Improving Raman spectroscopic identification of rice varieties by feature extraction

Sha M., Gui D., Zhang Z., Ji X., Shi X., Liu J., Zhang D.

The constituents of rice are heterogeneously distributed in a grain, collection of Raman spectra providing a better compositional representation of rice is an essential requirement for accurate discrimination of rice samples according to geographical origin. Homogeneity of rice flour with four different particle sizes was investigated by relative standard deviation (RSD) analysis and hierarchical clustering analysis (HCA) of Raman spectra. RSDs of Raman spectra of rice flour at 100–140 mesh were the smallest while HCA showed the highest similarities. Besides, Raman spectra of rice flour at 100–140 mesh were similar to those of rice flour with diameter below 0.6 mm. In addition, the experimental results were universally applicable for different batches and geographical origins of rice. The discrimination accuracy performed by support vector machine was obviously improved when using the Raman data of rice flour at the size of 100–140 mesh, hence, the recorded Raman spectra could provide reproducible and reliable data for discrimination the geographical origin of rice.

Li Z.J., Chen Y.H., Zhang D.J., Zhang G.F., Lu B.X.

Gore E., Mardon J., Cécile B., Lebecque A.

Li Y., Li F., Yang X., Guo L., Huang F., Chen Z., Chen X., Zheng S.

A rapid quantitative analysis model for determining the glycated albumin (GA) content based on Attenuated total reflectance (ATR)-Fourier transform infrared spectroscopy (FTIR) combining with linear SiPLS and nonlinear SVM has been developed. Firstly, the real GA content in human serum was determined by GA enzymatic method, meanwhile, the ATR-FTIR spectra of serum samples from the population of health examination were obtained. The spectral data of the whole spectra mid-infrared region (4000-600 cm-1) and GA's characteristic region (1800-800 cm-1) were used as the research object of quantitative analysis. Secondly, several preprocessing steps including first derivative, second derivative, variable standardization and spectral normalization, were performed. Lastly, quantitative analysis regression models were established by using SiPLS and SVM respectively. The SiPLS modeling results are as follows: root mean square error of cross validation (RMSECVT) = 0.523 g/L, calibration coefficient (RC) = 0.937, Root Mean Square Error of Prediction (RMSEPT) = 0.787 g/L, and prediction coefficient (RP) = 0.938. The SVM modeling results are as follows: RMSECVT = 0.0048 g/L, RC = 0.998, RMSEPT = 0.442 g/L, and Rp = 0.916. The results indicated that the model performance was improved significantly after preprocessing and optimization of characteristic regions. While modeling performance of nonlinear SVM was considerably better than that of linear SiPLS. Hence, the quantitative analysis model for GA in human serum based on ATR-FTIR combined with SiPLS and SVM is effective. And it does not need sample preprocessing while being characterized by simple operations and high time efficiency, providing a rapid and accurate method for GA content determination.

Zhu L., Sun J., Wu G., Wang Y., Zhang H., Wang L., Qian H., Qi X.

The processing and quality properties of rice are significantly influenced by its variety and region of origin. However, discriminating between varieties and geographic regions is an urgent but difficult and time-consuming endeavor in China. In this study, an effective and reliable identification method was established by combining Raman spectroscopy (RS) with multivariate data analysis methods. Numerous RS spectra were collected, and the sensitive fundamental vibrations of less polar groups and bonds in rice were analyzed. Principal component analysis (PCA) was used for preliminary identification. Subsequently, different modeling methods were compared and seemed to reliably identify rice types, varieties, and region of origin, with accuracies of between 80 and 100%. As a result, a soft independent modeling of class analogy (SIMCA) model was shown to be the superior model for rice identification. The SIMCA model can deliver high precision detection of adulterated rice (i.e., rice of high quality blended with rice of inferior quality), and this study lays the foundations for an advanced rice quality identification technology system.

Lim D.K., Mo C., Lee D., Long N.P., Lim J., Kwon S.W.

Cebi N., Dogan C.E., Develioglu A., Yayla M.E., Sagdic O.

l-Cysteine is deliberately added to various flour types since l-Cysteine has enabled favorable baking conditions such as low viscosity, increased elasticity and rise during baking. In Turkey, usage of l-Cysteine as a food additive isn't allowed in wheat flour according to the Turkish Food Codex Regulation on food additives. There is an urgent need for effective methods to detect l-Cysteine in wheat flour. In this study, for the first time, a new, rapid, effective, non-destructive and cost-effective method was developed for detection of l-Cysteine in wheat flour using Raman microscopy. Detection of l-Cysteine in wheat flour was accomplished successfully using Raman microscopy combined chemometrics of PCA (Principal Component Analysis) and HCA (Hierarchical Cluster Analysis). In this work, 500-2000cm-1 spectral range (fingerprint region) was determined to perform PCA and HCA analysis. l-Cysteine and l-Cystine were determined with detection limit of 0.125% (w/w) in different wheat flour samples.

Huo Y., Kamal G.M., Wang J., Liu H., Zhang G., Hu Z., Anwar F., Du H.

Food frauds related to the mislabeling and mixing of products of inferior quality with those of superior quality are a serious concern nowadays. NMR-based metabolomics has great potential in the authentication of foods for quality assurance and the tracing of fraudulent labeling. The present study was conducted to discriminate rice from geographically different provinces of China. The study reports the potential use of 1H NMR spectroscopy coupled with PCA and a discriminant analysis method, LDA for metabolomic fingerprinting of Chinese rice. A total of 106 rice samples from nine different provinces of China were analyzed for 1H NMR-based metabolomics. Both the whole variable analysis (heat map) and the Principal Component Analysis (PCA) showed a clear separation among the samples. Linear Discriminant Analysis (LDA) was conducted to extract the variables majorly responsible for this separation, such as sucrose, fructose, glucose, succinate, polyphenols, trigonelline and asparagine. The discrimination was explained on the basis of variations in latitude, temperature and rainfall in these provinces. The study highlights the application of 1H NMR for geographical discrimination of rice and its usefulness for consumers while choosing their desired variety of rice.

Bett-Garber K.L., Bryant R.J., Grimm C.C., Chen M., Lea J.M., McClung A.M.

There is a steady demand for imported basmati and jasmine rice in the United States. Rice varieties that can be domestically produced and compete with these imports have been developed from basmati, jasmine, and other aromatic germplasm sources. This study evaluated differences among eight U.S. aromatic varieties and imported basmati and jasmine samples. Basmati market types (Aromatic se2, Sierra, Dellmati, and Dellrose) and jasmine market types (JES, Jasmine 85, Jazzman, and Charleston Gold) grown in Arkansas and Texas were evaluated for descriptive flavors, apparent amylose, protein, and lipid contents, pasting profile, alkali spreading value, volatiles, grain color, grain dimensions, and agronomic traits. Seven natural flavor attributes and six volatiles differentiated the varieties within the jasmine and basmati classes along with several physicochemical traits, such as pasting profiles, grain dimensions, and grain color. U.S. varieties developed for either the basmati or jasmine market all had a flav...

Sha M., Zhang Z., Gui D., Wang Y., Fu L., Wang H.

A fingerprinting approach was developed by means of high-performance ion mobility spectrometry with direct electrospray ionization (ESI-HPIMS) for the quality consistency and authentication of apple essence which contain many analytes. Thirty-four apple essence samples of the same brand and commercialized as a same product but brewed in four different manufacturers were used to establish the fingerprints. Hierarchical clustering analysis (HCA) was performed to evaluate the similarity and variation of these samples. A combined data matrix was constructed with the use of individual data matrix of ion mobility spectrum in positive and negative ion modes. For comparison, ion mobility spectrum fingerprints in positive and negative ion modes respectively were also applied to the quality assessment of the same samples. Finally, our study demonstrated that the fusion of fingerprints did indeed provide more information and could be used to comprehensively conduct the quality consistency evaluation and discrimination of apple essences from similar products. It is suggested that this fingerprint approach is suitable for analysis of other complex, multianalyte substances.

Li Y., Zhang J., Li T., Liu H., Li J., Wang Y.

Chae Y.K., Kim S.H.

Rice is one of the most important crops that feed almost half of the world's population. With the increasing concern of consumers on the integrity of the product, efforts have been made to develop analytic techniques to discriminate rice products according to their origins or cultivars, but those efforts were mostly based on elemental analysis. We postulated that such discrimination would be possible with the global metabolite profiles. Nineteen metabolites of three different rice cultivars from four different geographical origins were identified from the extracts and compared with one another by the two-dimensional nuclear magnetic resonance (NMR) spectroscopy. NMR data were analyzed with the help of the metabolome database and the statistics software. The different rice samples were successfully separated in the principal component space, showing that the global metabolite profiles can be used to discriminate geographic origins. Our results show that the metabolite analysis via 1H–13C heteronuclear single quantum coherence spectra combined with the statistical method can be applied to discriminate the geographic origins or cultivars of rice samples, thus can provide a means to inspect and pick up fraudulent labeling or adulteration.

Borges E.M., Gelinski J.M., de Oliveira Souza V.C., Barbosa Jr. F., Batista B.L.

50 rice samples (18 organic and 32 ordinary) from Brazil were analyzed with inductively coupled plasma mass spectrometry (ICP-MS) for 20 elements (As, B, Ba, Ca, Cd, Ce, Co, Cr, Cu, Fe, K, La, Mg, Mn, Mo, P, Pb, Rb, Se and Zn) to identify significant differences between the two types (organic and ordinary) of rice. Concentrations of As, B, Ba, Co, Cr, Cu, Mn, P and Zn were found to be higher in ordinary versus organic rice, while K, Ca, Mo, Rb and Se concentrations were lower in ordinary versus organic samples. The remaining investigated elements (Cd, Ce, Fe, La, Mg and Pb) exhibited statistically equivalent concentration in the two types of rice. Principal Component Analysis (PCA), Soft Independent Modeling of Class Analogy (SIMCA), Hierarchical Cluster Analysis (HCA) and K-nearest neighbors (KNN) statistical techniques of the elemental fingerprints were readily able to discriminate organic from ordinary samples and can be used as alternative methods for adulteration evaluation.

Wu Z., Xu E., Long J., Wang F., Xu X., Jin Z., Jiao A.

Effective fermentation monitoring is a growing need during the manufacture of wine due to the rapid pace of change in the wine industry. Ethanol and reducing sugar are two most important process variables indicating the status of Chinese rice wine (CRW) fermentation process. In this study, the potentials of Raman spectroscopy (RS) as a rapid process analytical technique to monitor the evolution of these two chemical parameters involved in CRW fermentation process and to group samples according to different fermentation stages were investigated. The results demonstrated that compared with the PLS model using all wavelengths of Raman spectra, the prediction precision of model based on the spectral variables selected by competitive adaptive reweighted sampling (Cars) was significantly improved. In addition, nonlinear models outperformed linear models in predicting fermentation parameters. After systemically comparison and discussion, it was found that for both ethanol and glucose, Cars-support vector machine (Cars-SVM) models gave the best results with the highest prediction precisions. Moreover, the results obtained from discriminant partial least squares analysis (DPLS) showed that good performances were obtained with an average correct classification rate of 94.9% for different fermentation stages. The overall results indicated that RS combined with efficient variable selection algorithm and nonlinear regression tool could be utilized as a rapid method to monitor CRW fermentation process.

Wu Z., Li H., Long J., Xu E., Xu X., Jin Z., Jiao A.

Discrimination of Chinese rice wines from four geographical origins (‘Zhejiang’, ‘Jiangsu’, ‘Shanghai’ and “Fujian”) in China was investigated according to their UV–vis spectra. The UV–vis absorption spectra of 112 samples in the wavelength range of 190–800 nm were collected. Then the absorption data was subjected to principal component analysis to reveal differences between samples and the potential possibility of using multivariate methods to distinguish differences. Classification models were developed by soft independent modelling of class analogy, linear discriminate analysis (LDA), discriminant partial least squares and support vector machine. Seven mathematical pre-treatments were applied to improve the performance of the multivariate classification models. Among the seven pre-treatments, standard normal transformation (SNV) was superior to the other six methods. The results showed that, compared with other models, SNV-treated LDA models achieved better performances with an average correct classification rate of 98.96% in the training set and 100% in the testing set. The results demonstrate that UV spectroscopy combined with chemometric data analysis, as a rapid method to classify Chinese rice wines according to their geographical origins, is feasible. Copyright © 2015 The Institute of Brewing & Distilling

Zade S.V., Sahebi H., Alizadeh A.M., Jannat B., Rastegar H., Abedinzadeh S., Hashempour-Baltork F., Khaneghah A.M.

Li Z., Miao Z., Li C., Zhou Y., Qiu Y., Liu C., Teng X., Tan Y.

Kaur M., Khan U.M., Bayram I., Bhatia S., Gupta U.

Chai M., Hasi W., Ming X., Han S., Fang G., Yingaridi B.

Zhou Y., Qiu Y., Li Z., Miao Z., Li C., Liu C., Tan Y.

The storage time of rice determines its quality and nutritional value, and the longer the storage time, the greater the impact. In this study, different excitation wavelengths (405 nm, 365 nm, 310 nm) were used to detect the fluorescence spectrum of “Dongdao 12” brown rice. Support vector machine (SVM), K-nearest neighbor (KNN), and wide neural network (WNN) were used for modeling and analysis. Under the excitation of 310 nm, the accuracy of WNN classification is up to 99.2%. In order to reduce the scattering effect and other interference in the data, multiplicative scatter correction (MSC), standard normal variable (SNV), and Savitzky–Goray smoothing (SG) preprocessing methods were used. The results showed that SG + KNN classification achieved an accuracy of 99.3% under 310 nm excitation. In order to further improve the classification accuracy, the original spectrum and the preprocessed spectrum under different excitation light sources were fused. The classification accuracy of all methods was improved, and the original data fusion was combined with the WNN model to reach 100%. It shows that fluorescence spectroscopy has excellent potential in identifying rice storage years.

Khan U.M., Sameen A., Decker E.A., Shabbir M.A., Hussain S., Latif A., Abdi G., Aadil R.M.

Plant extracts have demonstrated the ability to act as coagulants for milk coagulation at an adequate concentration, wide temperatures and pH ranges. This research is focused on the use of different vegetative extracts such as Citrus aurnatium flower extract (CAFE), bromelain, fig latex, and melon extract as economical and beneficial coagulants in the development of plant-based cheddar-type cheese. The cheddar-type cheese samples were subjected to physicochemical analysis in comparison to controlled cheese samples made from acetic acid and rennet. The fat, moisture, protein, and salt contents remained the same over the storage period, but a slight decline was observed in pH. The Ferric reducing antioxidant power (FRAP) increased with the passage of the ripening period. The FTIR and Raman spectra showed exponential changes and qualitative estimates in the binding and vibrational structure of lipids and protein in plant-based cheeses. The higher FTIR and Raman spectra bands were observed in acid, rennet, bromelain, and CAFE due to their firm and strong texture of cheese while lower spectra were observed in cheese made from melon extract due to weak curdling and textural properties. These plant extracts are economical and easily available alternative sources for cheese production with higher protein and nutritional contents.

Li C., Tan Y., Liu C., Guo W.

The origin of agricultural products is crucial to their quality and safety. This study explored the differences in chemical composition and structure of rice from different origins using fluorescence detection technology. These differences are mainly affected by climate, environment, geology and other factors. By identifying the fluorescence characteristic absorption peaks of the same rice seed varieties from different origins, and comparing them with known or standard samples, this study aims to authenticate rice, protect brands, and achieve traceability. The study selected the same variety of rice seed planted in different regions of Jilin Province in the same year as samples. Fluorescence spectroscopy was used to collect spectral data, which was preprocessed by normalization, smoothing, and wavelet transformation to remove noise, scattering, and burrs. The processed spectral data was used as input for the long short-term memory (LSTM) model. The study focused on the processing and analysis of rice spectra based on NZ-WT-processed data. To simplify the model, uninformative variable elimination (UVE) and successive projections algorithm (SPA) were used to screen the best wavelengths. These wavelengths were used as input for the support vector machine (SVM) prediction model to achieve efficient and accurate predictions. Within the fluorescence spectral range of 475–525 nm and 665–690 nm, absorption peaks of nicotinamide adenine dinucleotide (NADPH), riboflavin (B2), starch, and protein were observed. The origin tracing prediction model established using SVM exhibited stable performance with a classification accuracy of up to 99.5%.The experiment demonstrated that fluorescence spectroscopy technology has high discrimination accuracy in tracing the origin of rice, providing a new method for rapid identification of rice origin.

AVCU F.M.

Cai Y., Yao Z., Cheng X., He Y., Li S., Pan J.

Rapid identification of unknown material samples using portable or handheld Raman spectroscopy detection equipment is becoming a common analytical tool. However, the design and implementation of a set of Raman spectroscopy-based devices for substance identification must include spectral sampling of standard reference substance samples, resolution matching between different devices, and the training process of the corresponding classification models. The process of selecting a suitable classification model is frequently time-consuming, and when the number of classes of substances to be recognised increases dramatically, recognition accuracy decreases dramatically. In this paper, we propose a fast classification method for Raman spectra based on deep metric learning networks combined with the Gramian angular difference field (GADF) image generation approach. First, we uniformly convert Raman spectra acquired at different resolutions into GADF images of the same resolution, addressing spectral dimension disparities induced by resolution differences in different Raman spectroscopy detection devices. Second, a network capable of implementing nonlinear distance measurements between GADF images of different classes of substances is designed based on a deep metric learning approach. The Raman spectra of 450 different mineral classes obtained from the RRUFF database were converted into GADF images and used to train this deep metric learning network. Finally, the trained network can be installed on an embedded computing platform and used in conjunction with portable or handheld Raman spectroscopic detection sensors to perform material identification tasks at various scales. A series of experiments demonstrate that our trained deep metric learning network outperforms existing mainstream machine learning models on classification tasks of different sizes. For the two tasks of Raman spectral classification of natural minerals of 260 classes and Raman spectral classification of pathogenic bacteria of 8 classes with significant noise, our suggested model achieved 98.05% and 90.13% classification accuracy, respectively. Finally, we also deployed the model in a handheld Raman spectrometer and conducted identification experiments on 350 samples of chemical substances attributed to 32 classes, achieving a classification accuracy of 99.14%. These results demonstrate that our method can greatly improve the efficiency of developing Raman spectroscopy-based substance detection devices and can be widely used in tasks of unknown substance identification.

Kharbach M., Alaoui Mansouri M., Taabouz M., Yu H.

In today’s era of increased food consumption, consumers have become more demanding in terms of safety and the quality of products they consume. As a result, food authorities are closely monitoring the food industry to ensure that products meet the required standards of quality. The analysis of food properties encompasses various aspects, including chemical and physical descriptions, sensory assessments, authenticity, traceability, processing, crop production, storage conditions, and microbial and contaminant levels. Traditionally, the analysis of food properties has relied on conventional analytical techniques. However, these methods often involve destructive processes, which are laborious, time-consuming, expensive, and environmentally harmful. In contrast, advanced spectroscopic techniques offer a promising alternative. Spectroscopic methods such as hyperspectral and multispectral imaging, NMR, Raman, IR, UV, visible, fluorescence, and X-ray-based methods provide rapid, non-destructive, cost-effective, and environmentally friendly means of food analysis. Nevertheless, interpreting spectroscopy data, whether in the form of signals (fingerprints) or images, can be complex without the assistance of statistical and innovative chemometric approaches. These approaches involve various steps such as pre-processing, exploratory analysis, variable selection, regression, classification, and data integration. They are essential for extracting relevant information and effectively handling the complexity of spectroscopic data. This review aims to address, discuss, and examine recent studies on advanced spectroscopic techniques and chemometric tools in the context of food product applications and analysis trends. Furthermore, it focuses on the practical aspects of spectral data handling, model construction, data interpretation, and the general utilization of statistical and chemometric methods for both qualitative and quantitative analysis. By exploring the advancements in spectroscopic techniques and their integration with chemometric tools, this review provides valuable insights into the potential applications and future directions of these analytical approaches in the food industry. It emphasizes the importance of efficient data handling, model development, and practical implementation of statistical and chemometric methods in the field of food analysis.

Wei Y., Yang C., He L., Wu F., Yu Q., Hu W.

The classification of GM and non-GM maize kernels is fundamental for further analysis of the gene action in maize. Therefore, a complete and novel detection scheme based on near-infrared spectra was designed to distinguish GM and non-GM maize kernels. Hyperspectral images (935–1720 nm) of 777 maize kernels from 3 kinds were captured, and the average spectra of the maize kernels were extracted for modeling analysis. The classical modeling methods based on feature engineering were first studied, and the backpropagation neural network–genetic algorithm model showed the best performance with a prediction accuracy of 0.861. Then, novel modeling methods based on deep learning were developed. To dig out the interactive information between different bands and match the application scenarios, the original spectra were transformed into two-dimensional matrices before establishing the deep learning models. A modified convolution neural network (i.e., VGG net) with dilated convolution was finally constructed to classify the maize kernels, and the prediction accuracy reached 0.961. This research provides a referential and novel way to detect GM maize kernels. Future research will improve the detection scheme for monitoring unauthorized GM organisms by introducing the visualization technology of deep learning.

Liu Y., Zhao S., Gao X., Fu S., Chao Song, Dou Y., Shaozhong Song, Qi C., Lin J.

Combined laser-induced breakdown spectroscopy (LIBS) and hyperspectral imaging (HSI) with machine learning algorithms can be used to identify rice quality and the place of origin of rice production rapidly and accurately.

Avcu F.M.

Abstract Honey is a food item that people consume because of its taste and positive effects on health. The importance of honey is increasing day by day because of the difficulties in production, the threat of the bee population due to environmental conditions and climate changes, and the increasing population. In this work, data obtained from Fourier transform infrared (FTIR) spectra of honey samples were used for clustering of honey data. First of all, the number of clusters was determined by applying elbow method to the spectrum data obtained from the samples. After this process, the data was divided into 5 clusters. The data were reduced to 2 dimensions with principal components analysis (PCA), clusters of samples were determined by applying Hierarchical clustering (HCA). 20% of the data whose clusters were determined were randomly selected to be used as test data. The rest of the data was used as training data in Deep Learning. After the training, the test data was checked and the accuracy was found to be 96.15%. The proposed method gives reliable results in clustering of honey samples with the advantages of being fast, cheap and not requiring preprocess procedure.

Long Y., Huang W., Wang Q., Fan S., Tian X.

Maize mildew is a common phenomenon and it is essential to detect the mildew of a single maize kernel and prevent mildew from spreading around. In this study, a line-scanning Raman hyperspectral imaging system was applied to detect fungal spore quantity of a single maize kernel. Raman spectra were extracted while textural features were obtained to depict the maize mildew. Three kinds of modeling algorithms were used to establish the quantitative model to determine the fungal spore quantity of a single maize kernel. Then competitive adaptive reweighted sampling (CARS) was used to optimize characteristic variables. The optimal detection model was established with variables selected from the combination of Raman spectra and textural variance feature by PLSR. Results indicated that it was feasible to detect the fungal spore quantity of a single maize kernel by Raman hyperspectral technique. The study provided an in-situ and nondestructive alternative to detect fungal spore quantity.

Pezzotti G., Zhu W., Hashimoto Y., Marin E., Masumura T., Sato Y., Nakazaki T.

Raman spectroscopy was applied to characterize at the molecular scale the nutritional quality of the Japanese Koshihikari rice cultivar in comparison with other renowned rice cultivars including Carnaroli from Italy, Calrose from the USA, Jasmine rice from Thailand, and Basmati from both India and Pakistan. For comparison, two glutinous (mochigome) cultivars were also investigated. Calibrated and validated Raman analytical algorithms allowed quantitative determinations of: (i) amylopectin and amylose concentrations, (ii) fractions of aromatic amino acids, and (iii) protein content and secondary structure. The Raman assessments non-destructively linked the molecular composition of grains to key nutritional parameters and revealed a complex intertwine of chemical properties. The Koshihikari cultivar was rich in proteins (but with low statistical relevance as compared to other investigated cultivars) and aromatic amino acids. However, it also induced a clearly higher glycemic impact as compared to long-grain cultivars from Asian countries. Complementary to genomics and wet-chemistry analyses, Raman spectroscopy makes non-destructively available factual and data-driven information on rice nutritional characteristics, thus providing customers, dietitian nutritionists, and producers with a solid science-consolidated platform.