Tomato Plant Leaf Disease Detection Using Transfer Learning-based ResNet110

Tomato is the second most often grown vegetable. The agricultural industry is severely endangered by plant diseases and dangerous insects. A vast number of images of plant leaves are employed in the artificial intelligence-based deep learning technique, which is essential in disease identification. To maximize agricultural yield, it is critical to quickly and efficiently identify plant diseases. However, utilizing deep learning techniques is difficult to diagnose diseases with small datasets. In this paper, a transfer learning-based ResNet110 was proposed for quickly diagnosing tomato leaf disease using limited plant data images. The plant village datasets collected from the field consisting of healthy and infected leaves of tomatoes, and the architecture of the convolutional neural network VGG16, VGG17, and ResNet110 are employed to train classifiers. Each classifier's performance was assessed by evaluating its ability to correctly classified the tomato crop. The existing methods such as Conditional Generative Adversarial Network (C-GAN), a novel fusion of superpixel and color balancing-based approach, Attention Embedded Residual CNN, lightweight CNN model, and Deep-CNN are used to justify the effectiveness of Transfer learning-based ResNet110. The proposed method achieves better accuracy of 99.7% when compared to the existing methods.