Review
Food Science & Technology
Yuanyuan Pu, Dolores Perez-Marin, Norah O'Shea, Ana Garrido-Varo
Summary: Quality and safety monitoring in the dairy industry is crucial to ensure products meet high standards. The development of miniaturised NIR spectrometers has provided a powerful capability for on-site or on-farm product measurements in the dairy sector.
Article
Food Science & Technology
Lorenzo Strani, Silvia Grassi, Cristina Alamprese, Ernestina Casiraghi, Roberta Ghiglietti, Francesco Locci, Nicolo Pricca, Anna De Juan
Summary: The effect of physicochemical factors and use of skim milk powder on milk rennet-coagulation was investigated using NIR spectroscopic monitoring and MCR-ALS models, revealing the significant impact of milk powder type on the coagulation process. The models successfully described the process evolution, explaining over 99.9% of variance and providing non-destructive and online tools for evaluating rennet-induced coagulation of reconstituted milks under various conditions.
Article
Food Science & Technology
Shijie Shi, Junheng Feng, Yingying Ma, Cougui Cao, Lina Li, Yang Jiang
Summary: This study used NIR spectroscopy and PLSR to rapidly and accurately predict the levels of illegal additives in wheat flour. The CARS-RF and RF algorithms showed great potential in accurately predicting the presence of illegal additives.
LWT-FOOD SCIENCE AND TECHNOLOGY
(2023)
Article
Agriculture, Dairy & Animal Science
Juliana S. Lima, Daniela C. S. Z. Ribeiro, Habib Asseiss Neto, Sergio V. A. Campos, Monica O. Leite, Marcia E. De R. Fortini, Beatriz Pinho Martins de Carvalho, Marcos Vinicius Oliveira Almeida, Leorges M. Fonseca
Summary: The addition of cheese whey to milk is a common form of fraud that has severe economic effects. Current methods for detecting this fraud are expensive and time consuming. This study evaluated the use of FTIR with machine learning methods to detect the addition of cheese whey to milk, and found it to be a highly efficient method.
JOURNAL OF DAIRY SCIENCE
(2022)
Article
Nutrition & Dietetics
Shaoli Liu, Ting Lei, Guipu Li, Shuming Liu, Xiaojun Chu, Donghai Hao, Gongnian Xiao, Ayaz Ali Khan, Taqweem Ul Haq, Manal Y. Sameeh, Tariq Aziz, Manal Tashkandi, Guanghua He
Summary: In this study, four methods were employed to preprocess the original spectra of infant formula milk powder, and two algorithms were used to extract characteristic wavelengths. PLSR and SVR models were established to predict the contents of micronutrient components. The results showed that the preprocessing methods and algorithms were effective in extracting feature wavelengths and achieving accurate predictions. This study is important for online detection and optimization control of nutritional components in infant formula production.
FRONTIERS IN NUTRITION
(2023)
Article
Engineering, Chemical
Pegah Sadeghi Vasafi, Olivier Paquet-Durand, Kim Brettschneider, Joerg Hinrichs, Bernd Hitzmann
Summary: Combining autoencoder neural network with near-infrared spectroscopy is a reliable method to monitor milk processing, detecting abnormal changes early, making process control easier, and ensuring product quality and safety.
JOURNAL OF FOOD ENGINEERING
(2021)
Article
Food Science & Technology
Jelena Muncan, Zoltan Kovacs, Bernhard Pollner, Kentarou Ikuta, Yoshihisa Ohtani, Fuminori Terada, Roumiana Tsenkova
Summary: This study aimed to explore the potential of using NIR spectroscopy and aquaphotomics to quantify common dietary fatty acids and fat content in cow's liquid milk. The results showed improved quantification accuracy when focusing on the 1600-1800 nm spectral region, with good predictions for certain fatty acids. Influential variables in regression models highlighted the importance of water-FAs interaction for self-organization into different assemblies.
Article
Food Science & Technology
Huseyin Ayvaz, Mustafa Mortas, Muhammed Ali Dogan, Mustafa Atan, Gulgun Yildiz Tiryaki, Yonca Karagul Yuceer
Summary: When evaluating quality parameters of commercial white cheese, NIR and MIR-ATR spectroscopy showed satisfactory performance in total protein, fat, salt, dry matter, moisture, ash content, pH, and titratable acidity. However, they performed poorly in predicting textural properties, nitrogen fractions, and ripening index. NIR spectroscopy was generally more accurate except for pH and titratable acidity, where MIR-ATR spectroscopy was superior.
JOURNAL OF FOOD SCIENCE AND TECHNOLOGY-MYSORE
(2021)
Article
Agriculture, Dairy & Animal Science
Ke Yang, Changqing An, Jieliang Zhu, Wenchuan Guo, Chang Lu, Xinhua Zhu
Summary: This study compared the performance of near-infrared spectroscopy (NIRS) and dielectric spectroscopy (DS) in quantitatively predicting the content of mature milk as an adulterant in bovine colostrum. The results showed that DS had better identification performance, providing important insights for the quantitative prediction of nonhomogeneous liquid food.
JOURNAL OF DAIRY SCIENCE
(2022)
Article
Materials Science, Ceramics
Yao Wang, Pei-feng Hsu, Yingsang Wu
Summary: Thermal barrier coatings (TBCs) using yttria-stabilized zirconium dioxide (YSZ) are widely used to protect metal components in gas turbines from high combustion temperatures. Radiative heat transfer becomes crucial in TBCs due to increasing combustion temperature and pressure. This study investigates the thermal radiative properties of YSZ films in the near-infrared wavelength range using a hybrid method that combines the Kubelka-Munk and discrete ordinates methods.
INTERNATIONAL JOURNAL OF APPLIED CERAMIC TECHNOLOGY
(2022)
Article
Agriculture, Dairy & Animal Science
Hulya Yaman, Didem P. Aykas, Rafael Jimenez-Flores, Luis E. Rodriguez-Saona
Summary: A rapid and simple method based on vibrational spectroscopic techniques was developed to understand biochemical changes during the ripening process of Turkish white cheese and generate predictive algorithms. The models showed good correlation between predicted values by vibrational spectroscopy and reference values, providing real-time tools for addressing deviations in cheese manufacturing. Portable vibrational spectroscopy units can be used for rapid in-situ monitoring of cheese quality during aging.
JOURNAL OF DAIRY SCIENCE
(2022)
Article
Food Science & Technology
Paulo Augusto Da Costa Filho, Yike Chen, Christophe Cavin, Roberto Galluzzo
Summary: This study demonstrates that MilkoScanTM FT1 is a rapid screening technology for detecting food adulterants in reconstituted skimmed milk powder. The study accurately identified the presence of adulterants in samples, including nitrogen-rich compounds and bulking agents, and showed the challenges in detecting lactose. However, specific calibration methods were able to detect the addition of lactose. This screening method has potential applications in various food businesses, ensuring better quality and safety for consumers.
Article
Pharmacology & Pharmacy
Natasha L. Velez-Silva, James K. Drennen III, Carl A. Anderson
Summary: This study focuses on assessing the impact of powder physical variation on near-infrared (NIR) spectroscopy and develops an in-line characterization method for powder stream density in a simulated continuous system. The results demonstrate that powder density is a significant source of spectral variability due to flow rate. This study is important for facilitating continuous process scale-up and ensuring the robustness of analytical models.
INTERNATIONAL JOURNAL OF PHARMACEUTICS
(2023)
Article
Biochemical Research Methods
C. Grelet, P. Dardenne, H. Soyeurt, J. A. Fernandez, A. Vanlierde, F. Stevens, N. Gengler, F. Dehareng
Summary: The development of models for predicting milk traits has been active, but there are challenges in implementing these models at a large scale. Factors such as data variability, spectral regions, and model complexity can impact prediction accuracy. Evaluation methods and indicators become important for assessing model and prediction quality.
Article
Engineering, Chemical
Sijun Wu, Xiaoyang Zhang, Guoming Zhou, Jiaheng Wu, Wen Song, Ying Zhang, Zheng Li, Wenlong Li
Summary: A method based on near-infrared spectroscopy was proposed to rapidly determine physical parameters of herbal medicine. The potential of direct standardization, partial least squares regression, and generalized regression neural network (GRNN) for physical fingerprint transformation was investigated. The results showed that the predictive capacity of GRNN models was the best.
ADVANCED POWDER TECHNOLOGY
(2023)