Pub Date : 2025-12-09DOI: 10.1016/j.jfoodeng.2025.112913
Priscila Alessio , Cibely S. Martin , Mateus D. Maximino , Matheus S. Pereira , Constantin Apetrei , Maria L. Rodriguez-Mendez
The immobilization of enzymes such as Tyrosinase (Tyr) in biomimetic nanostructured films is a promising approach to producing highly sensitive and selective biosensors for detecting phenolic compounds. In this study, an innovative approach to preserve the enzymatic functionality in a biomimetic environment is presented. Tyr was encapsulated within liposome structures formed by a lipid 1,2-dipalmitoyl-sn-glycero-3-phosphocholine (DPPC) using a straightforward technique, resulting in a grape-shaped morphology. The lutetium bisphthalocyanine (LuPc2) was also added to the liposome structure as electrocatalytic material and structured as thin film using the layer-by-layer (LbL) technique for biosensor application. The biosensor shows a Hill coefficient of 0.986 and a Michaelis–Menten constant of 21.96 μM, and a limit of detection of 44 nM for catechol detection. Measurements of polyphenolic contents performed in grape juice samples reveal that the encapsulation of Tyr inside liposomes preserves the functionality and improves the electrochemical performance for amperometric and voltametric detections.
{"title":"Encapsulation of tyrosinase in liposomes: A biomimetic approach to efficient biodetection of polyphenols in grape juice","authors":"Priscila Alessio , Cibely S. Martin , Mateus D. Maximino , Matheus S. Pereira , Constantin Apetrei , Maria L. Rodriguez-Mendez","doi":"10.1016/j.jfoodeng.2025.112913","DOIUrl":"10.1016/j.jfoodeng.2025.112913","url":null,"abstract":"<div><div>The immobilization of enzymes such as Tyrosinase (Tyr) in biomimetic nanostructured films is a promising approach to producing highly sensitive and selective biosensors for detecting phenolic compounds. In this study, an innovative approach to preserve the enzymatic functionality in a biomimetic environment is presented. Tyr was encapsulated within liposome structures formed by a lipid 1,2-dipalmitoyl-sn-glycero-3-phosphocholine (DPPC) using a straightforward technique, resulting in a grape-shaped morphology. The lutetium bisphthalocyanine (LuPc<sub>2</sub>) was also added to the liposome structure as electrocatalytic material and structured as thin film using the layer-by-layer (LbL) technique for biosensor application. The biosensor shows a Hill coefficient of 0.986 and a Michaelis–Menten constant of 21.96 μM, and a limit of detection of 44 nM for catechol detection. Measurements of polyphenolic contents performed in grape juice samples reveal that the encapsulation of Tyr inside liposomes preserves the functionality and improves the electrochemical performance for amperometric and voltametric detections.</div></div>","PeriodicalId":359,"journal":{"name":"Journal of Food Engineering","volume":"410 ","pages":"Article 112913"},"PeriodicalIF":5.8,"publicationDate":"2025-12-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145788870","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"农林科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-12-08DOI: 10.1016/j.jfoodeng.2025.112903
Eunji Ju , Jaehwi Seol , Sarang Kim , Sol Kim , Soo-Jung Kim , Hyoung Il Son
Three-dimensional food printing (3DFP) is receiving increasing attention for its ability to provide personalized nutrition and utilize alternative food sources, such as insects. However, food materials are highly sensitive to environmental factors such as temperature and humidity. This often causes defects such as nozzle clogging and uneven layer formation. These problems result in significant waste of materials and time, so continuous monitoring during the printing is essential. However, manual monitoring involves limitations owing to subjectivity and fatigue, monitoring must be done through an automated system. This paper proposes a deep learning-based monitoring system that can operate in real time and demonstrate robust performance despite environmental variations. The proposed system uses a camera to monitor the printing process in real time and detect both the type and location of defects. Since deep learning model has different characteristics, it is essential to compare models to select the most suitable model. In this study, three widely used segmentation models, U-Net, YOLOv8, and SegNet, were compared in terms of their performance and segmentation quality. A dataset was constructed using chocolate and three designs were printed three times in 27 cases. Using the dataset, the model learns sagging and thinning defects. The experimental results showed that U-Net demonstrated the best performance in defect detection. YOLOv8 displayed moderate performance with low sensitivity, highlighting its suitability for applications where speed is more important than accuracy. SegNet achieved the highest AUC value, suggesting that its performance can be enhanced via further optimization.
{"title":"Comparison of deep learning models for 3D food printing monitoring system","authors":"Eunji Ju , Jaehwi Seol , Sarang Kim , Sol Kim , Soo-Jung Kim , Hyoung Il Son","doi":"10.1016/j.jfoodeng.2025.112903","DOIUrl":"10.1016/j.jfoodeng.2025.112903","url":null,"abstract":"<div><div>Three-dimensional food printing (3DFP) is receiving increasing attention for its ability to provide personalized nutrition and utilize alternative food sources, such as insects. However, food materials are highly sensitive to environmental factors such as temperature and humidity. This often causes defects such as nozzle clogging and uneven layer formation. These problems result in significant waste of materials and time, so continuous monitoring during the printing is essential. However, manual monitoring involves limitations owing to subjectivity and fatigue, monitoring must be done through an automated system. This paper proposes a deep learning-based monitoring system that can operate in real time and demonstrate robust performance despite environmental variations. The proposed system uses a camera to monitor the printing process in real time and detect both the type and location of defects. Since deep learning model has different characteristics, it is essential to compare models to select the most suitable model. In this study, three widely used segmentation models, U-Net, YOLOv8, and SegNet, were compared in terms of their performance and segmentation quality. A dataset was constructed using chocolate and three designs were printed three times in 27 cases. Using the dataset, the model learns sagging and thinning defects. The experimental results showed that U-Net demonstrated the best performance in defect detection. YOLOv8 displayed moderate performance with low sensitivity, highlighting its suitability for applications where speed is more important than accuracy. SegNet achieved the highest AUC value, suggesting that its performance can be enhanced via further optimization.</div></div>","PeriodicalId":359,"journal":{"name":"Journal of Food Engineering","volume":"409 ","pages":"Article 112903"},"PeriodicalIF":5.8,"publicationDate":"2025-12-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145749274","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"农林科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-12-06DOI: 10.1016/j.jfoodeng.2025.112901
So-Yoon Park, Jae-Young Her
The creation of biodegradable packaging materials that mitigate environmental pollution and preserve food quality by valorizing agricultural waste is gaining increasing significance. In this study, microcrystalline cellulose (MCC) was isolated from corn husk powder (CHP) and incorporated into κ-carrageenan films to enhance their functional properties. MCC obtained via acid hydrolysis displayed significantly increased crystallinity, as determined by XRD, while effective removal of lignin, hemicellulose, and amorphous components was verified by FTIR analyses. Morphological (SEM) and thermal (TGA, DSC) characterizations demonstrated that the structural and thermal attributes of the extracted MCC were comparable to those of commercial MCC. The incorporation of 0.5 wt% (CAR/MCC0.5 %) resulted in a 67.2 % increase in tensile strength, a 49.2 % decrease in oxygen transmission rate, and improvements in both UV-blocking capability and thermal stability. This was demonstrated by a 29.7 % decrease in UV transmittance and an increase in the maximum degradation temperature (TPeak) from 220.86 °C to 225.59 °C, relative to the neat carrageenan film. During storage experiments using sunflower oil, CAR/MCC0.5 % achieved the lowest lipid oxidation rate constant (k = 3.326 meq O2/kg·day) and provided oxidative stability comparable to that of the positive control. A strong positive correlation between oxygen permeability and k validated the influence of barrier properties on oxidation inhibition. These results identify corn husk as a valuable feedstock for MCC production and confirm its suitability as a functional additive in biodegradable polymer composites, supporting its potential in sustainable food packaging applications.
{"title":"Corn husk-derived microcrystalline cellulose reinforced carrageenan films: Development, characterization, lipid oxidation kinetics, and food packaging application","authors":"So-Yoon Park, Jae-Young Her","doi":"10.1016/j.jfoodeng.2025.112901","DOIUrl":"10.1016/j.jfoodeng.2025.112901","url":null,"abstract":"<div><div>The creation of biodegradable packaging materials that mitigate environmental pollution and preserve food quality by valorizing agricultural waste is gaining increasing significance. In this study, microcrystalline cellulose (MCC) was isolated from corn husk powder (CHP) and incorporated into κ-carrageenan films to enhance their functional properties. MCC obtained via acid hydrolysis displayed significantly increased crystallinity, as determined by XRD, while effective removal of lignin, hemicellulose, and amorphous components was verified by FTIR analyses. Morphological (SEM) and thermal (TGA, DSC) characterizations demonstrated that the structural and thermal attributes of the extracted MCC were comparable to those of commercial MCC. The incorporation of 0.5 wt% (CAR/MCC<sup>0.5 %</sup>) resulted in a 67.2 % increase in tensile strength, a 49.2 % decrease in oxygen transmission rate, and improvements in both UV-blocking capability and thermal stability. This was demonstrated by a 29.7 % decrease in UV transmittance and an increase in the maximum degradation temperature (T<sub>Peak</sub>) from 220.86 °C to 225.59 °C, relative to the neat carrageenan film. During storage experiments using sunflower oil, CAR/MCC<sup>0.5 %</sup> achieved the lowest lipid oxidation rate constant (<em>k</em> = 3.326 meq O<sub>2</sub>/kg·day) and provided oxidative stability comparable to that of the positive control. A strong positive correlation between oxygen permeability and <em>k</em> validated the influence of barrier properties on oxidation inhibition. These results identify corn husk as a valuable feedstock for MCC production and confirm its suitability as a functional additive in biodegradable polymer composites, supporting its potential in sustainable food packaging applications.</div></div>","PeriodicalId":359,"journal":{"name":"Journal of Food Engineering","volume":"410 ","pages":"Article 112901"},"PeriodicalIF":5.8,"publicationDate":"2025-12-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145718993","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"农林科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-12-05DOI: 10.1016/j.jfoodeng.2025.112904
Zhili Qin , Jie Shi , Jinsong Tang , Gufen Zhang , Ruixue Huang , Chaoqun Yu
Perilla oil exhibits substantial nutritional value and has been widely applied in the food, cosmetic, and pharmaceutical industries. However, its practical application is restricted by its unpleasant odor, and its liquid state makes it inconvenient to carry and use. In this study, a nanocapsule formulation was developed using bovine serum albumin (BSA) as the wall material and perilla oil as the core. An emulsion was first prepared, followed by heat-induced denaturation of BSA at 120 °C for 30 min. Using lactose as a cryoprotectant, the nanocapsule powder was obtained via freeze-drying. For the optimized formulation, the particle size, zeta potential, encapsulation efficiency, and release rate were 438.2 ± 16.4 nm, −11.2 ± 0.44 mV, 83.87 ± 0.27 %, and 68.53 ± 2.58 %, respectively. Scanning electron microscopy (SEM) and transmission electron microscopy (TEM) confirmed that the nanocapsules were spherical with an average size of ∼500 nm. The lyophilized powder could be re-dissolved to form a homogeneous suspension. Electronic nose analysis and triangle sensory tests verified that the undesirable odor of perilla oil was effectively masked. This method enables simple and efficient encapsulation of perilla oil without the use of toxic crosslinkers, while enhancing the stability of perilla oil and expanding its application scope.
{"title":"Heat-induced bovine serum albumin nanocapsules for odor masking of perilla oil","authors":"Zhili Qin , Jie Shi , Jinsong Tang , Gufen Zhang , Ruixue Huang , Chaoqun Yu","doi":"10.1016/j.jfoodeng.2025.112904","DOIUrl":"10.1016/j.jfoodeng.2025.112904","url":null,"abstract":"<div><div>Perilla oil exhibits substantial nutritional value and has been widely applied in the food, cosmetic, and pharmaceutical industries. However, its practical application is restricted by its unpleasant odor, and its liquid state makes it inconvenient to carry and use. In this study, a nanocapsule formulation was developed using bovine serum albumin (BSA) as the wall material and perilla oil as the core. An emulsion was first prepared, followed by heat-induced denaturation of BSA at 120 °C for 30 min. Using lactose as a cryoprotectant, the nanocapsule powder was obtained via freeze-drying. For the optimized formulation, the particle size, zeta potential, encapsulation efficiency, and release rate were 438.2 ± 16.4 nm, −11.2 ± 0.44 mV, 83.87 ± 0.27 %, and 68.53 ± 2.58 %, respectively. Scanning electron microscopy (SEM) and transmission electron microscopy (TEM) confirmed that the nanocapsules were spherical with an average size of ∼500 nm. The lyophilized powder could be re-dissolved to form a homogeneous suspension. Electronic nose analysis and triangle sensory tests verified that the undesirable odor of perilla oil was effectively masked. This method enables simple and efficient encapsulation of perilla oil without the use of toxic crosslinkers, while enhancing the stability of perilla oil and expanding its application scope.</div></div>","PeriodicalId":359,"journal":{"name":"Journal of Food Engineering","volume":"409 ","pages":"Article 112904"},"PeriodicalIF":5.8,"publicationDate":"2025-12-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145749272","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"农林科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-12-04DOI: 10.1016/j.jfoodeng.2025.112902
Seunghun Lee , Sungmin Jeong , Suyong Lee
Hyperspectral imaging combined with artificial intelligence was applied to identify eight hydrocolloids representing diverse origins and structures relevant to food applications and to predict their viscosities under various processing conditions. Distinct hyperspectral patterns were observed among the hydrocolloids, with arabic gum and xanthan gum exhibiting high signal intensity due to their complex structures. Principal component analysis reduced the hyperspectral data to two components explaining 90.89 % of the total variance, enabling clear separation of the hydrocolloids into eight clusters. Convolutional neural network classification achieved accuracies above 96 %, with precision, recall, and F1-scores exceeding 0.95. Viscosities were experimentally measured at multiple concentrations (0.2–2.0 %, w/v) and temperatures (30–90 °C), and used to train three ensemble machine learning algorithms (XGBoost, Random Forest, and AdaBoost) for rapid viscosity prediction from hyperspectral features. XGBoost and Random Forest achieved superior performance with R2 values of 0.9969 and 0.9968 and root mean square errors of 0.0085 and 0.0087, respectively, while the AdaBoost model showed lower performance (R2 = 0.9256 and root mean square error = 0.0421). These findings demonstrate that hyperspectral imaging with artificial intelligence enables rapid and non-destructive hydrocolloid identification and accurate viscosity prediction, with dimensionality reduction enhancing prediction performance. This approach offers practical benefits for food formulation, process control, and quality improvement of hydrocolloid-containing products.
应用高光谱成像与人工智能相结合的方法，鉴定了与食品应用相关的8种不同来源和结构的水胶体，并预测了它们在不同加工条件下的粘度。在水胶体中观察到明显的高光谱模式，阿拉伯胶和黄原胶由于其复杂的结构而表现出高信号强度。主成分分析将高光谱数据简化为两个分量，解释了总方差的90.89%，从而将水胶体清晰地划分为8个簇。卷积神经网络分类准确率达到96%以上，准确率、查全率、f1得分均超过0.95。粘度在多种浓度（0.2 - 2.0%,w/v）和温度（30-90°C）下进行实验测量，并用于训练三种集成机器学习算法（XGBoost， Random Forest和AdaBoost），以便从高光谱特征中快速预测粘度。XGBoost和Random Forest模型表现较好，R2分别为0.9969和0.9968，均方根误差分别为0.0085和0.0087，而AdaBoost模型表现较差，R2 = 0.9256，均方根误差为0.0421。这些发现表明，人工智能的高光谱成像能够快速、无损地识别水胶体并准确预测粘度，而降维可以提高预测性能。这种方法为食品配方、工艺控制和含水胶体产品的质量改进提供了实际的好处。
{"title":"Hyperspectral imaging-based classification and viscosity prediction of hydrocolloids using convolutional neural network and ensemble models","authors":"Seunghun Lee , Sungmin Jeong , Suyong Lee","doi":"10.1016/j.jfoodeng.2025.112902","DOIUrl":"10.1016/j.jfoodeng.2025.112902","url":null,"abstract":"<div><div>Hyperspectral imaging combined with artificial intelligence was applied to identify eight hydrocolloids representing diverse origins and structures relevant to food applications and to predict their viscosities under various processing conditions. Distinct hyperspectral patterns were observed among the hydrocolloids, with arabic gum and xanthan gum exhibiting high signal intensity due to their complex structures. Principal component analysis reduced the hyperspectral data to two components explaining 90.89 % of the total variance, enabling clear separation of the hydrocolloids into eight clusters. Convolutional neural network classification achieved accuracies above 96 %, with precision, recall, and F1-scores exceeding 0.95. Viscosities were experimentally measured at multiple concentrations (0.2–2.0 %, w/v) and temperatures (30–90 °C), and used to train three ensemble machine learning algorithms (XGBoost, Random Forest, and AdaBoost) for rapid viscosity prediction from hyperspectral features. XGBoost and Random Forest achieved superior performance with R<sup>2</sup> values of 0.9969 and 0.9968 and root mean square errors of 0.0085 and 0.0087, respectively, while the AdaBoost model showed lower performance (R<sup>2</sup> = 0.9256 and root mean square error = 0.0421). These findings demonstrate that hyperspectral imaging with artificial intelligence enables rapid and non-destructive hydrocolloid identification and accurate viscosity prediction, with dimensionality reduction enhancing prediction performance. This approach offers practical benefits for food formulation, process control, and quality improvement of hydrocolloid-containing products.</div></div>","PeriodicalId":359,"journal":{"name":"Journal of Food Engineering","volume":"409 ","pages":"Article 112902"},"PeriodicalIF":5.8,"publicationDate":"2025-12-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145749273","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"农林科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-11-28DOI: 10.1016/j.jfoodeng.2025.112899
Amir Hosseinvand , Mahdi Ghafourivayghan
Food microbiology is critical to global health, especially given the substantial public health risks posed by foodborne illnesses. Rapid and accurate detection of pathogenic microorganisms in food is vital for ensuring safety and preventing outbreaks. This study presents a comprehensive review of biosensor-based approaches for identifying foodborne pathogens. Conventional detection methods including microbial culturing, colony counting, immunoassays, and PCR are often time consuming, requiring hours to days for results. These limitations have spurred the development of faster, more efficient alternatives. Biosensors, with their superior sensitivity and rapid response times, represent a transformative advancement in pathogen detection. By leveraging biorecognition elements and signal transduction mechanisms, biosensors enable real-time monitoring, significantly reducing analysis time while maintaining high accuracy. Their integration into food safety systems promises to enhance early contamination detection, improving both consumer protection and regulatory compliance. This review highlights the potential of biosensor technologies to revolutionize food microbiology diagnostics, addressing the urgent need for rapid, reliable pathogen screening in the food industry.
{"title":"Biosensor classification and microbiological applications: A food science and packaging perspective","authors":"Amir Hosseinvand , Mahdi Ghafourivayghan","doi":"10.1016/j.jfoodeng.2025.112899","DOIUrl":"10.1016/j.jfoodeng.2025.112899","url":null,"abstract":"<div><div><em>Food microbiology</em> is critical to global health, especially given the substantial public health risks posed by foodborne illnesses. Rapid and accurate detection of pathogenic microorganisms in food is vital for ensuring safety and preventing outbreaks. This study presents a comprehensive review of biosensor-based approaches for identifying foodborne pathogens. Conventional detection methods including microbial culturing, colony counting, immunoassays, and PCR are often time consuming, requiring hours to days for results. These limitations have spurred the development of faster, more efficient alternatives. Biosensors, with their superior sensitivity and rapid response times, represent a transformative advancement in pathogen detection. By leveraging biorecognition elements and signal transduction mechanisms, biosensors enable real-time monitoring, significantly reducing analysis time while maintaining high accuracy. Their integration into food safety systems promises to enhance early contamination detection, improving both consumer protection and regulatory compliance. This review highlights the potential of biosensor technologies to revolutionize food microbiology diagnostics, addressing the urgent need for rapid, reliable pathogen screening in the food industry.</div></div>","PeriodicalId":359,"journal":{"name":"Journal of Food Engineering","volume":"408 ","pages":"Article 112899"},"PeriodicalIF":5.8,"publicationDate":"2025-11-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145615908","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"农林科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-11-27DOI: 10.1016/j.jfoodeng.2025.112900
Zhengtao Xi , Tianhong Pan , Li Fang , Shan Chen
Near-infrared spectroscopy (NIRS) has emerged as highly effective analytical technique due to its rapid, non-destructive and non-polluting characteristics, and is increasingly used in various fields. However, traditional NIRS data analytical models are typically supervised, making them less suitable in scenarios where both labeled and unlabeled data coexist. To address this limitation, a self-training-based semi-supervised minimum redundancy and maximum relevance broad learning system integrated with deep deterministic policy gradient (SS-MRMR-DDPG-BLS) is proposed. In this framework, self-training enables simultaneous learning from labeled and unlabeled data, MRMR identifies salient spectral features, and DDPG optimizes the model's parameters. The present model was evaluated on the task of predicting sugar content in Huangshan Maofeng tea using NIRS data. Experimental results demonstrate that the SS-MRMR-DDPG-BLS significantly outperforms conventional machine learning algorithms in terms of prediction accuracy and generalization ability.
{"title":"Semi-supervised broad learning system for near-infrared spectroscopy","authors":"Zhengtao Xi , Tianhong Pan , Li Fang , Shan Chen","doi":"10.1016/j.jfoodeng.2025.112900","DOIUrl":"10.1016/j.jfoodeng.2025.112900","url":null,"abstract":"<div><div>Near-infrared spectroscopy (NIRS) has emerged as highly effective analytical technique due to its rapid, non-destructive and non-polluting characteristics, and is increasingly used in various fields. However, traditional NIRS data analytical models are typically supervised, making them less suitable in scenarios where both labeled and unlabeled data coexist. To address this limitation, a self-training-based semi-supervised minimum redundancy and maximum relevance broad learning system integrated with deep deterministic policy gradient (SS-MRMR-DDPG-BLS) is proposed. In this framework, self-training enables simultaneous learning from labeled and unlabeled data, MRMR identifies salient spectral features, and DDPG optimizes the model's parameters. The present model was evaluated on the task of predicting sugar content in Huangshan Maofeng tea using NIRS data. Experimental results demonstrate that the SS-MRMR-DDPG-BLS significantly outperforms conventional machine learning algorithms in terms of prediction accuracy and generalization ability.</div></div>","PeriodicalId":359,"journal":{"name":"Journal of Food Engineering","volume":"409 ","pages":"Article 112900"},"PeriodicalIF":5.8,"publicationDate":"2025-11-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145693378","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"农林科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
A rotating fluidized bed coater offers an alternative technology in which particles are fluidized by balancing centrifugal, buoyant, and gravitational forces, thereby improving coating homogeneity and mass transfer. The coating performance between the gradient orifice distributor (GOD) and the uniform orifice distributor (UOD) was compared at rotational speeds of 20, 70, or 120 rpm. The experiment was set up to coat a turmeric-extract solution onto the surface of jasmine rice. The coating substance was sprayed from the nozzle at 35 or 40 mL/min at an air-atomization pressure of 1.5 bar for 6 or 8 min. Hot air (55 or 60 °C) flows were passed at an inlet velocity of 2.9 m/s, with 85 % by volume of the air recycled. The quality of the sample was evaluated based on its final moisture content, percentage of fissure kernels, head-coated rice yield, color, coating efficiency, total phenolic content, and antioxidant activity. Based on the results, GOD was a suitable distributor plate for producing turmeric-extract-coated rice. For optimal coating conditions, the jasmine rice should be coated at a drying temperature of 55 °C, using a coating solution spray rate of 35 mL/min, a rotational speed of 70 rpm, a spraying coating time of 6 min, and a drying time of 10 s after the spraying stage. Using these conditions, the coating efficiency was 84.2 % and the antioxidant activity of the coated rice was 30.53 % higher than that of the referent white rice.
{"title":"Performance comparison between gradient orifice and uniform orifice distributors in rotating fluidization for coated jasmine rice","authors":"Preeda Prakotmak , Watchama Phothong , Chaiwat Rattanamechaiskul , Nittaya Junka","doi":"10.1016/j.jfoodeng.2025.112898","DOIUrl":"10.1016/j.jfoodeng.2025.112898","url":null,"abstract":"<div><div>A rotating fluidized bed coater offers an alternative technology in which particles are fluidized by balancing centrifugal, buoyant, and gravitational forces, thereby improving coating homogeneity and mass transfer. The coating performance between the gradient orifice distributor (GOD) and the uniform orifice distributor (UOD) was compared at rotational speeds of 20, 70, or 120 rpm. The experiment was set up to coat a turmeric-extract solution onto the surface of jasmine rice. The coating substance was sprayed from the nozzle at 35 or 40 mL/min at an air-atomization pressure of 1.5 bar for 6 or 8 min. Hot air (55 or 60 °C) flows were passed at an inlet velocity of 2.9 m/s, with 85 % by volume of the air recycled. The quality of the sample was evaluated based on its final moisture content, percentage of fissure kernels, head-coated rice yield, color, coating efficiency, total phenolic content, and antioxidant activity. Based on the results, GOD was a suitable distributor plate for producing turmeric-extract-coated rice. For optimal coating conditions, the jasmine rice should be coated at a drying temperature of 55 °C, using a coating solution spray rate of 35 mL/min, a rotational speed of 70 rpm, a spraying coating time of 6 min, and a drying time of 10 s after the spraying stage. Using these conditions, the coating efficiency was 84.2 % and the antioxidant activity of the coated rice was 30.53 % higher than that of the referent white rice.</div></div>","PeriodicalId":359,"journal":{"name":"Journal of Food Engineering","volume":"408 ","pages":"Article 112898"},"PeriodicalIF":5.8,"publicationDate":"2025-11-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145681639","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"农林科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-11-26DOI: 10.1016/j.jfoodeng.2025.112895
Yulun Chen , Yuliang Li , Kunfeng Liu , Yuhang Du , Hang Yu , Yahui Guo , Yuliang Cheng , Weirong Yao , Shaofeng Yuan
For the preservation of cherry tomatoes, this study reported the first utilization of lotus leaf essential oil (LEO) stabilized in gelatin (GL) nanoparticles in the development of a biomimetic film, a design that synergistically incorporates LEO's bioactivity, the stabilization effect of GL nanoparticles, and the functionality of the biomimetic structure. Carnauba wax was applied to achieve superhydrophobic properties. Compared to the bare GL film, the film exhibited improved mechanical strength, with a 42.94 % increase in water vapor barrier efficiency, a 75.12 % improvement in water resistance, and an 80 % boost in UV protection. Moreover, the controlled release of LEO in the film primarily driven by Fick diffusion. The film demonstrated superior antioxidant activity, with an 80 % increase compared to the bare GL film, and achieved antibacterial efficiency of over 93 %. When used to package cherry tomatoes under simulated harsh conditions, the film extended their shelf life by three days. This study presents an innovative approach to postharvest packaging of fruits.
{"title":"Multifunctional packaging film prepared by stabilizing lotus leaf essential oil with gelatin nanoparticles and biomimetic strategy for cherry tomato preservation","authors":"Yulun Chen , Yuliang Li , Kunfeng Liu , Yuhang Du , Hang Yu , Yahui Guo , Yuliang Cheng , Weirong Yao , Shaofeng Yuan","doi":"10.1016/j.jfoodeng.2025.112895","DOIUrl":"10.1016/j.jfoodeng.2025.112895","url":null,"abstract":"<div><div>For the preservation of cherry tomatoes, this study reported the first utilization of lotus leaf essential oil (LEO) stabilized in gelatin (GL) nanoparticles in the development of a biomimetic film, a design that synergistically incorporates LEO's bioactivity, the stabilization effect of GL nanoparticles, and the functionality of the biomimetic structure. Carnauba wax was applied to achieve superhydrophobic properties. Compared to the bare GL film, the film exhibited improved mechanical strength, with a 42.94 % increase in water vapor barrier efficiency, a 75.12 % improvement in water resistance, and an 80 % boost in UV protection. Moreover, the controlled release of LEO in the film primarily driven by Fick diffusion. The film demonstrated superior antioxidant activity, with an 80 % increase compared to the bare GL film, and achieved antibacterial efficiency of over 93 %. When used to package cherry tomatoes under simulated harsh conditions, the film extended their shelf life by three days. This study presents an innovative approach to postharvest packaging of fruits.</div></div>","PeriodicalId":359,"journal":{"name":"Journal of Food Engineering","volume":"408 ","pages":"Article 112895"},"PeriodicalIF":5.8,"publicationDate":"2025-11-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145615900","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"农林科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-11-25DOI: 10.1016/j.jfoodeng.2025.112896
Cara Anderton , Bettina Wolf , Lucio Cicerelli , Kristina Lodaitė , Ines Plasencia Gil , Fotis Spyropoulos
The influence of sucrose was investigated on the formation, microstructure and rheological behaviour of (1.5 wt%) agar fluid gels, with a focus on both the particulate and continuous phases. Rheology, texture analysis and phase contrast microscopy were used to assess the impact of sucrose concentration (0–60 wt%) and the timing of its addition. Increasing sucrose concentration resulted in a reduction of fluid gel particle size, increase in bulk viscosity, and slowed aggregation rate. At high concentrations of sucrose (>45 wt%) present during gelation, a significant increase in the gel particle stiffness and viscoelastic moduli was also observed. In contrast, sucrose addition after gel formation resulted in weakened rheological properties due to a reduction in effective volume fraction of the gel phase. Dilution studies revealed that unbound agar chains in the continuous phase contribute to maintaining interparticle interactions at reduced gel phase volume fractions. However, sucrose diminished the functional role of the continuous phase, likely due to contracted conformation of agar chains in the solution. These results provide insights into the functional relationship between agar and sucrose and offer practical guidance for the formulation of novel fluid gel-based foods where sucrose is a key component.
{"title":"Formulation of agar fluid gels in the presence of sugar for confectionery applications","authors":"Cara Anderton , Bettina Wolf , Lucio Cicerelli , Kristina Lodaitė , Ines Plasencia Gil , Fotis Spyropoulos","doi":"10.1016/j.jfoodeng.2025.112896","DOIUrl":"10.1016/j.jfoodeng.2025.112896","url":null,"abstract":"<div><div>The influence of sucrose was investigated on the formation, microstructure and rheological behaviour of (1.5 wt%) agar fluid gels, with a focus on both the particulate and continuous phases. Rheology, texture analysis and phase contrast microscopy were used to assess the impact of sucrose concentration (0–60 wt%) and the timing of its addition. Increasing sucrose concentration resulted in a reduction of fluid gel particle size, increase in bulk viscosity, and slowed aggregation rate. At high concentrations of sucrose (>45 wt%) present during gelation, a significant increase in the gel particle stiffness and viscoelastic moduli was also observed. In contrast, sucrose addition after gel formation resulted in weakened rheological properties due to a reduction in effective volume fraction of the gel phase. Dilution studies revealed that unbound agar chains in the continuous phase contribute to maintaining interparticle interactions at reduced gel phase volume fractions. However, sucrose diminished the functional role of the continuous phase, likely due to contracted conformation of agar chains in the solution. These results provide insights into the functional relationship between agar and sucrose and offer practical guidance for the formulation of novel fluid gel-based foods where sucrose is a key component.</div></div>","PeriodicalId":359,"journal":{"name":"Journal of Food Engineering","volume":"408 ","pages":"Article 112896"},"PeriodicalIF":5.8,"publicationDate":"2025-11-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145615903","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"农林科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
扫码关注我们
求助内容：
应助结果提醒方式：
京公网安备 11010802042870号