Near infrared spectroscopy for cooking time classification of cassava genotypes DOI Creative Commons
Massáine Bandeira e Sousa,

Cinara Fernanda Garcia Morales,

Edwige Gaby Nkouaya Mbanjo

и другие.

Frontiers in Plant Science, Год журнала: 2024, Номер 15

Опубликована: Июль 12, 2024

Cooking time is a crucial determinant of culinary quality cassava roots and incorporating it into the early stages breeding selection vital for breeders. This study aimed to assess potential near-infrared spectroscopy (NIRS) in classifying genotypes based on their cooking times. Five times (15, 20, 25, 30, 40 minutes) were assessed 888 evaluated over three crop seasons (2019/2020, 2020/2021, 2021/2022). Fifteen from five plants per plot, featuring diameters ranging 4 7 cm, randomly chosen analysis spectral data collection. Two root samples (15 slices each) genotype collected, with first set aside collection, processed, placed two petri dishes, while second was utilized assessment. classified binary multiclass variables (CT4C CT6C). NIRs devices, portable QualitySpec® Trek (QST) benchtop NIRFlex N-500 used collect data. Classification carried out using K-nearest neighbor algorithm (KNN) partial least squares (PLS) models. The split training (80%) an external validation (20%). For variables, classification accuracy notably high ( RCal2 0.72 0.99). Regarding remained consistent across classes, models, NIR instruments (~0.63). However, KNN model demonstrated slightly superior all except CT4C variable NoCook 25 min classes. Despite increased complexity associated classification, more efficient, offering higher facilitating most relevant or such as ≤ 30 minutes. optimal scenario minutes reached id="im2">RCal2 = 0.86 id="im3">RVal2 0.84, Kappa value 0.53. Overall, models exhibited robust fit times, showcasing significant high-throughput phenotyping tool time.

Язык: Английский

Cassava Breeding and Cultivation Challenges in Thailand: Past, Present, and Future Perspectives DOI Creative Commons
Pasajee Kongsil, Hernán Ceballos, Wanwisa Siriwan

и другие.

Plants, Год журнала: 2024, Номер 13(14), С. 1899 - 1899

Опубликована: Июль 10, 2024

Cassava (Manihot esculenta Crantz) was introduced to Southeast Asia in the 16th–17th centuries and has since flourished as an industrial crop. Since 1980s, Thailand emerged leading producer exporter of cassava products. This growth coincided with initiation breeding programs collaboration International Center for Tropical Agriculture (CIAT), focusing on root yield starch production. The success Thai can be attributed incorporation valuable genetic diversity from international germplasm resources cross local landraces, which become foundation many commercial varieties. Effective evaluation under diverse environmental conditions led release varieties high stability. A notable is development Kasetsart 50. However, extreme climate change poses significant challenges, including abiotic biotic stresses that threaten content, a potential decline starch-based industries. Future directions must include hybrid development, marker-assisted recurrent breeding, gene editing, along high-throughput phenotyping flower induction. These strategies are essential achieve objectives focused drought tolerance disease resistance, especially CMD CBSD.

Язык: Английский

Процитировано

6
Emerging Spectroscopic-Based Techniques for Rapid and Non-Destructive Quality Evaluation of Diverse Agricultural Seeds: A Review DOI
Marjun C. Alvarado,

Arsenio Bulfa

SSRN Electronic Journal, Год журнала: 2025, Номер unknown

Опубликована: Янв. 1, 2025

Язык: Английский

Процитировано

0
Comparative Analysis of XGB, CNN, and ResNet Models for Predicting Moisture Content in Porphyra yezoensis Using Near-Infrared Spectroscopy DOI Creative Commons
Wenwen Zhang,

Mingxuan Pan,

Peng Wang

и другие.

Foods, Год журнала: 2024, Номер 13(19), С. 3023 - 3023

Опубликована: Сен. 24, 2024

This study explored the performance and reliability of three predictive models—extreme gradient boosting (XGB), convolutional neural network (CNN), residual (ResNet)—for determining moisture content in Porphyra yezoensis using near-infrared (NIR) spectroscopy. We meticulously selected 380 samples from various sources to ensure a comprehensive dataset, which was then divided into training (300 samples) test sets (80 samples). The models were evaluated based on prediction accuracy stability, employing genetic algorithms (GA) partial least squares (PLS) for wavelength selection enhance interpretability feature extraction outcomes. results demonstrated that XGB model excelled with determination coefficient (R2) 0.979, root mean square error (RMSEP) 0.004, high ratio deviation (RPD) 4.849, outperforming both CNN ResNet models. A Gaussian process regression (GPR) employed uncertainty assessment, reinforcing our Considering model’s its implementation industrial settings quality assurance is recommended, particularly food industry where rapid non-destructive analysis essential. approach facilitates more efficient content, thereby enhancing product safety.

Язык: Английский

Процитировано

1
Near infrared spectroscopy for cooking time classification of cassava genotypes DOI Creative Commons
Massáine Bandeira e Sousa,

Cinara Fernanda Garcia Morales,

Edwige Gaby Nkouaya Mbanjo

и другие.

Frontiers in Plant Science, Год журнала: 2024, Номер 15

Опубликована: Июль 12, 2024

Cooking time is a crucial determinant of culinary quality cassava roots and incorporating it into the early stages breeding selection vital for breeders. This study aimed to assess potential near-infrared spectroscopy (NIRS) in classifying genotypes based on their cooking times. Five times (15, 20, 25, 30, 40 minutes) were assessed 888 evaluated over three crop seasons (2019/2020, 2020/2021, 2021/2022). Fifteen from five plants per plot, featuring diameters ranging 4 7 cm, randomly chosen analysis spectral data collection. Two root samples (15 slices each) genotype collected, with first set aside collection, processed, placed two petri dishes, while second was utilized assessment. classified binary multiclass variables (CT4C CT6C). NIRs devices, portable QualitySpec® Trek (QST) benchtop NIRFlex N-500 used collect data. Classification carried out using K-nearest neighbor algorithm (KNN) partial least squares (PLS) models. The split training (80%) an external validation (20%). For variables, classification accuracy notably high ( RCal2 0.72 0.99). Regarding remained consistent across classes, models, NIR instruments (~0.63). However, KNN model demonstrated slightly superior all except CT4C variable NoCook 25 min classes. Despite increased complexity associated classification, more efficient, offering higher facilitating most relevant or such as ≤ 30 minutes. optimal scenario minutes reached id="im2">RCal2 = 0.86 id="im3">RVal2 0.84, Kappa value 0.53. Overall, models exhibited robust fit times, showcasing significant high-throughput phenotyping tool time.

Язык: Английский

Процитировано

0