Publication: Automatic cattle identification based on fusion of texture features extracted from muzzle images
Issued Date
2018-04-27
Resource Type
Other identifier(s)
2-s2.0-85046942386
Rights
Mahidol University
Rights Holder(s)
SCOPUS
Bibliographic Citation
Proceedings of the IEEE International Conference on Industrial Technology. Vol.2018-February, (2018), 1484-1489
Suggested Citation
Worapan Kusakunniran, Anuwat Wiratsudakul, Udom Chuachan, Sarattha Kanchanapreechakorn, Thanandon Imaromkul Automatic cattle identification based on fusion of texture features extracted from muzzle images. Proceedings of the IEEE International Conference on Industrial Technology. Vol.2018-February, (2018), 1484-1489. doi:10.1109/ICIT.2018.8352400 Retrieved from: https://repository.li.mahidol.ac.th/handle/20.500.14594/45638
Research Projects
Organizational Units
Authors
Journal Issue
Thesis
Title
Automatic cattle identification based on fusion of texture features extracted from muzzle images
Other Contributor(s)
Abstract
© 2018 IEEE. Biometrics have been widely used for human identification, including fingerprint, face and iris because they cannot be easily duplicated. Recently, biometrics have been also used for animal (i.e. cattle in this paper) identification. Individual cattle identification is necessary for many important reasons including determining legal ownership, verifying transferred source, and implementing disease surveillance and control. Popular traditional methods for individual cattle identification are using plastic ear tags or microchips. However, the tag can be deduplicated and it can be dangerous and takes time for the human expert to place the microchip in the cattle. Also, it may hurt the cattle. Thus, in this paper, muzzle print is used as a biometric for automatic cattle identification. The fusion of texture features extracted from the muzzle image is used to represent individual cattle. They are Gabor feature and Local Binary Pattern (LBP) histogram. Gabor features were extracted at the different scales and orientations in specific frequencies, while LBP histogram was extracted for each local sub-image to preserve local spatial textures. Then, Support Vector Machine (SVM) is employed as a classifier. The proposed method is reported with the perfect accuracy.