University Sétif 1 FERHAT ABBAS Faculty of Sciences
Détail de l'auteur
Auteur Soulafa Chouarfa |
Documents disponibles écrits par cet auteur
Ajouter le résultat dans votre panier Affiner la recherche
Plant leaf Disease Classification Using Deep Learning Transformers Driven Bayesian Learning and Regularization. / Soulafa Chouarfa
Titre : Plant leaf Disease Classification Using Deep Learning Transformers Driven Bayesian Learning and Regularization. Type de document : texte imprimé Auteurs : Soulafa Chouarfa, Auteur ; Nesrine Bouchama, Auteur ; Abdelouahab Moussaoui, Directeur de thèse Année de publication : 2023 Importance : 1 vol (91 f .) Format : 29cm Langues : Français (fre) Catégories : Thèses & Mémoires:Informatique Mots-clés : Informatique Index. décimale : 004 Informatique Résumé : Countries depend on agriculture to help ensure food security for their citizens, but one
of the biggest challenges that countries can face is plant diseases, because they cause
the destruction of a large number of agricultural crops, which has a negative impact on
people like famines, as was the case with the famines in Ireland in 1845 and Bengal in 1943. So
early disease detection is necessary to enable farmers to maintain a healthy crop and increase
productivity. Therefore, in this paper We propose developing a deep learning models for the
automatic identification and classification of wheat and potato leaves diseases in order to make
accurate predictions. The current models, however, suffer from a number of problems thus we
suggest that enhancements be made by using a Bayesian learning and regularization. Bayesian
learning provides a principled framework for estimating uncertainty in predictions. Instead of
producing point estimates, Bayesian models can provide a distribution over possible outcomes.
This can be valuable in situations where understanding uncertainty or having probabilistic
predictions is important, such as plant diseases detection and classification in our case. The
proposed method has an accuracy rate of 99% for wheat plant and 99% for potato plant.Côte titre : MAI/0709 En ligne : https://drive.google.com/file/d/1-BP3jkVbHFa3zLpFUikjUlBZgE-bJYCW/view?usp=drive [...] Format de la ressource électronique : Plant leaf Disease Classification Using Deep Learning Transformers Driven Bayesian Learning and Regularization. [texte imprimé] / Soulafa Chouarfa, Auteur ; Nesrine Bouchama, Auteur ; Abdelouahab Moussaoui, Directeur de thèse . - 2023 . - 1 vol (91 f .) ; 29cm.
Langues : Français (fre)
Catégories : Thèses & Mémoires:Informatique Mots-clés : Informatique Index. décimale : 004 Informatique Résumé : Countries depend on agriculture to help ensure food security for their citizens, but one
of the biggest challenges that countries can face is plant diseases, because they cause
the destruction of a large number of agricultural crops, which has a negative impact on
people like famines, as was the case with the famines in Ireland in 1845 and Bengal in 1943. So
early disease detection is necessary to enable farmers to maintain a healthy crop and increase
productivity. Therefore, in this paper We propose developing a deep learning models for the
automatic identification and classification of wheat and potato leaves diseases in order to make
accurate predictions. The current models, however, suffer from a number of problems thus we
suggest that enhancements be made by using a Bayesian learning and regularization. Bayesian
learning provides a principled framework for estimating uncertainty in predictions. Instead of
producing point estimates, Bayesian models can provide a distribution over possible outcomes.
This can be valuable in situations where understanding uncertainty or having probabilistic
predictions is important, such as plant diseases detection and classification in our case. The
proposed method has an accuracy rate of 99% for wheat plant and 99% for potato plant.Côte titre : MAI/0709 En ligne : https://drive.google.com/file/d/1-BP3jkVbHFa3zLpFUikjUlBZgE-bJYCW/view?usp=drive [...] Format de la ressource électronique : Exemplaires (1)
Code-barres Cote Support Localisation Section Disponibilité MAI/0709 MAI/0709 Mémoire Bibliothéque des sciences Anglais Disponible
Disponible