Multi-Center Agent Loss for Visual Identification of Chinese Simmental in the Wild
Visual identification of cattle in the wild provides an essential way for real-time cattle monitoring applicable to precision livestock farming. Chinese Simmental exhibit a yellow or brown coat with individually characteristic white stripes or spots, which makes a biometric identifier for identifica...
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2022-02-01
|
| Series: | Animals |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2076-2615/12/4/459 |
| _version_ | 1797483458652536832 |
|---|---|
| author | Jianmin Zhao Qiusheng Lian Neal N. Xiong |
| author_facet | Jianmin Zhao Qiusheng Lian Neal N. Xiong |
| author_sort | Jianmin Zhao |
| collection | DOAJ |
| description | Visual identification of cattle in the wild provides an essential way for real-time cattle monitoring applicable to precision livestock farming. Chinese Simmental exhibit a yellow or brown coat with individually characteristic white stripes or spots, which makes a biometric identifier for identification possible. This work employed the observable biometric characteristics to perform cattle identification with an image from any viewpoint. We propose multi-center agent loss to jointly supervise the learning of DCNNs by SoftMax with multiple centers and the agent triplet. We reformulated SoftMax with multiple centers to reduce intra-class variance by offering more centers for feature clustering. Then, we utilized the agent triplet, which consisted of the features and the agents, to enforce separation among different classes. As there are no datasets for the identification of cattle with multi-view images, we created CNSID100, consisting of 11,635 images from 100 Chinese Simmental identities. Our proposed loss was comprehensively compared with several well-known losses on CNSID100 and OpenCows2020 and analyzed in an engineering application in the farming environment. It was encouraging to find that our approach outperformed the state-of-the-art models on the datasets above. The engineering application demonstrated that our pipeline with detection and recognition is promising for continuous cattle identification in real livestock farming scenarios. |
| first_indexed | 2024-03-09T22:48:17Z |
| format | Article |
| id | doaj.art-3f70697c7f69413f9902acfb465092ef |
| institution | Directory Open Access Journal |
| issn | 2076-2615 |
| language | English |
| last_indexed | 2024-03-09T22:48:17Z |
| publishDate | 2022-02-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Animals |
| spelling | doaj.art-3f70697c7f69413f9902acfb465092ef2023-11-23T18:25:26ZengMDPI AGAnimals2076-26152022-02-0112445910.3390/ani12040459Multi-Center Agent Loss for Visual Identification of Chinese Simmental in the WildJianmin Zhao0Qiusheng Lian1Neal N. Xiong2Institute of Information Science and Technology, Yanshan University, Qinhuangdao 066004, ChinaInstitute of Information Science and Technology, Yanshan University, Qinhuangdao 066004, ChinaDepartment of C.S., Colorado Technical University, Colorado Springs, CO 80907, USAVisual identification of cattle in the wild provides an essential way for real-time cattle monitoring applicable to precision livestock farming. Chinese Simmental exhibit a yellow or brown coat with individually characteristic white stripes or spots, which makes a biometric identifier for identification possible. This work employed the observable biometric characteristics to perform cattle identification with an image from any viewpoint. We propose multi-center agent loss to jointly supervise the learning of DCNNs by SoftMax with multiple centers and the agent triplet. We reformulated SoftMax with multiple centers to reduce intra-class variance by offering more centers for feature clustering. Then, we utilized the agent triplet, which consisted of the features and the agents, to enforce separation among different classes. As there are no datasets for the identification of cattle with multi-view images, we created CNSID100, consisting of 11,635 images from 100 Chinese Simmental identities. Our proposed loss was comprehensively compared with several well-known losses on CNSID100 and OpenCows2020 and analyzed in an engineering application in the farming environment. It was encouraging to find that our approach outperformed the state-of-the-art models on the datasets above. The engineering application demonstrated that our pipeline with detection and recognition is promising for continuous cattle identification in real livestock farming scenarios.https://www.mdpi.com/2076-2615/12/4/459cattle identificationdeep convolutional neural networks (DCNNs)deep metric learning (DML)open-set recognitionprecision livestock farming |
| spellingShingle | Jianmin Zhao Qiusheng Lian Neal N. Xiong Multi-Center Agent Loss for Visual Identification of Chinese Simmental in the Wild Animals cattle identification deep convolutional neural networks (DCNNs) deep metric learning (DML) open-set recognition precision livestock farming |
| title | Multi-Center Agent Loss for Visual Identification of Chinese Simmental in the Wild |
| title_full | Multi-Center Agent Loss for Visual Identification of Chinese Simmental in the Wild |
| title_fullStr | Multi-Center Agent Loss for Visual Identification of Chinese Simmental in the Wild |
| title_full_unstemmed | Multi-Center Agent Loss for Visual Identification of Chinese Simmental in the Wild |
| title_short | Multi-Center Agent Loss for Visual Identification of Chinese Simmental in the Wild |
| title_sort | multi center agent loss for visual identification of chinese simmental in the wild |
| topic | cattle identification deep convolutional neural networks (DCNNs) deep metric learning (DML) open-set recognition precision livestock farming |
| url | https://www.mdpi.com/2076-2615/12/4/459 |
| work_keys_str_mv | AT jianminzhao multicenteragentlossforvisualidentificationofchinesesimmentalinthewild AT qiushenglian multicenteragentlossforvisualidentificationofchinesesimmentalinthewild AT nealnxiong multicenteragentlossforvisualidentificationofchinesesimmentalinthewild |