基于深度神经网络的哈希算法研究综述

The Research Overview of Hash Algorithm Based on Deep Neural Network

随着社交媒体的快速发展,互联网上产生的图像、视频等多媒体数据迅速增长,如何快速地从海量图像数据库中找到有用的信息,已经成为大数据检索研究的主要内容之一。哈希检索技术能够将高维特征保相似地映射为哈希码,大幅度减少了检索过程中的存储与计算开销,因而得到了越来越多的关注。本文总结了深度哈希算法主要方法和前沿进展,并对未来的研究方向展开简要探讨。

With the rapid development of social media,multimedia data such as images and videos generated on the Internet are growing rapidly.How to quickly find useful information from massive image databases has become one of the main contents of big data retrieval research.Hash retrieval technology can similarly map high-dimensional features into hash codes,which greatly reduces the storage and computational overhead in the retrieval process,and thus has received more and more attention.this article summarizes the main methods and frontier progress of the deep hash algorithm,and briefly discusses the future research directions.