Browsing by Author "Castro, Wilson Manuel"

Now showing 1 - 7 of 7

Application of Machine Learning in the Discrimination of Citrus Fruit Juices: Uses of Dielectric Spectroscopy.
(Institute of Electrical and Electronics Engineers, 2020-10) Chuquizuta, Tony Steven; Oblitas, Jimy; Arteaga, Hubert; Castro, Wilson Manuel
Nowadays, process control in the juice industry requires fast, safe and easily applicable methods. In this regard, the use of dielectric spectroscopy is being coupled to statistical methods such as machine learning in order to develop new methods to identify adulteration. However, there is a small number of scientific reports above the application of the aforementioned methods when citric fruit juices is being identified. Therefore, the objective of this research was to evaluate dielectric spectroscopy and four different classification techniques (Support Vector Machine - SVM, K-nearest neighbor-KNN, Linear Discriminat -LD and Quadratic Discriminat-QD) to discriminate between three citrus juices. For this purpose, samples of Citrus limetta, Citrus limettioides and Citrus reticulata were evaluated; obtaining its dielectric spectral profiles in the range of 5 to 9 GHz. Then from the spectral profiles the loss factor (e”) was calculated using the reflection coefficient. Next e” value was pretreated, reducing noise through a savitzky golay filter, and new variables created through Principal Component Analysis (PCA). Finally, the models for classification were constructed with the previously mentioned techniques and the principal components. The results shown that using four components the variance can be explained in 97%; likewise, the discrimination values vary between 88.9 and 100.0%, with SVM, LD and QD the best discrimination techniques all successfully at 100.0 %. Therefore; It is concluded that the technique of dielectric spectroscopy and machine learning presents potential for the discrimination of citrus fruit juices.
Determination of hydration kinetic of pinto beans: A hyperspectral images application.
(Elsevier, 2024-03) Chuquizuta Trigoso, Tony Steven; Chavez, Segundo G.; Miano, Alberto Claudio; Castro-Giraldez, Marta; Fito, Pedro J.; Arteaga, Hubert; Castro, Wilson Manuel
Hydration is a typical operation applied to legumes before cooking, reducing time and the associated energy cost. To monitor the process, mass balance method is the most used methodology, despite this method is destructive, repetitive, and time-consuming. For that reason. hyperspectral techniques are presented as an alternative for assessing the hydration process since it is a noninvasive method. Therefore, the objective of this work was to evaluate the technique of hyperspectral imaging for studying the hydration kinetics of pinto beans. For this purpose, a sample of pinto beans was hydrated in distilled water, determining moisture content during the process and taking hyperspectral images by reflectance mode, in the range 400 to 800 nm until constant mass. The moisture content was modelled using Peleg and a sigmoidal model. Next, the images were pre-treated and the median spectral profile for each bean was obtained. Then, a regression model was fitted, using the wavelength that maximized the coefficient of determination (R2) and minimized the root mean square error (RMSE). The results show that Peleg model fit experimental data with R2 in the range of 0.974 to 0.989 while sigmoidal model of 0.997 to 0.999. On other hand, mean spectral profiles at 632 nm and sigmoidal model give the higher metrics 0.997 and 38.3 for R2 and RMSE respectively. The results showed that hyperspectral imaging in reflectance mode is a tool capable of measuring the hydration level of beans with higher performance at 632 nm, with a determination coefficient R2 higher than 0.98.
Dielectric Spectral Profiles for Andean Tubers Classification: A Machine Learning Techniques Application.
(Institute of Electrical and Electronics Engineers, 2021-09) Chuquizuta Trigoso, Tony Steven; Oblitas, Jimy; Arteaga, Hubert; Yarlequé Medina, Manuel A.; Castro, Wilson Manuel
Currently, the agri-food industry prioritizes the development of non-destructive methods, such as dielectric spectroscopy, for quality control. The obtained dielectric spectral properties can be coupled to multivariate statistical methods as "machine learning" when identification of attributes is wanted. However, these techniques have not been applied to andean tubers classification. Therefore, the objective of the present investigation is to evaluate the possibility of discriminating four andean tubers using dielectric spectra properties and machine learning techniques (Support Vector Machine - SVM, K-Nearest Neighbors-KNN, and Linear Discriminat - LD). For this purpose, samples of Tropaeolum tuberosum (Killu isañu), Solanum tuberosa (yellow) and two varieties of Oxalis tuberosa (Puka kamusa and Lari oqa) were acquired, 30 units per tuber. The dielectric spectral profile was extracted twice for each tubers sample, in the range from 2 to 8 GHz. Then, the dielectric constant (e') were calculated, and its dimensionality was reduced using principal component analysis. Finally, models for classification were built by employing KNN, SVM and LD techniques. The results showed that three components can explain the variance at 99.6 %. Likewise, the accuracy in the discrimination values varied between 79.17 - 83.04, being SVM the best discrimination technique. Consequently, it is concluded that the technique of dielectric spectroscopy and machine learning presents potential for andean tuber discrimination.
Dielectric spectroscopy for the prediction of pork quality during the post-mortem time
(Elsevier, 2025-08) Chuquizuta Trigoso, Tony Steven; Peralta, Magaly; Medina, Sideli; Arteaga, Hubert; Oblitas, Jimy; Chavez, Segundo G.; Castro, Wilson Manuel; Castro-Giraldez, Marta; Fito, Pedro Juan
Dielectric spectroscopy was used in this study to predict and classify pork quality during the post-mortem time. Eighty ~1 kg- longissimus dorsi muscles were collected and stored at 4 ± 1 ◦C and pH, instrumental color, and dielectric properties (ε’ and ε’’) were subsequently determined in the microwave range (0.5–9 GHz) at 3, 4, 5, 6, 7, 8, 9, 10 and 24 h post-mortem (hpm), as well as moisture at 8 hpm and drip weight loss at 24 hpm. Of the 80 pork samples, two types of meat were found. RFN (33) and DFD (47) between males and females. Quality parameters: RFN (pH=5.708–5.714; L*=43.341–43.692; moisture (%) = 68.857–69.604; drip loss = 1.655–1.833) and DFD (pH=6.154–6.177; L*=40.152–41.91; moisture (%) = 69.032–69.9; drip loss = 1.129–1.693). Quality parameter predictions during muscle-to-meat transformation showed R² of 0.743 (pH), 0.811 (L*) and 0.603 (C*) for DFD meats with PLSR (full) and R2 of 0.359 (pH), 0.558 (L*) and 0.284 (C*) for RNF meats with PLSR (optimized) from male pigs. R2 cv of 0.412–0.637 for pH, L* and c* for RFN and DFD meats from female pigs with PLSR (optimized). Dielectric spectroscopy predicts pork quality moderately well, but models that are more robust are needed to improve predictions of internal pork quality.
Impact of Magnetic Biostimulation and Environmental Conditions on the Agronomic Quality and Bioactive Composition of INIA 601 Purple Maize
(Multidisciplinary Digital Publishing Institute, 2025-06) Chuquizuta Trigoso, Tony Steven; Lobato, Cesar; Zirena Vilca, Franz; Huamán-Castilla, Nils Leander; Castro, Wilson Manuel; Castro-Giraldez, Marta; Fito, Pedro J.; Chavez, Segundo G.; Arteaga, Hubert
The utilization of magnetic fields in agricultural contexts has been demonstrated to exert a beneficial effect on various aspects of crop development, including germination, growth, and yield. The present study investigates the impact of magnetic biostimulation on seeds of purple maize (Zea mays L.), variety INIA 601, cultivated in Cajamarca, Peru, with a particular focus on their physical characteristics, yield, bioactive compounds, and antioxidant activity. The results demonstrated that seeds treated with pulsed (8 mT at 30 Hz for 30 min) and static (50 mT for 30 min) magnetic fields exhibited significantly longer cobs (16.89 and 16.53 cm, respectively) compared with the untreated control (15.79 cm). Furthermore, the application of these magnetic fields resulted in enhanced antioxidant activity in the bract, although the untreated samples exhibited higher values (110.56 µg/mL) compared with the pulsed (91.82 µg/mL) and static (89.61 µg/mL) treatments. The geographical origin of the samples had a significant effect on the physical development and the amount of total phenols, especially the antioxidant activity in the coronet and bract. Furthermore, a total of fourteen phenols were identified in various parts of the purple maize, with procyanidin B2 found in high concentrations in the bract and crown. Conversely, epicatechin, kaempferol, vanillin, and resveratrol were found in lower concentrations. These findings underscore the phenolic diversity of INIA 601 purple maize and its potential application in the food and pharmaceutical industries, suggesting that magnetic biostimulation could be an effective tool to improve the nutritional and antioxidant properties of crops.
Non-invasive monitoring of goldenberry freezing using infrared thermography and radiofrequency dielectric spectroscopy.
(Elsevier, 2025-07) Chuquizuta Trigoso, Tony Steven; Castro, Wilson Manuel; Castro-Giraldez, Marta; Fito, Pedro Juan
This study presents a non-invasive monitoring system combining infrared thermography and radiofrequency dielectric spectroscopy to characterize the freezing behavior of goldenberry (Physalis peruviana). The system enabled simultaneous acquisition of surface temperature profiles, internal dielectric responses, and emissivity changes during freezing at − 40 ◦C. Thermal imaging revealed distinct freezing stages, including subcooling, ice nucleation, and vitrification, with emissivity decreasing to 0.837 during initial dehydration and increasing to 0.951 near the glass transition (− 35.8 ◦C). Emissivity variations revealed key thermal transitions, while dielectric measurements identified α- and β-dispersions linked to ionic straight and surface tension of ice Ih formation, with relaxation frequencies decreasing progressively as freezing advanced. The integration of both techniques allowed the detection of critical phase transitions, including the onset and completion of ice crystallization, supported by differential scanning calorimetry. These findings provide insight into structural changes and water mobility in high-moisture fruits, enabling real-time assessment of freezing kinetics. The approach demonstrates significant potential for optimizing industrial freezing protocols, improving the preservation of delicate fruits by minimizing structural damage and degradation of bioactive compounds.
Predicción de atributos de calidad de leche fresca no pasteurizada mediante espectroscopia dieléctrica acoplada a herramientas quimiométricas.
(Institute of Electrical and Electronics Engineers, 2022-06) Chuquizuta Trigoso, Tony Steven; Colunche, Y.; Rubio, M.; Oblitas, Jimy; Arteaga, Hubert; Castro, Wilson Manuel
El objetivo de esta investigación es predecir los atributos de calidad de la leche fresca no pasteurizada mediante espectroscopia dieléctrica acoplada a herramientas quimiométricas. Para ello, se trabajó con leche fresca no pasteurizada de la raza Pardo Suizo, obtenida del establo “La Lechera”. Se obtuvieron diluciones de agua y leche fresca del 70 al 100 %.25∘do, seguida de la caracterización fisicoquímica (densidad, sólidos totales, punto de congelación, sólidos grasos, proteínas y agua añadida) y las propiedades dieléctricas en el rango de 0,5 a 9 GHz mediante una sonda coaxial de extremo abierto (N1501A-001), conectada a un Analizador de Redes Vectoriales, modelo N9915A-Keysight Technologies. Asimismo, se empleó la regresión de mínimos cuadrados parciales para correlacionar las propiedades fisicoquímicas con las propiedades dieléctricas. Los resultados obtenidos en la predicción del punto de congelación, las proteínas, los sólidos grasos y el agua añadida de leche fresca no pasteurizada presentaron un coeficiente de determinación y un error cuadrático medio en el rango de [0,95-0,98] y [2]..57 ×10− 7− 7,46 ×10− 2]En consecuencia, se concluye que la técnica de espectroscopia dieléctrica y aprendizaje automático presenta potencial para la predicción de las características fisicoquímicas de la leche fresca no pasteurizada, pudiendo implementarse en las líneas de producción para evaluar de forma rápida y fiable la calidad de la leche de vaca.