基于人工智能的实验动物中心信息化升级及应用实践

doi:10.12300/j.issn.1674-5817.2024.181

摘要/Abstract

摘要：

目的针对传统实验动物中心管理中存在的笼位调度效率低、人员行为监管不到位、设备老化难升级等问题，通过应用多模态大模型技术升级现有实验动物中心的信息系统，实现实验动物笼位状态的实时感知、实验人员行为的智能监管以及业务流程的自动化处理，从而提升管理效率与精细化水平。方法浙江大学医学院附属第一医院实验动物中心提出一个基于人工智能的实验动物中心信息化升级方式，能兼容不同饲养设备。该系统架构自下而上依次包括硬件设施层、核心算法层和应用功能层。硬件设施层配备摄像头和高速网络传输设备，用于笼位和人员信息采集；核心算法层通过多阶段图像预处理技术和多模态大模型识别技术，实现图片信息抽取和识别；应用功能层将识别结果和已有动物中心信息相整合，生成实时笼位占有热力图，直观、清晰地展示实验动物中心的笼位使用密度分布情况。结果基于人工智能的管理系统笼位识别准确率达98.5%，人员的实验服正确穿戴识别率为98.8%。每张图片平均处理时间约为3.7 s，笼位有效使用率提升23%，周转效率提高35%。此外，管理系统能够实时追踪并警告不规范行为，在通过智能化识别后，综合违规行为的发现次数暴露更多，违规行为发现率提升90.6%，持续使用3个月后，周平均违规行为较基线期下降54.0%。结论将多模态大模型应用于实验动物管理领域，可实现笼位标识实时监控与自动化管理，提升管理效率和精确度。系统整合视觉识别和行为分析等多源数据，可构建对实验人员的全方位智能监管体系，为科研机构提供高效、精准、价廉的管理工具，推动实验动物管理的智能化发展。

关键词: 实验动物设施管理, 多模态大模型, 智能识别, 信息系统

Abstract:

Objective In traditional laboratory animal centers, there are issues such as low efficiency in cage scheduling, insufficient supervision of personnel behavior, and difficulty in upgrading aging equipment. This study aims to upgrade the information system of existing laboratory animal centers by applying multimodal large language model technology. This upgrade intends to achieve real-time perception of the status of animal cages, intelligent supervision of experimental personnel behavior, and automated processing of business workflows, thereby improving management efficiency and precision. Methods An AI-based approach for upgrading laboratory animal center informatization was proposed by the First Affiliated Hospital of Zhejiang University School of Medicine,compatible with different breeding equipments. The system architecture, from the bottom up, consisted of three layers: hardware layer, core algorithm layer, and application layer. The hardware layer was equipped with cameras and high-speed network transmission devices for collecting information on cages and personnel. The core algorithm layer utilized multi-stage image preprocessing technology and multimodal large language model recognition technology to extract and identify image information. The application layer integrated the recognition results with the existing information of the animal center to generate real-time cage occupancy heatmaps, which visually and clearly showed the density distribution of cage usage in the laboratory animal center. Results The AI-based management system achieved a cage recognition accuracy of 98.5% and a correct wearing identification rate of laboratory coats of 98.8%. The average image processing time was 3.7 seconds per image, the effective utilization rate of cages increased by 23%, and the turnover efficiency improved by 35%. In addition, the management system could track and warn against non-compliant behaviors in real time. After intelligent recognition, the system detected more violations, with the violation detection rate increasing by 90.6%. After continuous use for three months, the weekly average number of violations decreased by 54.0% compared to the baseline period. Conclusion This study applies multimodal large language model to the field of laboratory animal management, achieving real-time monitoring and automated management of cage identification, thereby improving management efficiency and precision. The system integrates multi-source data such as visual recognition and behavior analysis, establishing a comprehensive intelligent supervision system for experimental personnel. It provides research institutions with efficient, accurate, and cost-effective management tools, promoting the intelligent development of laboratory animal management.

Key words: Laboratory animal facility management, Multimodal large language model, Intelligent recognition, Information system

中图分类号:

Q95-33

王庭君,罗浩,陈琦. 基于人工智能的实验动物中心信息化升级及应用实践[J]. 实验动物与比较医学, 2025, 45(4): 473-482. DOI: 10.12300/j.issn.1674-5817.2024.181.

WANG Tingjun,LUO Hao,CHEN Qi. Discussion on AI-Based Digital Upgrade and Application Practice of Laboratory Animal Centers[J]. Laboratory Animal and Comparative Medicine, 2025, 45(4): 473-482. DOI: 10.12300/j.issn.1674-5817.2024.181.

图/表 4

参考文献 26

[1]	国家市场监督管理总局, 国家标准化管理委员会. 实验动物环境及设施: [S]. 北京: 中国标准出版社, 2023.
	State Administration for Market Regulation, National Standardization Administration. Laboratory animal—Environment and housing facilities: [S]. Beijing: Standards Press of China, 2023.
[2]	刘莲莲, 张成梅, 张海艇, 等. 高校动物房屏障系统高效管理初探[J]. 医学动物防制, 2020, 36(7):661-662, 667. DOI: 10.7629/yxdwfz202007015 .
	LIU L L, ZHANG C M, ZHANG H T, et al. Primary study on efficient management of animal house barrier system in colleges and universities[J]. J Med Pest Contr, 2020, 36(7):661-662, 667. DOI: 10.7629/yxdwfz202007015 .
[3]	刘琪帅, 李朝, 任晓霞. "双一流"高校建设背景下新型实验动物中心的建设与管理[J]. 实验技术与管理, 2022, 39(9):266-268. DOI: 10.16791/j.cnki.sjg.2022.09.045 .
	LIU Q S, LI Z, REN X X. Construction and management of new laboratory animal center under background of "Double First-Class" university construction[J]. Exp Technol Manag, 2022, 39(9):266-268. DOI: 10.16791/j.cnki.sjg.2022.09.045 .
[4]	陈晓娟, 李巍, 汪洌. 高校实验动物管理流程信息化探索: 以浙江大学为例[J]. 实验动物与比较医学, 2021, 41(6):554-558. DOI: 10.12300/j.issn.1674-5817.2021.045 .
	CHEN X J, LI W, WANG L. Exploration on informatization of laboratory animal management process in colleges and universities: the case of Zhejiang University[J]. Lab Anim Comp Med, 2021, 41(6):554-558. DOI: 10.12300/j.issn.1674-5817.2021.045 .
[5]	国家质量监督检验检疫总局, 国家标准化管理委员会. 实验动物福利伦理审查指南: [S]. 北京: 中国标准出版社, 2018.
	General Administration of Quality Supervision, Inspection and Quarantine of the People's Republic of China, Standardization Administration of China. Laboratory animal—Guideline for ethical review of animal welfare: [S]. Beijing: Standards Press of China, 2018.
[6]	陆伟, 刘家伟, 马永强, 等. ChatGPT为代表的大模型对信息资源管理的影响[J]. 图书情报知识, 2023, 40(2):6-9, 70. DOI: 10.13366/j.dik.2023.02.006 .
	LU W, LIU J W, MA Y Q, et al. The influence of large language models represented by ChatGPT on information resources management[J]. Doc Inf Knowl, 2023, 40(2):6-9, 70. DOI: 10.13366/j.dik.2023.02.006 .
[7]	徐敏, 张振建, 许景东, 等. "一院多区"医院管理的探索与思考[J]. 中国卫生质量管理, 2017, 24(4):107-109. DOI: 10.13912/j.cnki.chqm.2017.24.4.35 .
	XU M, ZHANG Z J, XU J D, et al. Exploration and thinking of hospital management in "one hospital with multi-district"[J]. Chin Health Qual Manag, 2017, 24(4):107-109. DOI: 10.13912/j.cnki.chqm.2017.24.4.35 .
[8]	陈玲. 高校实验室安全管理面临的问题与对策[J]. 实验室研究与探索, 2017, 36(1):283-286. DOI: 10.3969/j.issn.1006-7167.2017.01.070 .
	CHEN L. Problems and countermeasures of laboratory safety management in colleges and universities in the new era[J]. Res Explor Lab, 2017, 36(1):283-286. DOI: 10.3969/j.issn.1006-7167.2017.01.070 .
[9]	胡永艳, 陶迎红, 孔申申. 一种基于互联网的实验动物管理系统介绍[J]. 实验动物与比较医学, 2020, 40(3):257-261. DOI: 10.3969/j.issn.1674-5817.2020.03.015 .
	HU Y Y, TAO Y H, KONG S S. Introduction of a web-based novel laboratory animal management system[J]. Lab Anim Comp Med, 2020, 40(3):257-261. DOI: 10.3969/j.issn.1674-5817.2020.03.015 .
[10]	宋予震, 郭素琴, 史洪涛, 等. 动物医学虚拟仿真实验教学平台的建设与应用[J]. 高教学刊, 2020, 6(21):79-81. DOI: 10.19980/j.cn23-1593/g4.2020.21.023 .
	SONG Y Z, GUO S Q, SHI H T, et al. Construction and application of virtual simulation experimental teaching platform for animal medicine[J]. J High Educ, 2020, 6(21):79-81. DOI: 10.19980/j.cn23-1593/g4.2020.21.023 .
[11]	黄术兵, 姚文茜, 侯豹, 等. 高校实验动物管理工作的现状和对策研究[J]. 实验室科学, 2023, 26(6):147-150. DOI: 10.3969/j.issn.1672-4305.2023.06.035 .
	HUANG S B, YAO W X, HOU B, et al. Current situation and countermeasures on the management of laboratory animals in universities[J]. Lab Sci, 2023, 26(6):147-150. DOI: 10.3969/j.issn.1672-4305.2023.06.035 .
[12]	李秀, 陈晓梅, 商可心, 等. 使用型高校实验动物中心信息化管理模式开发应用[J]. 上海畜牧兽医通讯, 2020(4):61-65, 68. DOI: 10.14170/j.cnki.cn31-1278/s.2020.04.025 .
	LI X, CHEN X M, SHANG K X, et al. Development and application of information management mode of experimental animal center in using university[J]. Shanghai J Anim Husb Vet Med, 2020(4):61-65, 68. DOI: 10.14170/j.cnki.cn31-1278/s.2020.04.025 .
[13]	胡瑚. 基于MySQL的科研信息管理系统数据库设计[J]. 信息与电脑(理论版), 2024, 36(4):169-171. DOI: 10.3969/j.issn.1003-9767.2024.04.054 .
	HU H. Database design of scientific research information management system based on MySQL[J]. Inf Comput, 2024, 36(4):169-171. DOI: 10.3969/j.issn.1003-9767.2024.04.054 .
[14]	郭全中, 苏刘润薇. 作为新基础设施的AGI: 以GPT-4O等新一代生成式人工智能为例[J]. 新闻爱好者, 2024(7):16-21. DOI: 10.16017/j.cnki.xwahz.20240527.001 .
	GUO Q Z, SU L R W. AGI as a new infrastructure: Taking GPT-4O and other new generation of generative artificial intelli-gence as an example[J]. Jour Lov, 2024(7):16-21. DOI: 10.16017/j.cnki.xwahz.20240527.001 .
[15]	郭全中, 张金熠. ChatGPT的技术特征与应用前景[J]. 中国传媒科技, 2023(1):159-160. DOI: 10.3969/j.issn.1671-0134.2023.01.032 .
	GUO Q Z, ZHANG J Y. Technical characteristics and application prospect of ChatGPT[J]. Media Sci Technol China, 2023(1):159-160. DOI: 10.3969/j.issn.1671-0134.2023.01.032 .
[16]	LI J J, GUAN Z Y, WANG J, et al. Integrated image-based deep learning and language models for primary diabetes care[J]. Nat Med, 2024, 30(10):2886-2896. DOI:10.1038/s41591-024-03139-8 .
[17]	大经中医. 岐黄问道大模型[EB/OL]. (2023-08-01)[2024-12-10]. .
	Dajing Chinese Medicine. Qihuang wendao large model[EB/OL]. (2023-08-01)[2024-12-10]. .
[18]	YANG S H, ZHAO H J, ZHU S B, et al. Zhongjing: enhancing the Chinese medical capabilities of large language model through expert feedback and real-world multi-turn dialogue[C]//AAAI Conference on Artificial Intelligence, 2025.
[19]	Interpreter X-Ray. Empower your medical choices with instant a X-ray insights[Z/OL]. [2025-08-01]. .
[20]	LI C Y, WONG C, ZHANG S, et al. LLaVA-med: training a large language-and-vision assistant for biomedicine in one day[C]. Proceedings of the 37th International Conference on Neural Information Processing Systems Article, 2023.
[21]	马驰, 袁粒星, 林玲, 等. 基于信息化的实验动物中心管理新模式实践[J]. 实验室研究与探索, 2020, 39(6):253-255, 277.
	MA C, YUAN L X, LIN L, et al. Practice of a new management model for laboratory animal centers based on information technology[J]. Res Explor Lab, 2020, 39(6): 253-255, 277.
[22]	朱玉, 秦嫘, 辛丹, 等. 高等院校实验动物的信息化管理体系建设探讨[J]. 畜牧兽医科技信息, 2022(6):19-21. DOI: 10.3969/J.ISSN.1671-6027.2022.06.007 .
	ZHU Y, QIN L, XIN D, et al. Discussion on the construction of information management system of experimental animals in colleges and universities[J]. Chin J Anim Husb Vet Med, 2022(6):19-21. DOI: 10.3969/J.ISSN.1671-6027.2022.06.007 .
[23]	赵朝阳, 朱贵波, 王金桥. ChatGPT给语言大模型带来的启示和多模态大模型新的发展思路[J]. 数据分析与知识发现, 2023, 7(3):26-35. DOI: 10.11925/infotech.2096-3467.2023.0216 .
	ZHAO C Y, ZHU G B, WANG J Q. The inspiration brought by ChatGPT to LLM and the new development ideas of multi-modal large model[J]. Data Anal Knowl Discov, 2023, 7(3):26-35. DOI: 10.11925/infotech.2096-3467.2023.0216 .
[24]	郭君斌, 于琳, 于传强. 改进YOLOv5s算法在交通标志检测识别中的应用[J]. 国防科技大学学报, 2024, 46(6):123-130. DOI: 10.11887/j.cn.202406013 .
	GUO J B, YU L, YU C Q. Application of improved YOLOv5s algorithm in traffic sign detection and recognition[J]. J Natl Univ Def Technol, 2024, 46(6):123-130. DOI: 10.11887/j.cn.202406013 .
[25]	韩冬, 李其花, 蔡巍, 等. 人工智能在医学影像中的研究与应用[J]. 大数据, 2019, 5(1):39-67. DOI: 10.11959/j.issn.2096-0271.2019004 .
	HAN D, LI Q H, CAI W, et al. Research and application of artificial intelligence in medical imaging[J]. Big Data Res, 2019, 5(1):39-67. DOI: 10.11959/j.issn.2096-0271.2019004 .
[26]	WANG C Y, BOCHKOVSKIY A, LIAO H M. YOLOv7: trainable bag-of-freebies sets new state-of-the-art for real-time object detectors[C] //2023 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). June 17-24,2023,Vancouver,BC,Canada.IEEE,2023:7464-7475. DOI:10.1109/CVPR52729.2023.00721 .

对比项目 Comparison items	本方案（多模态大模型） This solution (multimodal large language model)	人工监管方法 Manual supervision method	“信息化系统 + 传统人工智能笼位识别”方案 "Information system + You Only Look Once" solution
识别技术 Recognition technology	GPT-4o大模型	人工目视核查	YOLO目标检测
识别准确率 Recognition accuracy	98.5%（持续优化）	监管人员易疲劳	> 98%^*（理论推测值）
改造要求 Upgrade requirements	摄像头，网络系统，普通计算机	无	摄像头，网络系统，配备主流高性能显卡的计算机
改造成本 Upgrade cost	低	无	高
使用成本 Usage cost	低	高	低
工作效率 Work efficiency	高（3.7 s/笼架）	低（5～10 min/笼架）	极快（1 s以内）
适应性 Adaptability	强（可适应多种场景）	强	弱（需要与训练数据类似）
拓展性 Scalability	完善	无	支持的，但是成本高昂，需要重新训练模型

对比项目 Comparison items	本方案（多模态大模型） This solution (multimodal large language model)	人工监管方法 Manual supervision method	“信息化系统 + 传统人工智能笼位识别”方案 "Information system + You Only Look Once" solution
识别技术 Recognition technology	GPT-4o大模型	人工目视核查	YOLO目标检测
识别准确率 Recognition accuracy	98.5%（持续优化）	监管人员易疲劳	> 98%^*（理论推测值）
改造要求 Upgrade requirements	摄像头，网络系统，普通计算机	无	摄像头，网络系统，配备主流高性能显卡的计算机
改造成本 Upgrade cost	低	无	高
使用成本 Usage cost	低	高	低
工作效率 Work efficiency	高（3.7 s/笼架）	低（5～10 min/笼架）	极快（1 s以内）
适应性 Adaptability	强（可适应多种场景）	强	弱（需要与训练数据类似）
拓展性 Scalability	完善	无	支持的，但是成本高昂，需要重新训练模型