TSML: A New Pig Behavior Recognition Method Based on Two-Stream Mutual Learning Network
Changes in pig behavior are crucial information in the livestock breeding process, and automatic pig behavior recognition is a vital method for improving pig welfare. However, most methods for pig behavior recognition rely on human observation and deep learning. Human observation is often time-consu...
| Main Authors: | , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2023-05-01
|
| Series: | Sensors |
| Subjects: | |
| Online Access: | https://www.mdpi.com/1424-8220/23/11/5092 |
| _version_ | 1797596766423482368 |
|---|---|
| author | Wangli Hao Kai Zhang Li Zhang Meng Han Wangbao Hao Fuzhong Li Guoqiang Yang |
| author_facet | Wangli Hao Kai Zhang Li Zhang Meng Han Wangbao Hao Fuzhong Li Guoqiang Yang |
| author_sort | Wangli Hao |
| collection | DOAJ |
| description | Changes in pig behavior are crucial information in the livestock breeding process, and automatic pig behavior recognition is a vital method for improving pig welfare. However, most methods for pig behavior recognition rely on human observation and deep learning. Human observation is often time-consuming and labor-intensive, while deep learning models with a large number of parameters can result in slow training times and low efficiency. To address these issues, this paper proposes a novel deep mutual learning enhanced two-stream pig behavior recognition approach. The proposed model consists of two mutual learning networks, which include the red–green–blue color model (RGB) and flow streams. Additionally, each branch contains two student networks that learn collaboratively to effectively achieve robust and rich appearance or motion features, ultimately leading to improved recognition performance of pig behaviors. Finally, the results of RGB and flow branches are weighted and fused to further improve the performance of pig behavior recognition. Experimental results demonstrate the effectiveness of the proposed model, which achieves state-of-the-art recognition performance with an accuracy of 96.52%, surpassing other models by 2.71%. |
| first_indexed | 2024-03-11T02:57:39Z |
| format | Article |
| id | doaj.art-618567644a48460a8ad318f91addb5f9 |
| institution | Directory Open Access Journal |
| issn | 1424-8220 |
| language | English |
| last_indexed | 2024-03-11T02:57:39Z |
| publishDate | 2023-05-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Sensors |
| spelling | doaj.art-618567644a48460a8ad318f91addb5f92023-11-18T08:32:25ZengMDPI AGSensors1424-82202023-05-012311509210.3390/s23115092TSML: A New Pig Behavior Recognition Method Based on Two-Stream Mutual Learning NetworkWangli Hao0Kai Zhang1Li Zhang2Meng Han3Wangbao Hao4Fuzhong Li5Guoqiang Yang6School of Software, Shanxi Agricultural University, Jinzhong 030801, ChinaSchool of Software, Shanxi Agricultural University, Jinzhong 030801, ChinaSchool of Software, Shanxi Agricultural University, Jinzhong 030801, ChinaSchool of Software, Shanxi Agricultural University, Jinzhong 030801, ChinaYuncheng National Jinnan Cattle Genetic Resources and Gene Protection Center, Yongji 044099, ChinaSchool of Software, Shanxi Agricultural University, Jinzhong 030801, ChinaSchool of Software, Shanxi Agricultural University, Jinzhong 030801, ChinaChanges in pig behavior are crucial information in the livestock breeding process, and automatic pig behavior recognition is a vital method for improving pig welfare. However, most methods for pig behavior recognition rely on human observation and deep learning. Human observation is often time-consuming and labor-intensive, while deep learning models with a large number of parameters can result in slow training times and low efficiency. To address these issues, this paper proposes a novel deep mutual learning enhanced two-stream pig behavior recognition approach. The proposed model consists of two mutual learning networks, which include the red–green–blue color model (RGB) and flow streams. Additionally, each branch contains two student networks that learn collaboratively to effectively achieve robust and rich appearance or motion features, ultimately leading to improved recognition performance of pig behaviors. Finally, the results of RGB and flow branches are weighted and fused to further improve the performance of pig behavior recognition. Experimental results demonstrate the effectiveness of the proposed model, which achieves state-of-the-art recognition performance with an accuracy of 96.52%, surpassing other models by 2.71%.https://www.mdpi.com/1424-8220/23/11/5092pig breedingbehavior recognitioncomputer visiontwo stream mutual learninganimal welfare |
| spellingShingle | Wangli Hao Kai Zhang Li Zhang Meng Han Wangbao Hao Fuzhong Li Guoqiang Yang TSML: A New Pig Behavior Recognition Method Based on Two-Stream Mutual Learning Network Sensors pig breeding behavior recognition computer vision two stream mutual learning animal welfare |
| title | TSML: A New Pig Behavior Recognition Method Based on Two-Stream Mutual Learning Network |
| title_full | TSML: A New Pig Behavior Recognition Method Based on Two-Stream Mutual Learning Network |
| title_fullStr | TSML: A New Pig Behavior Recognition Method Based on Two-Stream Mutual Learning Network |
| title_full_unstemmed | TSML: A New Pig Behavior Recognition Method Based on Two-Stream Mutual Learning Network |
| title_short | TSML: A New Pig Behavior Recognition Method Based on Two-Stream Mutual Learning Network |
| title_sort | tsml a new pig behavior recognition method based on two stream mutual learning network |
| topic | pig breeding behavior recognition computer vision two stream mutual learning animal welfare |
| url | https://www.mdpi.com/1424-8220/23/11/5092 |
| work_keys_str_mv | AT wanglihao tsmlanewpigbehaviorrecognitionmethodbasedontwostreammutuallearningnetwork AT kaizhang tsmlanewpigbehaviorrecognitionmethodbasedontwostreammutuallearningnetwork AT lizhang tsmlanewpigbehaviorrecognitionmethodbasedontwostreammutuallearningnetwork AT menghan tsmlanewpigbehaviorrecognitionmethodbasedontwostreammutuallearningnetwork AT wangbaohao tsmlanewpigbehaviorrecognitionmethodbasedontwostreammutuallearningnetwork AT fuzhongli tsmlanewpigbehaviorrecognitionmethodbasedontwostreammutuallearningnetwork AT guoqiangyang tsmlanewpigbehaviorrecognitionmethodbasedontwostreammutuallearningnetwork |