Research of the Effectiveness of Convolutional Neural Networks by the Example of Determining the Gait of a Horse

摘要

We focused this article on the research on the efficiency of convolution neural networks using the example of solving the problem of horse gait classification. A theoretical research of the structure and principle of operation of convolution neural networks was carried out. The paper deals with methods for improving the accuracy and loss values of a neural network. Methods for assembling a data set for deep learning have been also researched. The input data are video converted to images that are sent to the neural network. An experiment was conducted to create a convolution neural network and analysis of effectiveness was used in the formulated problem.