Cow face identification based on CNN by using Channel Attention Module and Spatial Attention Module

Chengyun Liu1,+ , Feiyang Zhao1,+, Boya Huang1, Xintong Zhang1,2, Dequan Zhang1,2 and Hualin Li1,2,*
1School of Software and Artificial Intelligence, Chongqing Institute of Engineering, Chongqing, China.
2Shu Yi Xin Credit Management Co., LTD, Chongqing, China.
+ These authors contributed equally to this work and should be considered co-first authors.
* Corresponding author: Hualin Li. Tel.: 19968065404; email: hualinli@hotmail.com

 
Abstract. Cow face identification plays a crucial role in the cattle management system. Previous studies have primarily focused on radio frequency identification, and there is still a lack of comprehensive research on cow face identification. In this paper, instead of solely extracting features from individual images, we have constructed datasets for cow face identification. The datasets includes the facial images of all-black cow, all-white cow and mixed black-and-white cow. We apply convolutional neural networks method by utilizing ResNet backbone architectures, and additionally we incorporate different loss functions and attention modules to enhance the model's capacity. The results demonstrate that our methods have achieved the identification accuracy rate of 97.04% and FRR of 5.06%, which also improves identification speed and performance compared to other studies, marking a notable advancement in cow face identification.
Keywords-Cow face identification; Convolutional neural networks; ResNet; Attention module