<![CDATA[Maize productivity growth is expected to increase by the year. However, there are obstacles to achieving it. One of the causes is diseases attack. Generally, maize plant diseases are easily detected through the leaves. This article discusses maize leaf disease classification using computer vision with a convolutional neural network (CNN). It aims to compare the deep convolutional neural network (CNN) AlexNet and Squeezenet. The network also used optimization, stochastic gradient descent with momentum (SGDM). The dataset for this experiment was taken from PlantVillage with 3852 images with 4 classes i.e healthy, blight, spot, and rust. The data is divided into 3 parts: training, validation, and testing. Training and validation are 80%, the rest for testing. The results of training with cross-validation produce the best accuracy of 100% for AlexNet and Squeezenet. Furthermore, the best weights and biases are stored in the model for testing data classification. The recognition results using AlexNet showed 97.69% accuracy. While the results of Squeezenet 44.49% accuracy. From this experiment environment, it can be concluded that AlexNet is better than Squeezenet for maize leaf diseases classification.]]>
Copyrights © 2021
