Imaging Spectroscopy and Machine Learning for Intelligent Determination of Potato and Sweet Potato Quality

Recent advances on artificial intelligence-based approaches for food adulteration and fraud detection in the food industry: Challenges and opportunities

Food adulteration is the deceitful practice of misleading consumers about food to profit from it. The threat to public health and food quality or nutritional valuable make it a major issue. Food origin and adulteration should be considered to safeguard customers against fraud. It has been established that artificial intelligence is a cutting-edge technology in food science and engineering. In this study, it has been explained how AI detects food tampering. Applications of AI such as machine learning tools in food quality have been studied. This review covered several food quality detection web-based information sources. The methods used to detect food adulteration and food quality standards have been highlighted. Various comparisons between state-of-the-art techniques, datasets, and outcomes have been conducted. The outcomes of this investigation will assist researchers choose the best food quality method. It will help them identify of foods that have been explored by researchers and potential research avenues.

Automation and Optimization of Food Process Using CNN and Six-Axis Robotic Arm

The Food Process Robot Intelligent System (FPRIS) integrates a 3D-printed six-axis robotic arm with Artificial Intelligence (AI) and Computer Vision (CV) to optimize and automate the coffee roasting process. As an application of FPRIS coffee roasting, this system uses a Convolutional Neural Network (CNN) to classify coffee beans inside the roaster and control the roaster in real time, avoiding obstacles and empty spaces. This study demonstrates FPRIS's capability to precisely control the Degree of Roasting (DoR) by combining gas and image sensor data to assess coffee bean quality. A comparative analysis between the Preliminary Coffee Sample (PCS) and Validation Coffee Sample (VCS) revealed that increasing roast intensity resulted in consistent trends for both samples, including an increase in weight loss and Gas sensor Initial Difference (GID) and a decrease in Sum of Pixel Grayscale Values (SPGVs). This study underscores the potential of FPRIS to enhance precision and efficiency in coffee roasting. Future studies will expand on these findings by testing FPRIS across various food processes, potentially establishing a universal automation system for the food industry.

Modeling of moisture content during baking with different approaches for effective diffusivity and evaluation of quality variables in baked potato slices

Baking is a healthier alternative to frying, since texture, color, smell, and flavor are developed, without adding oil. The objective was to estimate the moisture content in potato slices, during baking using Fick's law of diffusion to model internal moisture transport and to assess the impact on quality attributes. Moisture transport kinetics were examined at three baking temperatures of 120, 130, and 140°C. Fick's law was employed to estimate average moisture content using different methods: considering both a constant (method of slopes by subperiods, MSS; and method of successive approximations, MSA) and a variable (represented as a quadratic function of time, QFT) behavior of effective diffusivity (De). Three quality variables were analyzed: water activity (aw, dew point hygrometry), total color difference (∆E, colorimetry), and fracturability (F, universal testing machine). The diffusivity estimated with the time-varying De method provided a more realistic description of moisture migration during baking. The aw, ∆E, and F for baked potato slices ranged from 0.234 to 0.276, 17.9 to 24.6, and 5.20 to 5.49 N, respectively. These attributes imply improved stability and extended shelf life, showing typical colors and texture changes for baked snacks. These changes are linked to variations in diffusivity, influenced by the size and quantity of micropores within the food structure. This study could allow an accurate prediction of mass transfer by considering variable De, facilitating the optimization of baking conditions. PRACTICAL APPLICATION: The analysis of the moisture content using Fick's law, considering a time-varying diffusivity, enables the optimization of the baking process for foods. This helps minimize the occurrence of defective products.

Frontiers of machine learning in smart food safety

Integration of machine learning (ML) technologies into the realm of smart food safety represents a rapidly evolving field with significant potential to transform the management and assurance of food quality and safety. This chapter will discuss the capabilities of ML across different segments of the food supply chain, encompassing pre-harvest agricultural activities to post-harvest processes and delivery to the consumers. Three specific examples of applying cutting-edge ML to advance food science are detailed in this chapter, including its use to improve beer flavor, using natural language processing to predict food safety incidents, and leveraging social media to detect foodborne disease outbreaks. Despite advances in both theory and practice, application of ML to smart food safety still suffers from issues such as data availability, model reliability, and transparency. Solving these problems can help realize the full potential of ML in food safety. Development of ML in smart food safety is also driven by social and industry impacts. The improvement and implementation of legal policies brings both opportunities and challenges. The future of smart food safety lies in the strategic implementation of ML technologies, navigating social and industry impacts, and adapting to regulatory changes in the AI era.

Machine Learning Application in Horticulture and Prospects for Predicting Fresh Produce Losses and Waste: A Review

The current review examines the state of knowledge and research on machine learning (ML) applications in horticultural production and the potential for predicting fresh produce losses and waste. Recently, ML has been increasingly applied in horticulture for efficient and accurate operations. Given the health benefits of fresh produce and the need for food and nutrition security, efficient horticultural production and postharvest management are important. This review aims to assess the application of ML in preharvest and postharvest horticulture and the potential of ML in reducing postharvest losses and waste by predicting their magnitude, which is crucial for management practices and policymaking in loss and waste reduction. The review starts by assessing the application of ML in preharvest horticulture. It then presents the application of ML in postharvest handling and processing, and lastly, the prospects for its application in postharvest loss and waste quantification. The findings revealed that several ML algorithms perform satisfactorily in classification and prediction tasks. Based on that, there is a need to further investigate the suitability of more models or a combination of models with a higher potential for classification and prediction. Overall, the review suggested possible future directions for research related to the application of ML in postharvest losses and waste quantification.

Effect of potassium fertilization on storage root number, yield, and appearance quality of sweet potato (Ipomoea batatas L.)

Increasing storage root number is a pivotal approach to enhance both storage root (SR) yield and appearance quality of sweet potato. Here, 2-year field experiments were conducted to investigate the effect of 0 (K0), 120 (K1), 240 (K2), and 360 (K3) kg ha-1 potassium fertilizer (K2O) on lignin metabolism, root growth, storage root yield, and uniformity. The results demonstrated that potassium (K) application led to a decrease in the activities of key enzymes involved in lignin biosynthesis, including phenylalanine deaminase (PAL), 4-coumarate coenzyme A ligase (4-CL), cinnamic acid dehydrogenase (CAD), polyphenol oxidase (PPO), and peroxidase (POD). This resulted in a significant reduction in lignin and G-type lignin contents in potential SRs compared to K0 treatment within 10-30 days after planting (DAP). BJ553 exhibited a significant decrease in PAL activity, as well as lignin and G-type contents at 10 DAP, whereas YS25 showed delayed effects until 20 DAP. However, the number and distribution of secondary xylem conduits as well as the mid-column diameter area in roots were increased in K2 treatment. Interestingly, K2 treatment exhibited significantly larger potential SR diameter than other treatments at 15, 20, and 25 DAP. At harvest, K2 treatment increased the SR number, the single SR weight, and overall yield greatly compared with K0 treatment, with an average increase of 19.12%, 16.54%, and 16.92% respectively. The increase of SR number in BJ553 was higher than that of YS25. Furthermore, K2 treatment exhibited the lowest coefficient of variation for both SR length and diameter, indicating a higher yield of middle-sized SRs. In general, appropriate potassium application could effectively suppress lignin biosynthesis, leading to a reduction in the degree of pericycle lignification in potential SRs. This promotes an increase in the number of storage roots and ultimately enhances both yield and appearance quality of sweet potato. The effect of potassium fertilizer on lignin metabolism in BJ553 roots was earlier and resulted in a greater increase in the SR number compared to YS25.

ScabyNet, a user-friendly application for detecting common scab in potato tubers using deep learning and morphological traits

Common scab (CS) is a major bacterial disease causing lesions on potato tubers, degrading their appearance and reducing their market value. To accurately grade scab-infected potato tubers, this study introduces "ScabyNet", an image processing approach combining color-morphology analysis with deep learning techniques. ScabyNet estimates tuber quality traits and accurately detects and quantifies CS severity levels from color images. It is presented as a standalone application with a graphical user interface comprising two main modules. One module identifies and separates tubers on images and estimates quality-related morphological features. In addition, it enables the extraction of tubers as standard tiles for the deep-learning module. The deep-learning module detects and quantifies the scab infection into five severity classes related to the relative infected area. The analysis was performed on a dataset of 7154 images of individual tiles collected from field and glasshouse experiments. Combining the two modules yields essential parameters for quality and disease inspection. The first module simplifies imaging by replacing the region proposal step of instance segmentation networks. Furthermore, the approach is an operational tool for an affordable phenotyping system that selects scab-resistant genotypes while maintaining their market standards.

Using machine learning for image-based analysis of sweetpotato root sensory attributes

The sweetpotato breeding process involves assessing different phenotypic traits, such as the sensory attributes, to decide which varieties to progress to the next stage during the breeding cycle. Sensory attributes like appearance, taste, colour and mealiness are important for consumer acceptability and adoption of new varieties. Therefore, measuring these sensory attributes is critical to inform the selection of varieties during breeding. Current methods using a trained human panel enable screening of different sweetpotato sensory attributes. Despite this, such methods are costly and time-consuming, leading to low throughput, which remains the biggest challenge for breeders. In this paper, we describe an approach to apply machine learning techniques with image-based analysis to predict flesh-colour and mealiness sweetpotato sensory attributes. The developed models can be used as high-throughput methods to augment existing approaches for the evaluation of flesh-colour and mealiness for different sweetpotato varieties. The work involved capturing images of boiled sweetpotato cross-sections using the DigiEye imaging system, data pre-processing for background elimination and feature extraction to develop machine learning models to predict the flesh-colour and mealiness sensory attributes of different sweetpotato varieties. For flesh-colour the trained Linear Regression and Random Forest Regression models attained R 2 values of 0.92 and 0.87, respectively, against the ground truth values given by a human sensory panel. In contrast, the Random Forest Regressor and Gradient Boosting model attained R 2 values of 0.85 and 0.80, respectively, for the prediction of mealiness. The performance of the models matched the desirable R 2 threshold of 0.80 for acceptable comparability to the human sensory panel showing that this approach can be used for the prediction of these attributes with high accuracy. The machine learning models were deployed and tested by the sweetpotato breeding team at the International Potato Center in Uganda. This solution can automate and increase throughput for analysing flesh-colour and mealiness sweetpotato sensory attributes. Using machine learning tools for analysis can inform and quicken the selection of promising varieties that can be progressed for participatory evaluation during breeding cycles and potentially lead to increased chances of adoption of the varieties by consumers.

Assessing and Quantifying the Surface Texture of Milk Powder Using Image Processing

Milk powders produced from similar spray dryers have different visual appearances, while the surface appearance of the powder is a key quality attribute because the smoothness of the milk powder also affects flowability and handling properties. Traditionally quantifying this nuanced visual metric was undertaken using sensory panelists, which is both subjective and time consuming. Therefore, it is advantageous to develop an on-line quick and robust appearance assessment tool. The aim of this work is to develop a classification model which can classify the milk powder samples into different surface smoothness groups. This work proposes a strategy for quantifying the relative roughness of commercial milk powder from 3D images. Photogrammetry equipment together with the software RealityCapture were used to build 3D models of milk powder samples, and a surface normal analysis which compares the area of the triangle formed by the 3 adjacent surface normals or compares the angle between the adjacent surface normals was used to quantify the surface smoothness of the milk powder samples. It was found that the area of the triangle of the smooth-surface milk powder cone is smaller than the area of the triangle of the rough-surface milk powder cone, and the angle between the adjacent surface normals of the rough-surface milk powder cone is larger than the angle between the adjacent surface normals of the smooth-surface milk powder cone, which proved that the proposed area metrics and angle metrics can be used as tools to quantify the smoothness of milk powder samples. Finally, the result of the support vector machine (SVM) classifier proved that image processing can be used as a preliminary tool for classifying milk powder into different surface texture groups.

Live-cell fluorescence spectral imaging as a data science challenge

Live-cell fluorescence spectral imaging is an evolving modality of microscopy that uses specific properties of fluorophores, such as excitation or emission spectra, to detect multiple molecules and structures in intact cells. The main challenge of analyzing live-cell fluorescence spectral imaging data is the precise quantification of fluorescent molecules despite the weak signals and high noise found when imaging living cells under non-phototoxic conditions. Beyond the optimization of fluorophores and microscopy setups, quantifying multiple fluorophores requires algorithms that separate or unmix the contributions of the numerous fluorescent signals recorded at the single pixel level. This review aims to provide both the experimental scientist and the data analyst with a straightforward description of the evolution of spectral unmixing algorithms for fluorescence live-cell imaging. We show how the initial systems of linear equations used to determine the concentration of fluorophores in a pixel progressively evolved into matrix factorization, clustering, and deep learning approaches. We outline potential future trends on combining fluorescence spectral imaging with label-free detection methods, fluorescence lifetime imaging, and deep learning image analysis.

A Rapid Non-Destructive Hyperspectral Imaging Data Model for the Prediction of Pungent Constituents in Dried Ginger

Ginger is best known for its aromatic odour, spicy flavour and health-benefiting properties. Its flavour is derived primarily from two compound classes (gingerols and shogaols), with the overall quality of the product depending on the interaction between these compounds. Consequently, a robust method for determining the ratio of these compounds would be beneficial for quality control purposes. This study investigated the feasibility of using hyperspectral imaging to rapidly determine the ratio of 6-gingerol to 6-shogoal in dried ginger powder. Furthermore, the performance of several pre-processing methods and two multivariate models was explored. The best-performing models used partial least squares regression (PSLR) and least absolute shrinkage and selection operator (LASSO), using multiplicative scatter correction (MSC) and second derivative Savitzky–Golay (2D-SG) pre-processing. Using the full range of wavelengths (~400–1000 nm), the performance was similar for PLSR (R² ≥ 0.73, RMSE ≤ 0.29, and RPD ≥ 1.92) and LASSO models (R² ≥ 0.73, RMSE ≤ 0.29, and RPD ≥ 1.94). These results suggest that hyperspectral imaging combined with chemometric modelling may potentially be used as a rapid, non-destructive method for the prediction of gingerol-to-shogaol ratios in powdered ginger samples.

Applications of Fluorescence Spectroscopy, RGB- and MultiSpectral Imaging for Quality Determinations of White Meat: A Review

Fluorescence spectroscopy, color imaging and multispectral imaging (MSI) have emerged as effective analytical methods for the non-destructive detection of quality attributes of various white meat products such as fish, shrimp, chicken, duck and goose. Based on machine learning and convolutional neural network, these techniques can not only be used to determine the freshness and category of white meat through imaging and analysis, but can also be used to detect various harmful substances in meat products to prevent stale and spoiled meat from entering the market and causing harm to consumer health and even the ecosystem. The development of quality inspection systems based on such techniques to measure and classify white meat quality parameters will help improve the productivity and economic efficiency of the meat industry, as well as the health of consumers. Herein, a comprehensive review and discussion of the literature on fluorescence spectroscopy, color imaging and MSI is presented. The principles of these three techniques, the quality analysis models selected and the research results of non-destructive determinations of white meat quality over the last decade or so are analyzed and summarized. The review is conducted in this highly practical research field in order to provide information for future research directions. The conclusions detail how these efficient and convenient imaging and analytical techniques can be used for non-destructive quality evaluation of white meat in the laboratory and in industry.