Near real-time animal action recognition and classification
In computer vision, identification of actions of an object is considered as a complex and relevant task. When solving the problem, one requires information on the position of key points of the object. Training models that determine the position of key points requires a large amount of data, includin...
| Main Authors: | , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Samara National Research University
2023-04-01
|
| Series: | Компьютерная оптика |
| Subjects: | |
| Online Access: | https://computeroptics.ru/eng/KO/Annot/KO47-2/470211e.html |
| _version_ | 1797363972362469376 |
|---|---|
| author | A.D. Egorov M.S. Reznik |
| author_facet | A.D. Egorov M.S. Reznik |
| author_sort | A.D. Egorov |
| collection | DOAJ |
| description | In computer vision, identification of actions of an object is considered as a complex and relevant task. When solving the problem, one requires information on the position of key points of the object. Training models that determine the position of key points requires a large amount of data, including information on the position of these key points. Due to the lack of data for training, the paper provides a method for obtaining additional data for training, as well as an algorithm that allows highly accurate recognition of animal actions based on a small number of data. The achieved accuracy of determining the key points positions within a test sample is 92%. Positions of the key points define the action of the object. Various approaches to classifying actions by key points are compared. The accuracy of identifying the action of the object in the image reaches 72.9 %. |
| first_indexed | 2024-03-08T16:28:40Z |
| format | Article |
| id | doaj.art-ce2b58dc957c486d85541dfd46a003ab |
| institution | Directory Open Access Journal |
| issn | 0134-2452 2412-6179 |
| language | English |
| last_indexed | 2024-03-08T16:28:40Z |
| publishDate | 2023-04-01 |
| publisher | Samara National Research University |
| record_format | Article |
| series | Компьютерная оптика |
| spelling | doaj.art-ce2b58dc957c486d85541dfd46a003ab2024-01-06T11:11:40ZengSamara National Research UniversityКомпьютерная оптика0134-24522412-61792023-04-0147227828610.18287/2412-6179-CO-1138Near real-time animal action recognition and classificationA.D. Egorov0M.S. Reznik1National Research Nuclear University MEPhINational Research Nuclear University MEPhIIn computer vision, identification of actions of an object is considered as a complex and relevant task. When solving the problem, one requires information on the position of key points of the object. Training models that determine the position of key points requires a large amount of data, including information on the position of these key points. Due to the lack of data for training, the paper provides a method for obtaining additional data for training, as well as an algorithm that allows highly accurate recognition of animal actions based on a small number of data. The achieved accuracy of determining the key points positions within a test sample is 92%. Positions of the key points define the action of the object. Various approaches to classifying actions by key points are compared. The accuracy of identifying the action of the object in the image reaches 72.9 %.https://computeroptics.ru/eng/KO/Annot/KO47-2/470211e.htmlcomputer visionmachine learninganimal recognitionaction recognitiondata augmentationkeypoint r-cnnmobile net |
| spellingShingle | A.D. Egorov M.S. Reznik Near real-time animal action recognition and classification Компьютерная оптика computer vision machine learning animal recognition action recognition data augmentation keypoint r-cnn mobile net |
| title | Near real-time animal action recognition and classification |
| title_full | Near real-time animal action recognition and classification |
| title_fullStr | Near real-time animal action recognition and classification |
| title_full_unstemmed | Near real-time animal action recognition and classification |
| title_short | Near real-time animal action recognition and classification |
| title_sort | near real time animal action recognition and classification |
| topic | computer vision machine learning animal recognition action recognition data augmentation keypoint r-cnn mobile net |
| url | https://computeroptics.ru/eng/KO/Annot/KO47-2/470211e.html |
| work_keys_str_mv | AT adegorov nearrealtimeanimalactionrecognitionandclassification AT msreznik nearrealtimeanimalactionrecognitionandclassification |