DualNet - A Novel Deep Learning Ensemble Approach for Tomato Leaf Disease Classification

Abstract

Tomato cultivation is a vital component of global agriculture, and the health of tomato plants is susceptible to various leaf diseases that can significantly impact yield and quality. In this study, we propose an innovative approach for accurate tomato leaf disease classification using a deep learning ensemble model named “DualNet.” Our model combines the strengths of two prominent architectures, GoogleNet and ResNet50, to achieve superior classification accuracy. We present a comprehensive analysis of individual model performance, where GoogleNet achieved an accuracy of 96.080%, and ResNet50 achieved 98.492% accuracy. To enhance classification accuracy, we employ an ensemble approach by fusing the predictions of GoogleNet and ResNet50. The resulting DualNet achieves an impressive accuracy of 99%, showcasing the effectiveness of our ensemble strategy. Our model not only demonstrates exceptional accuracy but also offers a robust solution for real-world tomato leaf disease classification, aiding farmers and experts in making informed decisions for crop management. This research underscores the potential of ensemble deep-learning techniques in agricultural applications and contributes to the advancement of precision agriculture. The proposed DualNet model holds promise for broader applications in plant disease classification, highlighting the significance of combining diverse models to achieve remarkable results.