Filters
SkutecznoĆÄ rozpoznawania proszkĂłw porzeczkowych za pomocÄ zespoĆĂłw klasyfikatorĂłw (classifier ensembles)
The application of convolutional neural networks, LF-NMR, and texture for microparticle analysis in assessing the quality of fruit powders: Case study â blackcurrant powders
Abstract It can be observed that dynamic developments in artificial intelligence contributing to the evolution of existing techniques used in food research. Currently, innovative methods are being sought to support unit processes such as food drying, while at the same time monitoring quality and extending their shelf life. The development of innovative technology using convolutional neural networks (CNNs) to assess the quality of fruit powders seems highly desirable. This will translate into obtaining homogeneous batches of powders based on the specific morphological structure of the obtained microparticles. The research aims to apply convolutional networks to assess the quality, consistency, and homogeneity of blackcurrant powders supported by comparative physical methods of low-field nuclear magnetic resonance (LF-NMR) and texture analysis. The results show that maltodextrin, inulin, whey milk proteins, microcrystalline cellulose, and gum arabic are effective carriers when identifying morphological structure using CNNs. The use of CNNs, texture analysis, and the effect of LF-NMR relaxation time together with statistical elaboration shows that maltodextrin as well as milk whey proteins in combination with inulin achieve the most favorable results. The best results were obtained for a sample containing 50% maltodextrin and 50% maltodextrin (MD50-MD70). The CNN model for this combination had the lowest mean squared error in the test set at 2.5741 Ă 10â4, confirming its high performance in the classification of blackcurrant powder microstructures.
Effects of Thickness of the Corn Seed Coat on the Strength of Processed Biological Materials
The strength and energy of processed biological materials depend, among others, on their properties. Despite the numerous studies available, the relationship between the internal structure of corn grains and their mechanical properties has not yet been explained. Hence, the aim of the work is to explore the relationship between the internal composition of maize kernels and its mechanical properties by studying the impact of the maize seed coat thickness on its breakage susceptibility. To achieve the assumed goal, selected physical properties (length, width, and thickness) of corn grains were distinguished, and a static compression test was carried out on the Insight 50 kN testing machine (MTS Systems Corporation, Eden Prairie, MN, USA) with a test system for experimental verification of the compression behavior of biological materials. Furthermore, after the compression test, the thickness of the seed coat was measured using a laboratory microscope. It was found that there is a correlation between the thickness of the maize seed coat and force, deformation, and mass-specific energy at the bioyield point. The presented data constitute a foundation for the development of a mechanistic breakage model considering the variable strength properties of the seed coat and endosperm as the structural elements of kernels. Further research should be focused on the determination of the strength properties under dynamic conditions and revealing the relationship between the loading rate, strength properties, and internal structure for several maize varieties, which better reflect the ranges of variability in the real nature of mechanical processing.
Applications MLP and other methods in artificial intelligence of fruit and vegetable in convective and spray drying
The seasonal nature of fruits and vegetables has an immense impact on the process of seeking methods that allow extending the shelf life in this category of food. It is observed that through continuous technological changes, it is also possible to notice changes in the methods used to examine and study food and its microbiological aspects. It should be added that a new trend of bioactive ingredient consumption is also on the increase, which translates into numerous attempts that are made to keep the high quality of those products for a longer time. New and modern methods are being sought in this area, where the main aim is to support drying processes and quality control during food processing. This review provides deep insight into the application of artificial intelligence (AI) using a multi-layer perceptron network (MLPN) and other machine learning algorithms to evaluate the effective prediction and classification of the obtained vegetables and fruits during convection as well as spray drying. AI in food drying, especially for entrepreneurs and researchers, can be a huge chance to speed up development, lower production costs, effective quality control and higher production efficiency. Current scientific findings confirm that the selection of appropriate parameters, among others, such as color, shape, texture, sound, initial volume, drying time, air temperature, airflow velocity, area difference, moisture content and final thickness, have an influence on the yield as well as the quality of the obtained dried vegetables and fruits. Moreover, scientific discoveries prove that the technology of drying fruits and vegetables supported by artificial intelligence offers an alternative in process optimization and quality control and, even in an indirect way, can prolong the freshness of food rich in various nutrients. In the future, the main challenge will be the application of artificial intelligence in most production lines in real time in order to control the parameters of the process or control the quality of raw materials obtained in the process of drying.
Analiza proszku owocowego z wykorzystaniem spektroskopii w podczerwieni z transformacjÄ Fouriera (FTIR ATR)
Application of artificial neural network for the quality-based classification of spray-dried rhubarb juice powders
AbstractThe aim of the study was to develop a neural model enabling classification of fruit spray dried powders, on the basis of graphic data acquired from a bitmap received in the process of spray drying. The neural model was developed with multi-layer perceptron topology. Input variables were expressed in 46 image descriptors based on RGB, YCbCr, HSV (B) and HSL models. Sensitivity analysis of input variables and principal component analysis determined the significance level of each attribute. The optimal model with the lowest error value root mean square, at the level of 0.04 contained 46 neurons in the input layer, 11 neurons in the hidden layer, 10 neurons in the output layer. The results allowed to show that dyeing force (color features) had influence on effective differentiation of the research material consisting of spray-dried powders of rhubarb juice with various dried juice content levels: 30, 40 and 50% as well as high (âHâ) and low (âLâ) level of saccharification a chosen carrier (potato maltodextrin).
Fruit Powder Analysis Using Machine Learning Based on Color and FTIR-ATR Spectroscopy - Case Study: Blackcurrant Powders
Fruits represent a valuable source of bioactivity, vitamins, minerals and antioxidants. They are often used in research due to their potential to extend sustainability and edibility. In this research, the currants were used to obtain currant powders by dehumidified air-assisted spray drying. In the research analysis of currant powders, advanced machine learning techniques were used in combination with Lab color space model analysis and Fourier transform infrared spectroscopy (FTIR). The aim of this project was to provide authentic information about the qualities of currant powders, taking into account their type and carrier content. In addition, the machine learning models were developed to support the recognition of individual blackcurrant powder samples based on Lab color. These results were compared using their physical properties and FTIR spectroscopy to determine the homogeneity of these powders; this will help reduce operating and energy costs while also increasing the production rate, and even the possibility of improving the available drying system.
Efficiency of Identification of Blackcurrant Powders Using Classifier Ensembles
In the modern times of technological development, it is important to select adequate methods to support various food and industrial problems, including innovative techniques with the help of artificial intelligence (AI). Effective analysis and the speed of algorithm implementation are key points in assessing the quality of food products. Non-invasive solutions are being sought to achieve high accuracy in the classification and evaluation of various food products. This paper presents various machine learning algorithm architectures to evaluate the efficiency of identifying blackcurrant powders (i.e., blackcurrant concentrate with a density of 67 °Brix and a color coefficient of 2.352 (E520/E420) in combination with the selected carrier) based on information encoded in microscopic images acquired via scanning electron microscopy (SEM). Recognition of blackcurrant powders was performed using texture feature extraction from images aided by the gray-level co-occurrence matrix (GLCM). It was evaluated for quality using individual single classifiers and a metaclassifier based on metrics such as accuracy, precision, recall, and F1-score. The research showed that the metaclassifier, as well as a single random forest (RF) classifier most effectively identified blackcurrant powders based on image texture features. This indicates that ensembles of classifiers in machine learning is an alternative approach to demonstrate better performance than the existing traditional solutions with single neural models. In the future, such solutions could be an important tool to support the assessment of the quality of food products in real time. Moreover, ensembles of classifiers can be used for faster analysis to determine the selection of an adequate machine learning algorithm for a given problem.
Analiza wpĆywu morfologii mikroczÄ steczek na stan jakoĆciowy suszonych rozpyĆowo owocĂłw z zastosowaniem uczenia maszynowego i gĆÄbokiego
Interactions of Oleanolic Acid, Apigenin, Rutin, Resveratrol and Ferulic Acid with Phosphatidylcholine Lipid Membranes - A Spectroscopic and Machine Learning Study
Biologically active compounds present in the diet can interact with biological membranes (such as cell membranes), changing their properties. Their mutual interactions can influence their respective activities. In this study, we analyzed the interactions of oleanolic acid and phenolic compounds such as apigenin, rutin, resveratrol and ferulic acid with phosphatidylcholine membranes. Spectroscopic methods (fluorescence spectroscopy, dynamic light scattering) and machine learning were applied. The results of structural studies were compared with the antioxidant activity of the investigated substances in lipid membranes. In liposomes loaded with oleanolic acid, the pro-oxidant activity of resveratrol arises from changes in membrane structure, leading to an increased exposure of its hydrophilic region to external radicals. A similar mechanism may be involved in the pro-oxidant action of oleanolic acid. By contrast, apigenin, rutin and ferulic acid are present at the membrane surface. Their presence in this region protects the bilayer from radicals generated in the aqueous phase. Lower antioxidant activity observed in the case of ferulic aid is probably related to weaker interactions of this compound with the membrane, compared to the investigated flavonoids. Appropriate machine learning models for predicting oleanolic acid and phenolic compounds have been developed for the future application of intelligent predictive systems to optimizing manufacturing processes involving liposomes. The most effective regression model turned out to be the MLP 1:1-100-50-50-6:1, identifying resveratrol with a determination index of 0.83.
Molecular Properties of Starch-Water Interactions in the Presence of Bioactive Compounds from Barley and Buckwheat-LF NMR Preliminary Study
The Prediction of Pectin Viscosity Using Machine Learning Based on Physical CharacteristicsâCase Study: Aglupectin HS-MR
In the era of technology development, the optimization of production processes, quality control and at the same time increasing production efficiency without wasting food, artificial intelligence is becoming an alternative tool supporting many decision-making processes. The work used modern machine learning and physical analysis tools to evaluate food products (pectins). Various predictive models have been presented to estimate the viscosity of pectin. Based on the physical analyses, the characteristics of the food product were isolated, including L*a*b* color, concentration, conductance and pH. Prediction was determined using the determination index and loss function for individual machine learning algorithms. As a result of the work, it turned out that the most effective estimation of pectin viscosity was using Decision Tree (R2 = 0.999) and Random Forest (R2 = 0.998). In the future, the prediction of pectin properties in terms of viscosity recognition may be significantly perceived, especially in the food and pharmaceutical industries. Predicting the natural pectin substrate may contribute to improving quality, increasing efficiency and at the same time reducing losses of the obtained final product.
Machine Learning in Sensory Analysis of MeadâA Case Study: Ensembles of Classifiers
The aim was to explore using machine learning (including cluster mapping and k-means methods) to classify types of mead based on sensory analysis and aromatic compounds. Machine learning is a modern tool that helps with detailed analysis, especially because verifying aromatic compounds is challenging. In the first stage, a cluster map analysis was conducted, allowing for the exploratory identification of the most characteristic features of mead. Based on this, k-means clustering was performed to evaluate how well the identified sensory features align with logically consistent groups of observations. In the next stage, experiments were carried out to classify the type of mead using algorithms such as Random Forest (RF), adaptive boosting (AdaBoost), Bootstrap aggregation (Bagging), K-Nearest Neighbors (KNN), and Decision Tree (DT). The analysis revealed that the RF and KNN algorithms were the most effective in classifying mead based on sensory characteristics, achieving the highest accuracy. In contrast, the AdaBoost algorithm consistently produced the lowest accuracy results. However, the Decision Tree algorithm achieved the highest accuracy value (0.909), demonstrating its potential for precise classification based on aroma characteristics. The error matrix analysis also indicated that acacia mead was easier for the algorithms to identify than tilia or buckwheat mead. The results show the potential of combining an exploratory approach (cluster map with the k-means method) with machine learning. It is also important to focus on selecting and optimizing classification models used in practice because, as the results so far indicate, choosing the right algorithm greatly affects the success of mead identification.
Explainable AI: Machine Learning Interpretation in Blackcurrant Powders
Recently, explainability in machine and deep learning has become an important area in the field of research as well as interest, both due to the increasing use of artificial intelligence (AI) methods and understanding of the decisions made by models. The explainability of artificial intelligence (XAI) is due to the increasing consciousness in, among other things, data mining, error elimination, and learning performance by various AI algorithms. Moreover, XAI will allow the decisions made by models in problems to be more transparent as well as effective. In this study, models from the âglass boxâ group of Decision Tree, among others, and the âblack boxâ group of Random Forest, among others, were proposed to understand the identification of selected types of currant powders. The learning process of these models was carried out to determine accuracy indicators such as accuracy, precision, recall, and F1-score. It was visualized using Local Interpretable Model Agnostic Explanations (LIMEs) to predict the effectiveness of identifying specific types of blackcurrant powders based on texture descriptors such as entropy, contrast, correlation, dissimilarity, and homogeneity. Bagging (Bagging_100), Decision Tree (DT0), and Random Forest (RF7_gini) proved to be the most effective models in the framework of currant powder interpretability. The measures of classifier performance in terms of accuracy, precision, recall, and F1-score for Bagging_100, respectively, reached values of approximately 0.979. In comparison, DT0 reached values of 0.968, 0.972, 0.968, and 0.969, and RF7_gini reached values of 0.963, 0.964, 0.963, and 0.963. These models achieved classifier performance measures of greater than 96%. In the future, XAI using agnostic models can be an additional important tool to help analyze data, including food products, even online.
Application of Machine Learning to Assess the Quality of Food Products - Case Study: Coffee Bean
Modern machine learning methods were used to automate and improve the determination of an effective quality index for coffee beans. Machine learning algorithms can effectively recognize various anomalies, among others factors, occurring in a food product. The procedure for preparing the machine learning algorithm depends on the correct preparation and preprocessing of the learning set. The set contained coded information (i.e., selected quality coefficients) based on digital photos (input data) and a specific class of coffee bean (output data). Because of training and data tuning, an adequate convolutional neural network (CNN) was obtained, which was characterized by a high recognition rate of these coffee beans at the level of 0.81 for the test set. Statistical analysis was performed on the color data in the RGB color space model, which made it possible to accurately distinguish three distinct categories of coffee beans. However, using the Lab* color model, it became apparent that distinguishing between the quality categories of under-roasted and properly roasted coffee beans was a major challenge. Nevertheless, the Lab* model successfully distinguished the category of over-roasted coffee beans.