Title :
Weighted grid Principal Component Analysis hashing
Author :
Xiancheng Zhou ; Zhiqian Huang ; Ng, Wing W. Y.
Abstract :
Principal Component Analysis (PCA) is one of the most widely used components of hashing. In this paper, we propose three PCA-based hashing methods to improve the performance of the Principal Component Hashing (PCH). Different principal components have different among of variances of data. In the PCH, each principal component corresponds to a hash function. Hence, the PCH treats each principal component to have the same importance which will lead to the loss of much information in constructing hashing table. To deal with this shortage, we propose the weighted PCH (WPCH), the grid PCH (GPCH) and the weighted grid PCH (WGPCH).
Keywords :
file organisation; image retrieval; principal component analysis; PCA-based hashing methods; WGPCH; WPCH; grid PCH; hash function; hashing table; weighted PCH; weighted grid PCH; weighted grid principal component analysis hashing; Abstracts; Lead; Mobile communication; Principal component analysis; Hashing; Large Scale Image Retrieval; PCA;
Conference_Titel :
Machine Learning and Cybernetics (ICMLC), 2014 International Conference on
Conference_Location :
Lanzhou
Print_ISBN :
978-1-4799-4216-9
DOI :
10.1109/ICMLC.2014.7009117