.png){kind=icon}
.png){kind=icon}
Return
Merging model-based two-dimensional principal component analysis
.png){kind=icon}
PRE
AI
DOI：10.1016/j.neucom.2015.05.002.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Two-dimensional Principal Component Analysis (2DPCA) is a prevalent way to represent images with widespread applications. However, its performance will remarkably degenerate when directly using it to online learning and big data analysis. In this paper we present a new constructive approach to merge multiple eigenspaces of 2DPCA for successively adding new observations and then provide an efficient way to solve the projection matrix of incremental 2DPCA. The proposed method takes into account the change of mean of data, which is important for classification, and significantly reduces the computation complexity and storage space. Experimental results on the FERET, AR, and PIE face databases show the efficiency of the proposed method for online learning and big data analysis. (C) 2015 Elsevier B.V. All rights reserved.
Keywords:
2DPCA
Dimensionality reduction
Model merging
Big data