Firefly Algorithm-Integrated Convolutional Neural Network for Advanced Detection and Classification of Tomato Leaf Disease

Abstract

Effective identification and categorization of diseases in tomato plant leaves are critical for enhancing crop productivity and quality. This study introduces an advanced FiFy-CNN model, combining Convolutional Neural Networks (CNN) with Firefly Optimization, to automate and optimize the diagnosis of prevalent tomato leaf diseases. FiFy-CNN leverages Firefly Optimization to fine-tune CNN learning rates, and extracts texture features using the Grey Level Co-occurrence Matrix (GLCM), resulting in superior performance in metrics like accuracy, sensitivity, specificity, and precision compared to traditional methods. A comprehensive image dataset of tomato diseases, including Mosaic virus, Yellow Leaf Curl virus, and early blight, was used for evaluation, highlighting FiFy-CNN’s accuracy in disease detection. Experimental findings demonstrate that this framework outperforms conventional CNN models and architectures such as VGG16 and GoogLeNet, offering a reliable and efficient solution for automated plant disease monitoring—marking a substantial advancement for precision agriculture.