A Novel CNN Approach for Accurate Tomato Disease Classification
Ovidiu Cosma, Laura Cosma

Abstract. Plant diseases pose significant threats to agriculture, leading to substantial yield losses worldwide. Tomato crops are particularly affected by a wide range of pathogens, making accurate and timely disease detection critical for sustainable production. This paper proposes a novel convolutional neural network (CNN) approach for tomato disease classification, trained and tested on an image dataset featuring prevalent tomato diseases from the Plant Village dataset. The proposed architecture is designed to balance high classification accuracy with reduced computational requirements, making it suitable for deployment on resource-constrained devices used in smart agriculture, such as edge devices and mobile applications. Experimental results demonstrate the effectiveness of the model in distinguishing between different tomato diseases, outperforming several established CNN baselines in both accuracy and efficiency.

Keywords: convolutional neural networks; tomato disease classification; deep learning; smart agriculture; image classification

Ovidiu Cosma, Laura Cosma, "A Novel CNN Approach for Accurate Tomato Disease Classification", Proc. 30th ICE/IEEE ITMC Conf., IEEE, 2024, 2023. DOI: https://doi.org/10.1109/ICE/ITMC61926.2024.10794256.

Reference: Proc. 30th ICE/IEEE ITMC Conf., IEEE, 2024. DOI: 10.1109/ICE/ITMC61926.2024.10794256

Other publications