Application of Deep Transfer Learning in Plant Leaf Diseases Detection
Fuente:
Zenodo
Guardado en:
| Autores principales: | , , |
|---|---|
| Formato: | Recurso digital |
| Publicado: |
Zenodo
2025
|
| Materias: | |
| Acceso en línea: | |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| _version_ | 1866901183135744000 |
|---|---|
| author | Azlou Jamal Bah Ibrahima Mehdi Khamir |
| author_facet | Azlou Jamal Bah Ibrahima Mehdi Khamir |
| contents | Extraction of plant features, identification, and classification of unhealthy plants are critical for food security. This paper describes a Transfer Learning (TL) approach for identifying Plant Diseases (PD) that combines feature extraction and fine-tuning with classification using traditional Machine Learning (ML). Our work mainly focuses on comparing the classification of Plant Diseases using traditional Machine Learning classifiers on top of Deep Learning (DL) extracted features with MobileNet and using deep Transfer Learning fine-tuning techniques to do the classification task. For the traditional ML classifier training, Cross-Validation is used. However, the PlantVillage dataset was downsampled for all of the experimentations, focusing more on the utility of using TL. The augmentation technique has been used in this study to deal with the problem of overfitting. We concluded that applying Transfer Learning in conjunction with data augmentation produces the best results in classifying plant leaf diseases. |
| format | Recurso digital |
| id | zenodo_https___doi_org_10_5281_zenodo_18114725 |
| institution | Zenodo |
| language | |
| publishDate | 2025 |
| publisher | Zenodo |
| record_format | zenodo |
| spellingShingle | Application of Deep Transfer Learning in Plant Leaf Diseases Detection Azlou Jamal Bah Ibrahima Mehdi Khamir Plant Diseases (PD) Deep Learning (DL) Convolutional Neural Network (CNN) MobileNet Transfer Learning (TL) Logistic Regression (LR) Cross-Validation (CV). Extraction of plant features, identification, and classification of unhealthy plants are critical for food security. This paper describes a Transfer Learning (TL) approach for identifying Plant Diseases (PD) that combines feature extraction and fine-tuning with classification using traditional Machine Learning (ML). Our work mainly focuses on comparing the classification of Plant Diseases using traditional Machine Learning classifiers on top of Deep Learning (DL) extracted features with MobileNet and using deep Transfer Learning fine-tuning techniques to do the classification task. For the traditional ML classifier training, Cross-Validation is used. However, the PlantVillage dataset was downsampled for all of the experimentations, focusing more on the utility of using TL. The augmentation technique has been used in this study to deal with the problem of overfitting. We concluded that applying Transfer Learning in conjunction with data augmentation produces the best results in classifying plant leaf diseases. |
| title | Application of Deep Transfer Learning in Plant Leaf Diseases Detection |
| topic | Plant Diseases (PD) Deep Learning (DL) Convolutional Neural Network (CNN) MobileNet Transfer Learning (TL) Logistic Regression (LR) Cross-Validation (CV). |
| url | https://doi.org/10.5281/zenodo.18114725 |