面向高校图书馆的智慧搜索模式研究

Research on Smart Search Models for University Libraries

为了实现高校图书馆对智能化、数字化、动态化、全面化的智慧搜索模式的需求,并提高高校图书馆检索服务体验,增强高校图书馆竞争力。研究基于移动视觉搜索技术提出了满足用户多元化需求的高校图书馆智慧搜索系统,并使用注意力机制对YOLOv5网络进行改进。实验结果表明,YOLOv5+CA识别精度为0.904,运行时间为42.5 ms,内存占比为1.2%,有较好的计算效能和较小的内存占用,可以适配于图书馆中的数据智能检索服务。

In order to satisfy the needs of university libraries for intelligent,digital,dynamic and comprehensive intelligent search,and improve the experience of university library retrieval services and enhance the competitiveness of university libraries,the study proposes an intelligent search system for university libraries based on mobile visual search technology to meet the diversified needs of users.The YOLOv5 network is improved by using the attention mechanism.The experimental results show that YOLOv5+CA has a recognition accuracy of 0.904,a running time of 42.5 ms and a memory occupation of 1.2%,which has a better computational performance and a smaller memory occupation,and can be adapted to the intelligent data retrieval service in libraries.