Classification Disease in Tomato Leaf Based on the Mix Convolution Neural Network

Tomatoes are among the most common crops we can find everywhere in all culinary backgrounds. Tomatoes diseases can seriously affect crop output and productivity. Therefore, diagnosing and detecting diseases is crucial to help farmers improve tomato crop productivity. The article uses deep learning, specifically convolutional neural networks, to detect and identify leaf diseases. We designed a MIX-CNN network with a mixed network structure and a small-size model, only 62.4K with 29 layers. The network design includes enhancing two consecutive maxpooling layers at the input and two fully connected layers at the output to mitigate overfitting. The results demonstrate that the proposed network effectively classified the leaf disease dataset acquired from the PlantVillage dataset, achieving an accuracy of 97.6%.