AI Tools in Food Safety and Quality Analysis: Comparison
View Latest Version
Please note this is a comparison between Version 2 by Wendy Huang and Version 1 by Hind Raki.

On a global scale, food safety and security aspects entail consideration throughout the farm-to-fork continuum, considering food’s supply chain. Generally, the agrifood system is a multiplex network of interconnected features and processes, with a hard predictive rate, where maintaining the food’s safety is an indispensable element and is part of the Sustainable Development Goals (SDGs). It has led the scientific community to develop advanced applied analytical methods, such as machine learning (ML) and deep learning (DL) techniques applied for assessing foodborne diseases. 

  • spectroscopy
  • food safety
  • quality
  • AI tools
  • machine learning
  • deep learning

1. Introduction

One Health is a very distinct concept unifying soil, plant, and human health for a flourishing and sustainable ecosystem. This approach can also be applied on ensuring food safety and security. Ending hunger and ensuring access to safe, nutritious, and sufficient food all year round by all people are part of the Sustainable Development Goals (SDGs) to be achieved by 2030, according to the United Nations SDGs agenda, specifically SDG3 and SDG2, respectively (Good Health and Well-being, Zero Hunger). Nonetheless, multiple challenges and issues are increasingly making this mission impossible to accomplish [1]. As a matter of fact and according to statistical studies, humanity is expecting an enormous increase in the global population, reaching 9.7 billion individuals by 2025. Surely, ensuring that they all have access to safe and nutritious food becomes more challenging. The increasing demand for crop-based food in the market is due to multiple factors, mainly the propaganda of plant-based diets [2], the transition of human diets in society, and the factor of poor diversity in crop-food consumption while thousands of crops exist. These factors have impacted diversity concerning crops, and lately, scientists have been shedding light on orphan crops, which are crops not traded internationally, but would have been important for regional food security [3], hence ourthe objective to target food of crop origins.
Concretely, it has always been a big challenge; the food system dates back several decades, with a tendency to evolve continuously, depending on its economic, social, cultural, and environmental factors, along with several external and internal variables [4]. These complex interdependent and interconnected factors impact the global agrifood system, encompassing the farm-to-fork continuum and various environmental and socioeconomic factors [5]. In this dynamic context, the integration of crop classification using deep learning emerges as a significant consideration, providing advanced tools for precise and efficient analysis within the intricate web of factors influencing the agrifood system, supported by numerous studies dedicated to the development of advanced tools [6,7][6][7]. Hence, food security, quality, and safety, now more than ever, are depending on the entire food supply chain from early production to market accessibility. A food-borne disease is referred to as food contamination, on account of the presence of hazardous contaminants that can cause human body illnesses. They are divided into three groups, namely biological, chemical, and physical contaminants, based on the pollutant and the process by which they enter the food product [8], for example, during crop cultivation, due to contaminated soil (animal manure or chemical fertilizers), as well as the water used for irrigation of produce (groundwater, recovered rainfall, surface water, or re-utilized wastewater) [9,10][9][10]. Food contamination represents a big challenge because of its large impact, being related to the whole food supply chain. In order to protect consumers from unsafe foods, standards are required to establish a monitoring system to reduce chemical and microbiological food contamination [11], including all food chain participants, such as farmers, processors, transporters, retailers, and consumers [4]. Even though there are many guidelines to follow, only a fraction of them is followed. In many cases, individuals neglect simple practices, such as proper hand-washing methods and the good use of gloves, which leads to serious food poisoning [12].
It was said that the transmission of viruses through food matrices is less likely compared to direct person-to-person contact, respiratory droplets or contaminated surfaces. The COVID-19 pandemic has forced us to think about the technological preparedness to use these smart technologies, which avoid human-to-human and human-to-food contact during food processing. Hence, there is a need for technological innovation combining analytical strategies with AI tools, taking into consideration economic and feasibility challenges [13]. Artificial intelligence (AI) is primordial for technological advances, developing computational tools to provide sustainable solutions for food security and safety [14]. There are potential applications of machine vision systems in addition to analytical strategies for more accurate and lower-cost techniques for contaminant detection in food [15].
Spectroscopic measurement methods are versatile tools applied across diverse scientific domains. They play crucial roles in pharmaceutical quality control, environmental monitoring, materials science, medical diagnostics, forensic analysis, agricultural research and much more. The adaptability and precision of these methods make them invaluable for characterizing composition, structure, and properties in various scientific disciplines. Recent advances have emerged in chemical analytical strategies, depending on the variety of studied food matrices. There is a constant quest to find the most suitable technique to investigate certain aspects of a compound and ascertain its consistency or structure.

2. Spectroscopy for Food Safety and Quality

Spectroscopic methods serve as analytical tools to identify food’s composition, germs, pests, diseases, and adulteration [16]. The following listing presents the most common techniques within the literature: 1. infrared spectroscopy, 2. Raman spectroscopy, 3. nuclear magnetic resonance (NMR) spectroscopy,4. ultraviolet–visible spectroscopy (UV-vis). Infrared spectroscopy using Fourier transform infrared (FTIR) is often employed. The mid-infrared (MIR) region covers an area between 4000 and 400 cm−1−1 [17]. A soil’s composition, characteristics, and organic matter may all be found via MIR spectroscopy. Also, the diffuse reflectance infrared Fourier transform (DRIFT) method may identify the chemical characteristics of humus and soil. Moreover, attenuated total reflectance (ATR) can identify organic materials in soil [18]. Anisidine, which is produced during the oxidation of food, is frequently measured using UV-visible spectroscopy to assess the quality of oil [19]. Moreover, fluorescence is a property shared by a large number of microorganisms, including their colonies, making it simple to identify any bacteria by looking at their fluorescence spectra [20,21][20][21]. Tryptophan, riboflavin, and lumichrome are three different forms of fluorophores found in yogurt, and their presence enables fluorescence spectroscopy to assess the yogurt’s quality [22]. Honey is a substance produced by bees from floral nectar, containing phenolic chemicals that are byproducts of the phenolic acid found in the flower, and as it is packaged and transported, its qualities alter [23]. Fluorescence spectroscopy can quantify the concentration of phenolic chemicals. During processing and storage, mycotoxins and fungus are found in grains using MIR spectroscopy [24]. On the other hand, carbohydrates’ structure can change during storage, especially in the presence of water, which can be identified using Raman spectroscopy. This technique characterizes and quantifies the lipid content of foods [25]. NMR spectroscopy can monitor the ripening, drying, and adulteration of food components as well as determining the genotype responsible for a certain phenotype of the grapes used to create wine. As a result, NMR spectroscopy may offer information on mixtures of metabolites [26]. Infrared or Raman fingerprints are the outcomes of observations made on a large number of objects or samples with a wide variety of characteristics (variables, like the absorbance at various wave-numbers or wave-number shifts) from a monochromatic light source for FTIR and Raman, respectively [27]. Therefore, since the exponential rise in computing power and the capacity to gather, store, and analyze enormous volumes of data, machine learning (ML) systems can improve the potential to extrapolate information from complicated spectrum data.

3. Integrating AI Tools for Food Safety and Quality Analysis

The development of analytical strategies for food and beverage assessments is crucial for ensuring food safety and public health. Various technologies, including imaging, odor, and taste, have been developed [28]. Spectroscopic techniques, such as infrared spectroscopy and Raman spectroscopy, have proven to be rapid and nondestructive for microorganism detection [29]. Magnetic surface-enhanced Raman scattering nanoprobes show high specificity in separating and detecting multiple pathogens in complex food matrices [30]. The usage of electronic noses, coupled with data acquisition cards and classification methods, enhances success rates in ensuring the originality of saffron products [28]. The integration of more effective algorithms is essential for spectral data processing and microorganism reference database building [29]. Hyperspectral imaging (HSI) systems, including computer vision systems (CVS), are widely applied in the food industry for nonintrusive quality control [28]. HSI covers various food-processing phases, providing the ability to control the quality and safety of processed foods [31]. Challenges lie in the classification, especially for crops within the same family, and the high cost of these technologies [31]. Algorithms and chemometric methods should focus on reducing the dimensionality of data and improving computational efficiency while enhancing performance and robustness [31]. Terahertz spectroscopy techniques, coupled with machine learning tools, ensure quality and security inspection of agricultural products and food [32]. Classical methods of spectral preprocessing, such as smoothing, standard normal variate, and Fourier transformation, can be integrated into multivariate calibration steps for more efficiency [33]. In the quantification of honey adulteration, spectroscopy and hyperspectral imaging, when coupled with machine learning models and optic fiber sensors, provide fast and nondestructive detection [34]. AI development in data mining has made significant breakthroughs, particularly with the application of deep learning (DL) in the analysis of spectral data from food and agricultural products [35]. DL approaches offer a less laborious yet more precise method for this purpose. Combining infrared spectroscopy (IRS) and hyperspectral imaging (HSI) techniques with AI tools shows potential in advancing the quality evaluation of cereals, which are among the top consumed crops globally [36]. The integration of convolutional neural networks (CNNs) in the qualitative and quantitative analysis of spectra involves extracting micro- and macro-features through multiple convolution and pooling layers. DL-spectroscopic sensing techniques have demonstrated promising results in the quality evaluation of food and agro-products, encompassing identification, geographical origin detection, adulteration recognition, bruise detection, and component content prediction for crops [35]. The application of CNNs helps avoid secondary workloads, although challenges persist, such as determining optimal network scale, selecting parameters, addressing overfitting, and enhancing model interpretability—a current dilemma in AI research [33]. Table 1 highlights recent and high-quality papers that couple analytical strategies with machine learning approaches across diverse food safety and quality purposes.
Table 1.
Latest applications combining AI tools with analytical approaches for food safety and quality.
Analytical Approach AI Tool Problematic Ref.
Spectroscopy Python-based portable system using Jetson TX2 Module Food classification of four classes of coffee and purées [37]
Near-infrared spectroscopy Block sparse Bayesian learning (BSBL) with fast marginalized likelihood maximization (FMLM) Computational cost reduction for calculating the inverse of a large matrix containing absorption peak information [38]
Impedance spectroscopy A fuzzy logic model applied on the parameters extracted from distribution of relaxation times (DRT) Meat-based food classification according to its freshness for different types of muscles [39]
TeraHertz (THz) spectroscopy and chemometrics Interval partial least squares (iPLS) for optimizing the THz frequency and other preprocessing techniques combined with extreme learning machine (ELM), genetic algorithm support vector machine (GA-SVM), and artificial bee colony algorithm support vector machine (ABC-SVM) for decision making Three typical soybean origins’ identification [40]
Fourier transform infrared (FTIR) spectroscopy FTIR and multispectral imaging (MSI) coupled with support vector machines (SVM) for regression Meat quality assessment, specifically minced pork patties stored under modified atmosphere packaging (MAP) conditions, by estimating the microbial population [41]
Raman spectroscopy A single convolutional neural network (CNN) model development where hyperparameters, activity functions, and loss functions were optimized Spectral data preprocessing simplification [42]
Dielectric spectroscopy Principal component analysis (PCA) for preprocessing and four models, namely support vector machine—SVM, K-nearest neighbor—KNN, linear discriminant—LD and quadratic discriminant—QD, for classification purposes Discrimination between three citrus juices in order to develop new technologies to identify adulteration [43]

References

  1. Garcia, S.N.; Osburn, B.I.; Jay-Russell, M.T. One health for food safety, food security, and sustainable food production. Front. Sustain. Food Syst. 2020, 4, 1.
  2. Hargreaves, S.M.; Rosenfeld, D.L.; Moreira, A.V.B.; Zandonadi, R.P. Plant-based and vegetarian diets: An overview and definition of these dietary patterns. Eur. J. Nutr. 2023, 62, 1109–1121.
  3. Matthews, P.J.; Ghanem, M.E. Perception gaps that may explain the status of taro (Colocasia esculenta) as an “orphan crop”. Plants People Planet 2021, 3, 99–112.
  4. Neufeld, L.M.; Hendriks, S.; Hugas, M. Healthy Diet: A Definition for the United Nations Food Systems Summit 2021. In A Paper from the Scientific Group of the UN Food Systems Summit; Springer International Publishing: Cham, Switzerland, 2021.
  5. FAO. Thinking about the Future of Food Safety—A Foresight Report; FAO: Rome, Italy, 2022.
  6. Mohanty, S.P.; Hughes, D.P.; Salathé, M. Using deep learning for image-based plant disease detection. Front. Plant Sci. 2016, 7, 1419.
  7. Raki, H.; González-Vergara, J.; Aalaila, Y.; Elhamdi, M.; Bamansour, S.; Guachi-Guachi, L.; Peluffo-Ordoñez, D.H. Crop Classification Using Deep Learning: A Quick Comparative Study of Modern Approaches. In Proceedings of the International Conference on Applied Informatics, Serdang, Malaysia, 18–19 May 2022; Springer: Cham, Switzerland, 2022; pp. 31–44.
  8. Kamala, K.; Kumar, V.P. Food products and food contamination. In Microbial Contamination and Food Degradation; Elsevier: Amsterdam, The Netherlands, 2018; pp. 1–19.
  9. Bintsis, T. Microbial pollution and food safety. AIMS Microbiol. 2018, 4, 377.
  10. Turek, C.; Stintzing, F.C. Stability of essential oils: A review. Compr. Rev. Food Sci. Food Saf. 2013, 12, 40–53.
  11. Clapp, J.; Moseley, W.G.; Burlingame, B.; Termine, P. The case for a six-dimensional food security framework. Food Policy 2021, 106, 102164.
  12. Chatterjee, A.; Abraham, J. Microbial contamination, prevention, and early detection in food industry. In Microbial Contamination and Food Degradation; Elsevier: Amsterdam, The Netherlands, 2018; pp. 21–47.
  13. Chitrakar, B.; Zhang, M.; Bhandari, B. Improvement strategies of food supply chain through novel food processing technologies during COVID-19 pandemic. Food Control 2021, 125, 108010.
  14. Surampalli, R.Y.; Zhang, T.C.; Goyal, M.K.; Brar, S.K.; Tyagi, R. Sustainability: Fundamentals and Applications; John Wiley & Sons: Hoboken, NJ, USA, 2020.
  15. Teena, M.; Manickavasagan, A.; Mothershaw, A.; El Hadi, S.; Jayas, D. Potential of machine vision techniques for detecting fecal and microbial contamination of food products: A review. Food Bioprocess Technol. 2013, 6, 1621–1634.
  16. Craig, A.P.; Franca, A.S.; Irudayaraj, J. Surface-enhanced Raman spectroscopy applied to food safety. Annu. Rev. Food Sci. Technol. 2013, 4, 369–380.
  17. Brousseau, J.L.; Vidon, S.; Leblanc, R. Investigation of the chemical nature of two-dimensional polymerized octadecyltrimethoxysilane Langmuir films by inelastic electron tunneling spectroscopy. J. Chem. Phys. 1998, 108, 7391–7396.
  18. Ojanen, P.; Mäkiranta, P.; Penttilä, T.; Minkkinen, K. Do logging residue piles trigger extra decomposition of soil organic matter? For. Ecol. Manag. 2017, 405, 367–380.
  19. Świeca, M.; Sęczyk, Ł.; Gawlik-Dziki, U.; Dziki, D. Bread enriched with quinoa leaves—The influence of protein–phenolics interactions on the nutritional and antioxidant quality. Food Chem. 2014, 162, 54–62.
  20. Aghayee, S.; Benadiba, C.; Notz, J.; Kasas, S.; Dietler, G.; Longo, G. Combination of fluorescence microscopy and nanomotion detection to characterize bacteria. J. Mol. Recognit. 2013, 26, 590–595.
  21. Guo, R.; McGoverin, C.; Swift, S.; Vanholsbeeck, F. A rapid and low-cost estimation of bacteria counts in solution using fluorescence spectroscopy. Anal. Bioanal. Chem. 2017, 409, 3959–3967.
  22. Christensen, J.; Becker, E.M.; Frederiksen, C. Fluorescence spectroscopy and PARAFAC in the analysis of yogurt. Chemom. Intell. Lab. Syst. 2005, 75, 201–208.
  23. Gismondi, A.; De Rossi, S.; Canuti, L.; Novelli, S.; Di Marco, G.; Fattorini, L.; Canini, A. From Robinia pseudoacacia L. nectar to Acacia monofloral honey: Biochemical changes and variation of biological properties. J. Sci. Food Agric. 2018, 98, 4312–4322.
  24. Su, W.H.; He, H.J.; Sun, D.W. Non-destructive and rapid evaluation of staple foods quality by using spectroscopic techniques: A review. Crit. Rev. Food Sci. Nutr. 2017, 57, 1039–1051.
  25. Grundas, S.; Stępniewski, A. Advances in Agrophysical Research; BoD–Books on Demand: Norderstedt, Germany, 2013.
  26. Kim, H.K.; Choi, Y.H.; Verpoorte, R. NMR-based metabolomic analysis of plants. Nat. Protoc. 2010, 5, 536–549.
  27. Lindgren, F.; Rännar, S. Alternative partial least-squares (PLS) algorithms. In 3D QSAR in Drug Design: Recent Advances; Springer: Dordrecht, The Netherlands, 1998; pp. 105–113.
  28. Tan, W.K.; Husin, Z.; Yasruddin, M.L.; Ismail, M.A.H. Recent technology for food and beverage quality assessment: A review. J. Food Sci. Technol. 2022, 60, 1681–1694.
  29. Wang, K.; Pu, H.; Sun, D.W. Emerging spectroscopic and spectral imaging techniques for the rapid detection of microorganisms: An overview. Compr. Rev. Food Sci. Food Saf. 2018, 17, 256–273.
  30. Wang, Y.; Ravindranath, S.; Irudayaraj, J. Separation and detection of multiple pathogens in a food matrix by magnetic SERS nanoprobes. Anal. Bioanal. Chem. 2011, 399, 1271–1278.
  31. Liu, Y.; Pu, H.; Sun, D.W. Hyperspectral imaging technique for evaluating food quality and safety during various processes: A review of recent applications. Trends Food Sci. Technol. 2017, 69, 25–35.
  32. Jiang, Y.; Li, G.; Ge, H.; Wang, F.; Li, L.; Chen, X.; Lu, M.; Zhang, Y. Machine learning and application in terahertz technology: A review on achievements and future challenges. IEEE Access 2022, 10, 53761–53776.
  33. Wang, H.P.; Chen, P.; Dai, J.W.; Liu, D.; Li, J.Y.; Xu, Y.P.; Chu, X.L. Recent advances of chemometric calibration methods in modern spectroscopy: Algorithms, strategy, and related issues. TrAC Trends Anal. Chem. 2022, 153, 116648.
  34. Al-Awadhi, M.A.; Deshmukh, R.R. A Review on Modern Analytical Methods for Detecting and Quantifying Adulteration in Honey. In Proceedings of the 2021 International Conference of Modern Trends in Information and Communication Technology Industry (MTICTI), Sana’a, Yemen, 4–6 December 2021; IEEE: Piscataway, NJ, USA, 2021; pp. 1–6.
  35. Zhang, X.; Yang, J.; Lin, T.; Ying, Y. Food and agro-product quality evaluation based on spectroscopy and deep learning: A review. Trends Food Sci. Technol. 2021, 112, 431–441.
  36. An, D.; Zhang, L.; Liu, Z.; Liu, J.; Wei, Y. Advances in infrared spectroscopy and hyperspectral imaging combined with artificial intelligence for the detection of cereals quality. Crit. Rev. Food Sci. Nutr. 2022, 63, 9766–9796.
  37. Buzura, L.; Budileanu, M.L.; Potarniche, A.; Galatus, R. Python based portable system for fast characterisation of foods based on spectral analysis. In Proceedings of the 2021 IEEE 27th International Symposium for Design and Technology in Electronic Packaging (SIITME), Timisoara, Romania, 27–30 October 2021; IEEE: Piscataway, NJ, USA, 2021; pp. 275–280.
  38. Pan, T.; Wu, C.; Chen, Q. Sparse Reconstruction Using Block Sparse Bayesian Learning With Fast Marginalized Likelihood Maximization for Near-Infrared Spectroscopy. IEEE Trans. Instrum. Meas. 2021, 71, 2500410.
  39. Nouri, H.; Guermazi, M.; Kallel, A.Y.; Hao, W.; Kanoun, O. Meat Freshness Assessment based on Impedance Spectroscopy and Distribution of Relaxation Times (DRT). In Proceedings of the 2022 International Workshop on Impedance Spectroscopy (IWIS), Chemnitz, Germany, 27–30 September 2022; IEEE: Piscataway, NJ, USA, 2022; pp. 41–45.
  40. Wei, X.; Zhu, S.; Zhou, S.; Zheng, W.; Li, S. Identification of soybean origin by terahertz spectroscopy and chemometrics. IEEE Access 2020, 8, 184988–184996.
  41. Fengou, L.C.; Mporas, I.; Spyrelli, E.; Lianou, A.; Nychas, G.J. Estimation of the microbiological quality of meat using rapid and non-invasive spectroscopic sensors. IEEE Access 2020, 8, 106614–106628.
  42. Shen, J.; Li, M.; Li, Z.; Zhang, Z.; Zhang, X. Single convolutional neural network model for multiple preprocessing of Raman spectra. Vib. Spectrosc. 2022, 121, 103391.
  43. Trigoso, T.C.; Cruz, J.O.; Miñano, H.A.; Silupu, W.C. Application of machine learning in the discrimination of citrus fruit juices: Uses of dielectric spectroscopy. In Proceedings of the 2020 IEEE Engineering International Research Conference (EIRCON), Lima, Peru, 21–23 October 2020; IEEE: Piscataway, NJ, USA, 2020; pp. 1–4.
More
© Text is available under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/)
Video Production Service
Feedback