Exploring Transfer Learning Models for Multi-Class Classification of Infected Date Palm Leaves
Fagun Patel, Shubbh Mewada, Sheshang Degadwala, Dhairya Vyas
Abstract
This research explores the feasibility of leveraging transfer learning models, including AlexNet, VGGNet, and ResNet, alongside a newly devised Convolutional Neural Network (CNN) architecture, for the accurate multi-class classification of infected date palm leaves. The research harnesses the power of transfer learning by adapting these established models to the task of identifying various types of infections in date palm leaves. Alongside, a custom CNN is introduced, designed to capture intricate disease-related patterns and outperform generic architectures. Rigorous experimentation on a carefully annotated dataset enables a comprehensive performance comparison in terms of accuracy, precision, recall, F1 score, and computational efficiency. This investigation not only advances automated plant disease detection but also emphasizes the significance of specialized models in agricultural technology, contributing to sustainable food production and resource preservation.