USAGE OF NEURAL NETWORK FOR IMAGE CLASSIFICATION
V.S. DOLGOPYATOV, E.V. MORAR · Applied Mathematics and Fundamental Informatics · 2025
The paper provides a brief theoretical overview of the features of a popular method for solving the problem of image classification in the form of neural networks. A suitable dataset with images of plant diseases has been found and described. Specialized layers of a convolutional neural network are considered, as well as a class weighting method to solve the problem of class imbalance in the training sample. As a result of the study, a small but fairly accurate convolutional neural network was created and trained from a set of convolutional and fully connected layers to solve the problem of classifying types of plant diseases by image.