Comparative Analysis of Pre-Trained CNN and Feature Selection Algorithms for Plant Leaf Disease Identification

Plant disease identification through leaves is an important aspect of crop health management for achieving the highest productivity. This research work compares the use of pre-trained Convolutional Neural Networks (CNNs) and feature selection algorithms for identifying plant leaf diseases. Pre-trained CNNs are evaluated against traditional classifiers. Their performance is enhanced using feature selection methods such as Genetic Algorithm and Recursive Feature Elimination. The study evaluates the model's performance using key metrics, including accuracy, precision, recall, F1-score and computational efficiency on a dataset consist of healthy and unhealthy leaf images. The results show that pre-trained CNNs offer high accuracy. Meanwhile, feature selection algorithms deliver competitive performance with reduced model complexity, providing valuable insights for agricultural technology applications.