Deep Transfer Learning for Automatic Plant Species Recognition.

Ouahab Abdelwhab; Lazreg Taibaoui; Boubakeur Zegnini

doi:10.5935/jetia.v11i51.1464

Authors

Ouahab Abdelwhab University of Adrar http://orcid.org/0000-0003-0648-2947 (unauthenticated)
Lazreg Taibaoui Department of Electrical Engineering, Laboratoire d'Etudes et Développement des Matériaux Semi-Conducteurs et Diélectriques, Amar Telidji University of Laghouat, BP 37 G, Route de Ghardaïa, Laghouat03000, Algeria https://orcid.org/0000-0002-1598-4546 (unauthenticated)
Boubakeur Zegnini Department of Electrical Engineering, Laboratoire d'Etudes et Développement des Matériaux Semi-Conducteurs et Diélectriques, Amar Telidji University of Laghouat, BP 37 G, Route de Ghardaïa, Laghouat03000, Algeria https://orcid.org/0000-0003-0937-188X (unauthenticated)

DOI:

https://doi.org/10.5935/jetia.v11i51.1464

Abstract

Image processing has emerged as a promising tool for plant species recognition, allowing individuals to capture images with their mobile phones in the field and identify plant species or a list of closely related plants. Deep learning, particularly Convolutional Neural Networks (CNNs), has become the leading approach in image recognition tasks. This study explores the use of transfer learning, a deep learning technique, for automatic plant species recognition. Transfer learning involves using pre-trained CNN models, originally trained on large datasets like ImageNet, and fine-tuning them for specific tasks with smaller datasets. In this research, six pre-trained CNN models—VGG16, VGG19, DenseNet121, InceptionResNetV2, MobileNet, and MobileNetV2—were evaluated on a dataset comprising 30 plant species. The goal is to determine which transfer learning model performs best for plant species recognition.