基于深度学习的以图搜图技术在照片档案管理中的应用研究

摘要

传统档案管理中识别照片档案内容完全依靠手工著录文字说明,但实际管理中存在文字说明著录不全导致照片内容展示不全等缺点。在回顾照片档案管理现状和图像内容搜索技术发展的基础上,本文提出基于深度学习的以图搜图照片档案管理系统架构,并运用Keras深度学习框架和VGG16网络模型成功实现照片档案以图搜图实验。运用深度学习方法来对照片档案进行图片内容识别,对照片档案中的内容进行特征提取,实现以图搜图,为照片档案的管理利用和资源开发提供理论支持和实践思路。

Abstract

In traditional archives management, the content recognition of photo archives completely relies on the manual description text, but in the actual management, there are some shortcomings such as incomplete description text leading to the incomplete display of the photo contents. Based on the review of the current situation of photo archives management and the development of image search technology, this paper puts forward the architecture of photo archives management system based on deep learning model, and uses the Keras deep learning framework and VGG16 network model to successfully implements photo archives search. This paper uses deep learning method to recognize the image contents of the photo archives , extract the features of the contents of the photo archives, search the image by the image, and wants to theoretical support and practical ideas for the management, utilization and resource development of photo archives.

引用格式