A generalizable model for the facial recognition of sika deer with enhanced cross-domain performance

Animal recognition technology is recognized as a fundamental tool for ecological conservation, species monitoring, and scientific research. The population management of sika deer, a first-class protected species in China, heavily depends on precise individual identification. However, challenges such...

Full description

Saved in:
Bibliographic Details
Main Authors: Ye Mu, Jinghuan Hu, Zhipeng Li, Heyang Wang, He Gong, Yu Sun, Tianli Hu
Format: Article
Language:English
Published: Elsevier 2025-08-01
Series:Smart Agricultural Technology
Subjects:
Online Access:http://www.sciencedirect.com/science/article/pii/S2772375525002047
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849312679346307072
author Ye Mu
Jinghuan Hu
Zhipeng Li
Heyang Wang
He Gong
Yu Sun
Tianli Hu
author_facet Ye Mu
Jinghuan Hu
Zhipeng Li
Heyang Wang
He Gong
Yu Sun
Tianli Hu
author_sort Ye Mu
collection DOAJ
description Animal recognition technology is recognized as a fundamental tool for ecological conservation, species monitoring, and scientific research. The population management of sika deer, a first-class protected species in China, heavily depends on precise individual identification. However, challenges such as complex backgrounds, occlusions, and age-related facial variations pose significant obstacles to robust facial recognition. To address these domain-specific limitations, an enhanced YOLOv11-based architecture, DeerFace-YOLO, is proposed. The model is validated through transfer learning across developmental stages using a curated dataset of 40 individuals with diverse postures in complex environments. To assess scalability, the framework is extended to pig facial recognition and cattle individual detection tasks, demonstrating cross-species generalization capabilities. A dual-channel architecture is implemented, integrating Vision Transformer (ViT) and Convolutional Neural Networks (CNNs) to synergize local feature extraction with global contextual modeling, thereby mitigating CNN’s limited receptive fields and ViT’s computational complexity. Background interference is suppressed through adaptive attention mechanisms. For multi-scale feature optimization, a Multi-Scale Feature Enhancement (MSFE) module is designed to dynamically prioritize task-relevant features. Additionally, a Reparameterizable Lightweight Detection Head (RLDH) is developed to maintain accuracy while reducing parameter redundancy by 33 %. Experimental results demonstrate state-of-the-art performance, achieving 98.7 % recognition accuracy and a threefold reduction in computational load compared to baseline models. The framework establishes a standardized dataset for non-invasive animal identification and provides a scalable solution for ecological conservation and precision livestock management.
format Article
id doaj-art-9d9476a3539c4b229c4f3deb4c2a98f5
institution Kabale University
issn 2772-3755
language English
publishDate 2025-08-01
publisher Elsevier
record_format Article
series Smart Agricultural Technology
spelling doaj-art-9d9476a3539c4b229c4f3deb4c2a98f52025-08-20T03:53:01ZengElsevierSmart Agricultural Technology2772-37552025-08-011110097110.1016/j.atech.2025.100971A generalizable model for the facial recognition of sika deer with enhanced cross-domain performanceYe Mu0Jinghuan Hu1Zhipeng Li2Heyang Wang3He Gong4Yu Sun5Tianli Hu6College of Information Technology, Jilin Agricultural University, Changchun 130118, China; Jilin Province Agricultural Internet of Things Technology Collaborative Innovation Center, Changchun 130118, China; Jilin Province Intelligent Environmental Engineering Research Center, Changchun 130118, ChinaCollege of Information Technology, Jilin Agricultural University, Changchun 130118, ChinaFaculty of Animal Science and Technology, Jilin Agricultural University, Changchun 130118, ChinaCollege of Information Technology, Jilin Agricultural University, Changchun 130118, ChinaCollege of Information Technology, Jilin Agricultural University, Changchun 130118, China; Jilin Province Agricultural Internet of Things Technology Collaborative Innovation Center, Changchun 130118, China; Jilin Province Intelligent Environmental Engineering Research Center, Changchun 130118, ChinaCollege of Information Technology, Jilin Agricultural University, Changchun 130118, ChinaCollege of Information Technology, Jilin Agricultural University, Changchun 130118, China; Jilin Province Agricultural Internet of Things Technology Collaborative Innovation Center, Changchun 130118, China; Jilin Province Intelligent Environmental Engineering Research Center, Changchun 130118, China; Corresponding author at: College of Information Technology, Jilin Agricultural University, Changchun 130118, China.Animal recognition technology is recognized as a fundamental tool for ecological conservation, species monitoring, and scientific research. The population management of sika deer, a first-class protected species in China, heavily depends on precise individual identification. However, challenges such as complex backgrounds, occlusions, and age-related facial variations pose significant obstacles to robust facial recognition. To address these domain-specific limitations, an enhanced YOLOv11-based architecture, DeerFace-YOLO, is proposed. The model is validated through transfer learning across developmental stages using a curated dataset of 40 individuals with diverse postures in complex environments. To assess scalability, the framework is extended to pig facial recognition and cattle individual detection tasks, demonstrating cross-species generalization capabilities. A dual-channel architecture is implemented, integrating Vision Transformer (ViT) and Convolutional Neural Networks (CNNs) to synergize local feature extraction with global contextual modeling, thereby mitigating CNN’s limited receptive fields and ViT’s computational complexity. Background interference is suppressed through adaptive attention mechanisms. For multi-scale feature optimization, a Multi-Scale Feature Enhancement (MSFE) module is designed to dynamically prioritize task-relevant features. Additionally, a Reparameterizable Lightweight Detection Head (RLDH) is developed to maintain accuracy while reducing parameter redundancy by 33 %. Experimental results demonstrate state-of-the-art performance, achieving 98.7 % recognition accuracy and a threefold reduction in computational load compared to baseline models. The framework establishes a standardized dataset for non-invasive animal identification and provides a scalable solution for ecological conservation and precision livestock management.http://www.sciencedirect.com/science/article/pii/S2772375525002047Facial recognitionSika deer detectionYOLOv11Transfer learningAnimal recognition
spellingShingle Ye Mu
Jinghuan Hu
Zhipeng Li
Heyang Wang
He Gong
Yu Sun
Tianli Hu
A generalizable model for the facial recognition of sika deer with enhanced cross-domain performance
Smart Agricultural Technology
Facial recognition
Sika deer detection
YOLOv11
Transfer learning
Animal recognition
title A generalizable model for the facial recognition of sika deer with enhanced cross-domain performance
title_full A generalizable model for the facial recognition of sika deer with enhanced cross-domain performance
title_fullStr A generalizable model for the facial recognition of sika deer with enhanced cross-domain performance
title_full_unstemmed A generalizable model for the facial recognition of sika deer with enhanced cross-domain performance
title_short A generalizable model for the facial recognition of sika deer with enhanced cross-domain performance
title_sort generalizable model for the facial recognition of sika deer with enhanced cross domain performance
topic Facial recognition
Sika deer detection
YOLOv11
Transfer learning
Animal recognition
url http://www.sciencedirect.com/science/article/pii/S2772375525002047
work_keys_str_mv AT yemu ageneralizablemodelforthefacialrecognitionofsikadeerwithenhancedcrossdomainperformance
AT jinghuanhu ageneralizablemodelforthefacialrecognitionofsikadeerwithenhancedcrossdomainperformance
AT zhipengli ageneralizablemodelforthefacialrecognitionofsikadeerwithenhancedcrossdomainperformance
AT heyangwang ageneralizablemodelforthefacialrecognitionofsikadeerwithenhancedcrossdomainperformance
AT hegong ageneralizablemodelforthefacialrecognitionofsikadeerwithenhancedcrossdomainperformance
AT yusun ageneralizablemodelforthefacialrecognitionofsikadeerwithenhancedcrossdomainperformance
AT tianlihu ageneralizablemodelforthefacialrecognitionofsikadeerwithenhancedcrossdomainperformance
AT yemu generalizablemodelforthefacialrecognitionofsikadeerwithenhancedcrossdomainperformance
AT jinghuanhu generalizablemodelforthefacialrecognitionofsikadeerwithenhancedcrossdomainperformance
AT zhipengli generalizablemodelforthefacialrecognitionofsikadeerwithenhancedcrossdomainperformance
AT heyangwang generalizablemodelforthefacialrecognitionofsikadeerwithenhancedcrossdomainperformance
AT hegong generalizablemodelforthefacialrecognitionofsikadeerwithenhancedcrossdomainperformance
AT yusun generalizablemodelforthefacialrecognitionofsikadeerwithenhancedcrossdomainperformance
AT tianlihu generalizablemodelforthefacialrecognitionofsikadeerwithenhancedcrossdomainperformance