Artificial intelligence and machine learning

Research Article

DOI 10.25209/2079-3316-2024-15-4-79-96

Neural network classification of videos based on a small number of frames

Alexander Vladimirovich Smirnov¹, Dmitry Denisovich Parfenov², Igor Petrovich Tishchenko³

1,3	Ailamazyan Program Systems Institute of RAS, Ves'kovo, Russia
2	Admiral Makarov State University of Maritime and Inland Shipping, St. Petersburg, Russia
1	asmirnov_1991@mail.ru

Abstract. The article proposes a method for neural network classification of short videos. The classification problem is considered from the point of view of reducing the number of operations required to categorize videos. The proposed solution consists of using a small number of frames (no more than 10) to perform classification using the lightest neural network architecture of the ResNet family of models. As part of the work, a proprietary training dataset was created, consisting of three classes: “animals”, “cars” and “people”. As a result, a classification accuracy of 79% was obtained, a database of classified videos was formed, and an application with GUI elements was developed for interacting with the classifier and viewing the results. (In Russian).

Keywords: Video classification, dataset, neural networks, graphical user interface

MSC-2020

68T10; 68T45

For citation: Alexander V. Smirnov, Dmitry D. Parfenov, Igor P. Tishchenko. Neural network classification of videos based on a small number of frames. Program Systems: Theory and Applications, 2024, 15:4, pp. 79–96. (In Russ.). https://psta.psiras.ru/2024/4_79-96.

Full text of article (PDF): https://psta.psiras.ru/read/psta2024_4_79-96.pdf.

The article was submitted 01.10.2024; approved after reviewing 23.10.2024; accepted for publication 04.11.2024; published online 20.11.2024.