摘要
针对失物寻找信息分散的问题,探索了图像检索技术在失物寻找系统中的应用。前端通过小程序实现图像上传和预处理,后端利用深度学习完成特征值提取、相似度计算和数据库比对等操作,从而实现高效的图像检索功能。经数据集测试结果表明,Top-5准确率达到92.3%,单次检索响应时间不超过1.2 s,验证了该方法的可行性与高效性。
In response to the problem of scattered information in lost and found items,the application of image retrieval technology in the lost and found item search system is explored.The front-end uses mini programs to upload and preprocess images,while the back-end utilizes deep learning to extract feature values,calculate similarity,and perform database comparison operations,thereby achieving efficient image retrieval functionality.The test results of the dataset show that the Top-5 accuracy reaches 92.3%,and the response time for a single retrieval does not exceed 1.2 seconds,verifying the feasibility and efficiency of this method.
作者
张丽香
沈杰颖
ZHANG Lixiang;SHEN Jieying(Private Hualian College,Guangzhou 510663,China)
基金
2024年广东省科技创新战略专项资金项目(大学生科技创新培育)
SK(用于寻找的小程序)(pdjh2024b696)。
关键词
图像检索
小程序
特征提取
数据存储
image retrieval
mini program
feature extraction
data storage
