AN EFFICIENT PLANT LEAF DISEASE DETECTION MODEL USING SHALLOW-CONVNET

In recent time, deep convolution neural networks have seen an exponential growth in their use in phytopathology. However, deep convolutional neural network needs a lot of processing power because of its intricate structure consisting of a large stack of layers. In this article, authors have introduced a novel lightweight sequential CNN architecture-Shallow-ConvNet for the diagnosis of leaf diseases. The suggested approach contains fewer layers and around 75% fewer attributes than pre-trained CNN-based approaches. For the experiments and performance evaluation, authors utilized a hybrid dataset consisting of 7166 images of tomato & potato with real field and laboratory-conditions affected with early and late blight diseases. The performance of the proposed architecture is compared against three recently priorly trained CNN architectures such as ResNet-50, VGG-16, and VGG-19. The average accuracy percentage reported by the proposed architecture is 97.22, and the time consumed in training is also much better. The experimental findings demonstrate that the suggested approach outperforms the recent existing trained CNN approaches and has a very less number of layers and parameters which significantly reduces the number of computing resources and time needed to train the model which could be a better choice for real-time plant disease diagnosis applications on resource constrained computing devices.