University Sétif 1 FERHAT ABBAS Faculty of Sciences
Détail de l'auteur
Auteur Zakaria Bouzour |
Documents disponibles écrits par cet auteur
Ajouter le résultat dans votre panier Affiner la recherche
Vision Transformer Based Deep Learning Models for Plant Disease Detection and Diagnosis / Rayene Amina Boukabouya
Titre : Vision Transformer Based Deep Learning Models for Plant Disease Detection and Diagnosis Type de document : texte imprimé Auteurs : Rayene Amina Boukabouya, Auteur ; Zakaria Bouzour, Auteur ; Abdelouahab Moussaoui, Directeur de thèse Année de publication : 2022 Langues : Français (fre) Catégories : Thèses & Mémoires:Informatique Mots-clés : Deep learning
Vision TransformersIndex. décimale : 004 Informatique Résumé :
ne particular use of deep learning is classification. In this thesis, we want to
Compare multiple classification models based on classical neural networks and
vision Transformers to classify and identify different plant diseases at early
stages with high accuracy that outperform SOTA models. Plant health is one of the
most crucial things in a natural cycle, it needs to be preserved to maintain the life
of the individuals. The diagnosed at late stages, there is almost no chance to reverse
agricultural crops, which means the increase of lesions of hunger food in the world and
the farmers will lose many dollars, loss their time and their hard work. We propose
deep convolutional neural network architectures and vision transformers and tuned on
tomato and grape dataset, base model CNN, InceptionV3, VGG-16, CNN with Attention
and ViT. we compare our models for the tomatoes leaf and, base CNN, VGG16, Efficient3,
and InceptionV3 Vit for the second Dataset. The recent vision transformers model gives
the performance more than the previously published works using the same data for the
Tomatoes leaf, where we obtained an accuracy up to 99.7% and for the grape we achieved
98%.
This technology might help speed up the classification and treatment of treatable
illnesses, allowing for early treatment and better agricultural resultsCôte titre : MAI/0589 En ligne : https://drive.google.com/file/d/1luRodvYBcwPD52z6rtYFhSGLjHj0yzFN/view?usp=share [...] Format de la ressource électronique : Vision Transformer Based Deep Learning Models for Plant Disease Detection and Diagnosis [texte imprimé] / Rayene Amina Boukabouya, Auteur ; Zakaria Bouzour, Auteur ; Abdelouahab Moussaoui, Directeur de thèse . - 2022.
Langues : Français (fre)
Catégories : Thèses & Mémoires:Informatique Mots-clés : Deep learning
Vision TransformersIndex. décimale : 004 Informatique Résumé :
ne particular use of deep learning is classification. In this thesis, we want to
Compare multiple classification models based on classical neural networks and
vision Transformers to classify and identify different plant diseases at early
stages with high accuracy that outperform SOTA models. Plant health is one of the
most crucial things in a natural cycle, it needs to be preserved to maintain the life
of the individuals. The diagnosed at late stages, there is almost no chance to reverse
agricultural crops, which means the increase of lesions of hunger food in the world and
the farmers will lose many dollars, loss their time and their hard work. We propose
deep convolutional neural network architectures and vision transformers and tuned on
tomato and grape dataset, base model CNN, InceptionV3, VGG-16, CNN with Attention
and ViT. we compare our models for the tomatoes leaf and, base CNN, VGG16, Efficient3,
and InceptionV3 Vit for the second Dataset. The recent vision transformers model gives
the performance more than the previously published works using the same data for the
Tomatoes leaf, where we obtained an accuracy up to 99.7% and for the grape we achieved
98%.
This technology might help speed up the classification and treatment of treatable
illnesses, allowing for early treatment and better agricultural resultsCôte titre : MAI/0589 En ligne : https://drive.google.com/file/d/1luRodvYBcwPD52z6rtYFhSGLjHj0yzFN/view?usp=share [...] Format de la ressource électronique : Exemplaires (1)
Code-barres Cote Support Localisation Section Disponibilité MAI/0589 MAI/0589 Mémoire Bibliothéque des sciences Anglais Disponible
Disponible