Sparse Unsupervised Dimensionality Reduction for Multiple View Data

• Published in: IEEE transactions on circuits and systems for video technology Vol. 22; no. 10; pp. 1485 - 1496
• Main Authors: Yahong Han, Fei Wu, Dacheng Tao, Jian Shao, Yueting Zhuang, Jianmin Jiang
• Format: Journal Article
• Language: English
• Published: New York, NY IEEE 01.10.2012; Institute of Electrical and Electronics Engineers
• Subjects: Applied sciences; Detection, estimation, filtering, equalization, prediction; Educational institutions; Exact sciences and technology; Feature extraction; Image processing; Information, signal and communications theory; Loading; Multiple view data; Principal component analysis; Signal and communications theory; Signal processing; Signal representation. Spectral analysis; Signal, noise; Sparse matrices; structured sparsity; Telecommunications and information theory; Vectors; video and image classification; Visualization; Performance evaluation; Low-dimensional system; Discriminant analysis; Load distribution; Image processing; Multiple view data; Resource sharing; Unsupervised classification; video and image classification; Algorithm; Signal classification; Flexibility; Adaptive method; Multiple image; Learning; Dimension reduction; Multi-view technology; Signal processing; Feature extraction; Information exchange; Gain; structured sparsity; Image classification
• ISSN: 1051-8215; 1558-2205
• DOI: 10.1109/TCSVT.2012.2202075

Different kinds of high-dimensional visual features can be extracted from a single image. Images can thus be treated as multiple view data when taking each type of extracted high-dimensional visual feature as a particular understanding of images. In this paper, we propose a framework of sparse unsupervised dimensionality reduction for multiple view data. The goal of our framework is to find a low-dimensional optimal consensus representation from multiple heterogeneous features by multiview learning. In this framework, we first learn low-dimensional patterns individually from each view, considering the specific statistical property of each view. We construct a low-dimensional optimal consensus representation from those learned patterns, the goal of which is to leverage the complementary nature of the multiple views. We formulate the construction of the low-dimensional consensus representation to approximate the matrix of patterns by means of a low-dimensional consensus base matrix and a loading matrix. To select the most discriminative features for the spectral embedding of multiple views, we propose to add an l 1 -norm into the loading matrix's columns and impose orthogonal constraints on the base matrix. We develop a new alternating algorithm, i.e., spectral sparse multiview embedding, to efficiently obtain the solution. Each row of the loading matrix encodes structured information corresponding to multiple patterns. In order to gain flexibility in sharing information across subsets of the views, we impose a novel structured sparsity-inducing norm penalty on the loading matrix's rows. This penalty makes the loading coefficients adaptively load shared information across subsets of the learned patterns. We call this method structured sparse multiview dimensionality reduction. Experiments on a toy benchmark image data set and two real-world Web image data sets demonstrate the effectiveness of the proposed algorithms.