Classification of tomato leaf disease based on leaf image using convolutional neural network algorithm

The Tomato plant is a horticultural commodity that is easily cultivated in the highlands and lowlands. Although easy to cultivate, Tomato plants are susceptible to disease, especially on the leaves. The emergence of early symptoms of Tomato leaf disease can be used as a form of prevention by farmers so that the disease is not contagious on healthy leaves. So it takes accuracy in seeing the form of leaf damage to be able to take the right decisions in preventing Tomato plants. One solution to this problem is to classify Tomato leaf diseases based on Tomato leaf images using the Convolutional Neural Network algorithm. The results of the experiment on the Tomato leaf disease classification process which was carried out by dividing the image into training data and testing data of 80% and 20% resulted in the best performance at epoch 50 epochs, batch sizes 16 and 95 at step per epoch. The accuracy produced in this study was 96.6%, and precision, recall, and F1-Score were 96.3%.