) increased. It was observed that the quality deterioration of the fish fillets developed more quickly when the samples suffered frequent temperature fluctuations than they stored at higher but stable temperatures. The changes of T22 and T23 of both the samples stored at stable and fluctuated temperatures showed good correlations with sensory, total volatile basic nitrogen and thiobarbituric acid reactive substances values (R2 > 0.9, p < 0.05). Therefore, low-field nuclear magnetic resonance technology can be a potential tool to monitor the quality changes of hairtail fish during cold chain logistics.

Identification of Soybean Varieties Using Hyperspectral Imaging Coupled with Convolutional Neural Network

Soybean variety is connected to stress resistance ability, as well as nutritional and commercial value. Near-infrared hyperspectral imaging was applied to classify three varieties of soybeans (Zhonghuang37, Zhonghuang41, and Zhonghuang55). Pixel-wise spectra were extracted and preprocessed, and average spectra were also obtained. Convolutional neural networks (CNN) using the average spectra and pixel-wise spectra of different numbers of soybeans were built. Pixel-wise CNN models obtained good performance predicting pixel-wise spectra and average spectra. With the increase of soybean numbers, performances were improved, with the classification accuracy of each variety over 90%. Traditionally, the number of samples used for modeling is large. It is time-consuming and requires labor to obtain hyperspectral data from large batches of samples. To explore the possibility of achieving decent identification results with few samples, a majority vote was also applied to the pixel-wise CNN models to identify a single soybean variety. Prediction maps were obtained to present the classification results intuitively. Models using pixel-wise spectra of 60 soybeans showed equivalent performance to those using the average spectra of 810 soybeans, illustrating the possibility of discriminating soybean varieties using few samples by acquiring pixel-wise spectra.

Combining near-infrared hyperspectral imaging with elemental and isotopic analysis to discriminate farm-raised pacific white shrimp from high-salinity and low-salinity environments

White shrimp (Litopenaeus vannamei) raised in low-salinity farm are considered inferior to those in seawater. In order to develop a rapid discrimination method for the food industry, we investigated the potential of using near-infrared hyperspectral imaging to discriminate shrimp muscle samples from freshwater and seawater farms. We constructed 3 different discrimination models with 4 optimal wavelength selection methods and compared the performance of each model. The results showed that sequential forward selection combined with partial least squares discriminant analysis (SFS-PLS-DA) generated the best discrimination performance with an overall accuracy of 99.2%. The elemental and isotopic analysis indicated a high correlation between 918 and 925 nm region (which was selected by SFS) and ¹³C concentration. This agrees with the fact that there is more ¹³C in shrimp of salty water compared to those of freshwater. The results demonstrated (hyperspectral imaging) HSI is promising to discriminate L. vannamei raised in fresh and seawater environments.

Deep Learning Meets Hyperspectral Image Analysis: A Multidisciplinary Review

Modern hyperspectral imaging systems produce huge datasets potentially conveying a great abundance of information; such a resource, however, poses many challenges in the analysis and interpretation of these data. Deep learning approaches certainly offer a great variety of opportunities for solving classical imaging tasks and also for approaching new stimulating problems in the spatial-spectral domain. This is fundamental in the driving sector of Remote Sensing where hyperspectral technology was born and has mostly developed, but it is perhaps even more true in the multitude of current and evolving application sectors that involve these imaging technologies. The present review develops on two fronts: on the one hand, it is aimed at domain professionals who want to have an updated overview on how hyperspectral acquisition techniques can combine with deep learning architectures to solve specific tasks in different application fields. On the other hand, we want to target the machine learning and computer vision experts by giving them a picture of how deep learning technologies are applied to hyperspectral data from a multidisciplinary perspective. The presence of these two viewpoints and the inclusion of application fields other than Remote Sensing are the original contributions of this review, which also highlights some potentialities and critical issues related to the observed development trends.