PB-STR: A spatiotemporal transformer network for multi-behavior recognition of pigs
Pig behavior is a reliable indicator of health status, accurate recognition is vital for effective health surveillance and management. This study proposes PB-STR, a behavior recognition model based on the integration of video spatiotemporal feature fusion. The model addresses challenges in recognizi...
Saved in:
| Main Authors: | , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Elsevier
2025-12-01
|
| Series: | Smart Agricultural Technology |
| Subjects: | |
| Online Access: | http://www.sciencedirect.com/science/article/pii/S2772375525003636 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1849472686218018816 |
|---|---|
| author | Yufan Hu Xiaobo Wang Rui Mao Yusen Guo Xianyao Zhu Meili Wang |
| author_facet | Yufan Hu Xiaobo Wang Rui Mao Yusen Guo Xianyao Zhu Meili Wang |
| author_sort | Yufan Hu |
| collection | DOAJ |
| description | Pig behavior is a reliable indicator of health status, accurate recognition is vital for effective health surveillance and management. This study proposes PB-STR, a behavior recognition model based on the integration of video spatiotemporal feature fusion. The model addresses challenges in recognizing multiple behaviors within a single frame and handling dynamically changing behaviors. It develops a Time Series Prediction Module (UnetTSF) and a Context Anchor Attention (CAA) module, enhancing the PB-STR framework's ability to capture feature evolution over time and fully utilize contextual information. To enhance the model's proficiency in detecting and recognizing behaviors within overlapping regions, the detection head employs Minimum Points Distance Intersection over Union (MPDIoU) as its bounding box loss function, improving adaptability to variations in pig positions. The PB-STR model was evaluated on a proprietary dataset of 294 videos covering seven pig behaviors. With a mean Average Precision of 94.2 %, recall of 90.8 %, and precision of 87.5 %, the PB-STR model can concurrently recognize five dynamic and two static behaviors in pigs. By outperforming models such as DETR, DAB-DETR, Deformable DETR, CenterNet, and DINO, the proposed approach not only enhances detection accuracy but also serves as a technological foundation for intelligent, welfare-oriented pig farming, facilitating in the sector's modernization. |
| format | Article |
| id | doaj-art-86ff077ba13d4e93b356ef7ecf88c4cf |
| institution | Kabale University |
| issn | 2772-3755 |
| language | English |
| publishDate | 2025-12-01 |
| publisher | Elsevier |
| record_format | Article |
| series | Smart Agricultural Technology |
| spelling | doaj-art-86ff077ba13d4e93b356ef7ecf88c4cf2025-08-20T03:24:26ZengElsevierSmart Agricultural Technology2772-37552025-12-011210113110.1016/j.atech.2025.101131PB-STR: A spatiotemporal transformer network for multi-behavior recognition of pigsYufan Hu0Xiaobo Wang1Rui Mao2Yusen Guo3Xianyao Zhu4Meili Wang5College of Information Engineering, Northwest A&F University, Yangling 712100, ChinaCollege of Information Engineering, Northwest A&F University, Yangling 712100, ChinaCollege of Information Engineering, Northwest A&F University, Yangling 712100, China; Shaanxi Engineering Research Center of Agriculture Information Intelligent Perception and Analysis, Yangling 712100, China; Corresponding author.College of Information Engineering, Northwest A&F University, Yangling 712100, ChinaCollege of Information Engineering, Northwest A&F University, Yangling 712100, ChinaCollege of Information Engineering, Northwest A&F University, Yangling 712100, China; Shaanxi Engineering Research Center of Agriculture Information Intelligent Perception and Analysis, Yangling 712100, ChinaPig behavior is a reliable indicator of health status, accurate recognition is vital for effective health surveillance and management. This study proposes PB-STR, a behavior recognition model based on the integration of video spatiotemporal feature fusion. The model addresses challenges in recognizing multiple behaviors within a single frame and handling dynamically changing behaviors. It develops a Time Series Prediction Module (UnetTSF) and a Context Anchor Attention (CAA) module, enhancing the PB-STR framework's ability to capture feature evolution over time and fully utilize contextual information. To enhance the model's proficiency in detecting and recognizing behaviors within overlapping regions, the detection head employs Minimum Points Distance Intersection over Union (MPDIoU) as its bounding box loss function, improving adaptability to variations in pig positions. The PB-STR model was evaluated on a proprietary dataset of 294 videos covering seven pig behaviors. With a mean Average Precision of 94.2 %, recall of 90.8 %, and precision of 87.5 %, the PB-STR model can concurrently recognize five dynamic and two static behaviors in pigs. By outperforming models such as DETR, DAB-DETR, Deformable DETR, CenterNet, and DINO, the proposed approach not only enhances detection accuracy but also serves as a technological foundation for intelligent, welfare-oriented pig farming, facilitating in the sector's modernization.http://www.sciencedirect.com/science/article/pii/S2772375525003636Pig behavior recognitionDeep learningSpatiotemporal transformer networkPB-STR |
| spellingShingle | Yufan Hu Xiaobo Wang Rui Mao Yusen Guo Xianyao Zhu Meili Wang PB-STR: A spatiotemporal transformer network for multi-behavior recognition of pigs Smart Agricultural Technology Pig behavior recognition Deep learning Spatiotemporal transformer network PB-STR |
| title | PB-STR: A spatiotemporal transformer network for multi-behavior recognition of pigs |
| title_full | PB-STR: A spatiotemporal transformer network for multi-behavior recognition of pigs |
| title_fullStr | PB-STR: A spatiotemporal transformer network for multi-behavior recognition of pigs |
| title_full_unstemmed | PB-STR: A spatiotemporal transformer network for multi-behavior recognition of pigs |
| title_short | PB-STR: A spatiotemporal transformer network for multi-behavior recognition of pigs |
| title_sort | pb str a spatiotemporal transformer network for multi behavior recognition of pigs |
| topic | Pig behavior recognition Deep learning Spatiotemporal transformer network PB-STR |
| url | http://www.sciencedirect.com/science/article/pii/S2772375525003636 |
| work_keys_str_mv | AT yufanhu pbstraspatiotemporaltransformernetworkformultibehaviorrecognitionofpigs AT xiaobowang pbstraspatiotemporaltransformernetworkformultibehaviorrecognitionofpigs AT ruimao pbstraspatiotemporaltransformernetworkformultibehaviorrecognitionofpigs AT yusenguo pbstraspatiotemporaltransformernetworkformultibehaviorrecognitionofpigs AT xianyaozhu pbstraspatiotemporaltransformernetworkformultibehaviorrecognitionofpigs AT meiliwang pbstraspatiotemporaltransformernetworkformultibehaviorrecognitionofpigs |