Article Details

Title · Authors · Abstract · Keywords · DOI · References

Research on Cattle Facial Recognition Method Based on Improved MobileNetV2

TIAN Huijuan, CAO Mengqi, RU Chunrui, REN Zhaohui

(Yangling Vocational and Technical College, Xianyang 712199, China)

Section:Artificial Intelligence · 2025, Vol. 9, No. 11 · Pages:20-24 · 2096-4706(2025)11-0020-05 · CLC number:TP391.4 · Document code:A · DOI:10.19850/j.cnki.2096-4706.2025.11.005

Abstract

In view of the problems that cattle are easily frightened and consume manpower and material resources in the process of traditional cattle individual identification, this paper proposes a non-contact cattle individual identification method by means of computer vision technology. Firstly, the cattle face photos are taken in the real cattle farm environment, and the cattle face recognition dataset is constructed. Secondly, combined with the actual network conditions and computing power level of the cattle farm, the lightweight Neural Network model MobileNetV2 is selected as the basic network model. Finally, the Coordinate Attention Mechanism is introduced to improve the MobileNetV2 model, so as to improve the accuracy of the model and enhance its feature extraction ability for key positions. The experimental results show that the improved MobileNetV2 model performs well in the cattle face recognition task. The model size is only 2.86 MB, and the recognition accuracy is as high as 95.81%, which can fully meet the needs of non-contact cattle individual recognition in the actual cattle farm environment.

Keywords

MobileNetV2; Attention Mechanism; cattle identification

References

  1. [1] Huang Alwen, Qianbo, Guan Feifan, etc Sheep face recognition model based on wavelet transform and convolutional neural network [J]. Journal of Agricultural Machinery, 2023, 54 (5): 278-287
  2. [2] Statement Multi angle facial expression recognition based on improved VGGNet and its application [D]. Shenyang: Shenyang University of Technology, 2020
  3. [3] Xu Xiaohui Research on Image based Lightweight Cow Face Recognition Method [D]. Changchun: Jilin University, 2024
  4. [4] Qi Yongsheng, Zhang Xinze, Zhang Jiaying, etc Local occlusion cow face recognition method based on feature mask [J]. Journal of Agricultural Machinery, 2024, 55 (11): 93-102
  5. [5] Luan Haotian, Qi Yongsheng, Liu Liqiang, etc VanillaFaceNet: A high-precision and fast inference based cow face recognition method [J]. Journal of Agricultural Engineering, 2024, 40 (18): 120-131
  6. [6] Fan Huanxi, Ye Yaoyao Two Stage MobileNet: A Lightweight Model for Facial Attribute Recognition [J]. Modern Computer, 2020 (24): 54-59
  7. [7] SANDLER M,HOWARD A,ZHU M L,et al. Mobilenetv2: Inverted Residuals and Linear Bottlenecks [C]//2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition.Salt Lake City:IEEE,2018:4510-4520.
  8. [8] YUAN Z W,ZHANG J. Feature Extraction and Image Retrieval Based on AlexNet [EB/OL].[2024-12-28].https://www.deepdyve.com/lp/spie/feature-extraction-and-image-retrieval-based-on-alexnet-uW74B4tctW?articleList=/search-related?to=EHXCUcg06H&page=10.
  9. [9] MA N N,ZHANG X Y,ZHENG H T,et al. Shufflenet v2: Practical Guidelines for Efficient CNN Architecture Design [J/OL].arXiv:1807.11164 [cs.CV].[2024-12-28].https://arxiv.org/abs/1807.11164.
  10. [10] YU W H,ZHOU P,YAN S C,et al. Inceptionnext: When Inception Meets Convnext [C]//2024 IEEE/CVF Conference on Computer Vision and Pattern Recognition.Seattle:IEEE,2024:5672-5683.

Funding

Yangling Vocational and Technical College 2023 Science and Technology Innovation Project (zk23-02)

About the authors

Tian Huijuan (1995-), female, Han ethnicity, Shanxi University Tongren, teaching assistant, master's student, research direction: computer application technology; Cao Mengqi (1996-), female, Han ethnicity, from Lantian, Shaanxi, teaching assistant, master's student, research direction: information visualization; Ru Chunrui (1996-), female, Han ethnicity, from Pingliang, Gansu Province, lecturer, master's student, research direction: network security technology; Ren Chaohui (1996-), male, Han ethnicity, born in Yang Lingren, Shaanxi, lecturer, master's student, research direction: network system management.