Convolutional neural network allows amylose content prediction in yam (Dioscorea alata L.) flour using near infrared spectroscopy

Houngbo Mahugnon Ezekiel, Desfontaines Lucienne, Diman Jean-Louis, Arnau Gemma, Mestres Christian, Davrieux Fabrice, Rouan Lauriane, Beurier Grégory, Marie-Magdeleine Carine, Meghar Karima, Alamu Emmanuel Oladeji, Otegbayo Bolanle Omolara, Cornet Denis. 2024. Convolutional neural network allows amylose content prediction in yam (Dioscorea alata L.) flour using near infrared spectroscopy. Journal of the Science of Food and Agriculture, 104 (8), n.spéc. Tropical roots, tubers and bananas: New breeding tools and methods to meet consumer preferences : 4915-4921.

Résumé : Background: Yam (Dioscorea alata L.) is the staple food of many populations in the intertropical zone where it is grown. The lack of phenotyping methods for tuber quality hinders the adoption of new genotypes from the breeding programs. Recently, near infrared spectroscopy (NIRS) has been used as a reliable tool to characterize the chemical composition of the yam tuber. However, it failed to predict the amylose content, although this trait is strongly involved in the quality of the product. Results: This study used NIRS to predict the amylose content from 186 yam flour samples. Two calibration methods were developed and validated on an independent dataset: Partial Least Square (PLS) and Convolutional Neural Network (CNN). To evaluate final model performances, the coefficient of determination (R2), the root mean square error (RMSE), and the Ratio of Performance to Deviation (RPD) were calculated using predictions on an independent validation dataset. Tested models showed contrasting performances (i.e. R2 of 0.72 and 0.89, RMSE of 1.33 and 0.81, RPD of 2.13 and 3.49 respectively, for the PLS and the CNN model). Conclusion: According to the quality standard for NIRS model prediction used in food science, the PLS method proved unsuccessful (RPD<3 and R2<0.8) for predicting amylose content from yam flour, while the CNN proved reliable and efficient method. With the application of deep learning method, this study established the proof of concept that amylose content, a key driver of yam textural quality and acceptance, could be predicted accurately using NIRS as a high throughput phenotyping method.

Mots-clés Agrovoc : Dioscorea alata, composition des aliments, spectroscopie infrarouge

Mots-clés libres : Consumer acceptability, Tuber quality, Amylose content, Near infrared spectrometry, Convolutional neural network, High throughput phenotyping

Classification Agris : Q04 - Composition des produits alimentaires

Champ stratégique Cirad : CTS 3 (2019-) - Systèmes alimentaires

Agences de financement hors UE : Consortium of International Agricultural Research Centers, Centre de Coopération Internationale en Recherche Agronomique pour le Développement, Bill and Melinda Gates Foundation

Projets sur financement : (FRA) Breeding RTB Products for End User Preferences

Auteurs et affiliations

• Houngbo Mahugnon Ezekiel, CIRAD-BIOS-UMR AGAP (FRA)
• Desfontaines Lucienne, INRAE (FRA)
• Diman Jean-Louis, INRAE (FRA)
• Arnau Gemma, CIRAD-BIOS-UMR AGAP (FRA) ORCID: 0000-0003-0714-6830
• Mestres Christian, CIRAD-PERSYST-UMR Qualisud (FRA)
• Davrieux Fabrice, CIRAD-PERSYST-UMR Qualisud (REU)
• Rouan Lauriane, CIRAD-BIOS-UMR AGAP (FRA)
• Beurier Grégory, CIRAD-BIOS-UMR AGAP (FRA)
• Marie-Magdeleine Carine, INRAE (FRA)
• Meghar Karima, CIRAD-PERSYST-UMR Qualisud (FRA)
• Alamu Emmanuel Oladeji, IITA (NGA)
• Otegbayo Bolanle Omolara, Bowen university (NGA)
• Cornet Denis, CIRAD-BIOS-UMR AGAP (FRA) ORCID: 0000-0001-9297-2680 - auteur correspondant

Source : Cirad-Agritrop (https://agritrop.cirad.fr/605417/)

Métriques

[ Page générée et mise en cache le 2024-12-18 ]