Diffusion Hashing
Diffusion Hashing
复制标题
DOI：
--
复制
发表时间：
2011
期刊：
影响因子：
--
通讯作者：
Atsushi Tatsuma;Masaki Aono
中科院分区：
文献类型：
--
作者：
Atsushi Tatsuma;Masaki Aono
随着宽带互联网在全球范围内的普及，包括文本、图像和视频在内的海量多媒体数据呈爆炸式增长，并可用于互联网上的交互应用。与此同时，针对海量多媒体数据库的快速检索也越来越受到人们的重视。基于散列的近似最近邻（ANN）搜索是一种将散列键作为检索索引来实现快速检索的技术，其中数据的相似性被保持并嵌入在散列键的邻域中。换句话说，哈希键之间的汉明码越接近，数据就越相似。一般来说，短二进制代码更适合存储哈希键和值。困难在于定义数据之间的相似性并将其反映在二进制代码中。在本文中，我们提出了扩散哈希（DH）作为一种新的人工神经网络搜索技术的基础上散列各向异性扩散核。DH的目标是将搜索索引转换为尽可能短的二进制代码，保持高维空间中数据流形上随机游走引起的相似性。从比较实验中，我们将证明DH优于以前已知的基于散列的ANN搜索技术，包括局部敏感散列和频谱散列。
With the worldwide spread of the broadband Internet, massive multimedia data including texts, images, and videos are increasing explosively and available for interactive applications over the Internet. At the same time, more and more attention has been paid to aiming at fast retrieval from massive multimedia databases. Hash-based Approximate Nearest Neighbor (ANN) search is a technology that achieves fast retrieval by regarding the hash key as a retrieval index, where the similarity of data is maintained and embedded in the neighborhood of the hash key. In other words, the closer the Hamming codes between hash keys, the more similar the data become. In general, short binary codes are preferred for storing hash keys and values. The difficulty is to define the similarity between data and reflect it in binary codes. In this paper, we propose Diffusion Hashing (DH) as a novel ANN search technique based on hashing with an anisotropic diffusion kernel. DH aims to transform the search index into as short binary codes as possible, preserving the similarity induced by random walk on the data manifold in higher dimensional space. From comparative experiments, we will demonstrate that DH outperforms previously known hash-based ANN search techniques including Locality Sensitive Hashing and Spectral Hashing.