Search
Abstract
The aim of this article is to present a neural system based on stock architecture for recognizing emotional behavior in dogs. Our considerations are inspired by the original work of Franzoni et al. on recognizing dog emotions. An appropriate set of photographic data has been compiled taking into account five classes of emotional behavior in dogs of one breed, including joy, anger, licking, yawning, and sleeping. Focusing on a particular breed makes it easier to experiment and recognize the emotional behavior of dogs. To broaden our conclusions, in our research study we compare our system with other systems of different architectures. In addition, we also use modern transfer learning with augmentation and data normalization techniques. The results show that VGG16 and VGG19 are the most suitable backbone networks. Therefore, a certain deep neural network, named mVGG16, based on the suboptimal VGG16 has been created, trained and fine-tuned with transfer (without augmentation and normalization). The developed system is then tested against an internal test dataset. In addition, to show the robustness of the system, a set of external data outside the breed is also taken into account. Being able to detect unsafe dog behavior and rely on a generalization for other breeds is worth popularizing. Equally important are the possible applications of the system to monitor the behavior of pets in the absence of their owners.
Citations
-
2
CrossRef
-
0
Web of Science
-
2
Scopus
Authors (3)
Cite as
Full text
- Publication version
- Accepted or Published Version
- DOI:
- Digital Object Identifier (open in new tab) 10.1111/coin.12559
- License
- Copyright (2022 Wiley Periodicals LLC)
Keywords
Details
- Category:
- Articles
- Type:
- artykuły w czasopismach
- Published in:
-
COMPUTATIONAL INTELLIGENCE
no. 38,
pages 2116 - 2133,
ISSN: 0824-7935 - Language:
- English
- Publication year:
- 2022
- Bibliographic description:
- Kowalczuk Z., Czubenko M., Żmuda-Trzebiatowska W.: Categorization of emotions in dog behavior based on the deep neural network// COMPUTATIONAL INTELLIGENCE -Vol. 38,iss. 6 (2022), s.2116-2133
- DOI:
- Digital Object Identifier (open in new tab) 10.1111/coin.12559
- Sources of funding:
-
- Free publication
- Verified by:
- Gdańsk University of Technology
seen 97 times