Uncorrelated Discriminant Locality Preserving Projections

• Published in: IEEE signal processing letters Vol. 15; pp. 361 - 364
• Main Authors: Xuelian Yu, Xuelian Yu, Xuegang Wang, Xuegang Wang
• Format: Journal Article
• Language: English
• Published: New York IEEE 2008; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Algorithm design and analysis; Between-class distance; Face recognition; Feature extraction; Hilbert space; Kernel; manifold learning; Pattern analysis; Radar scattering; Signal processing algorithms; Target recognition; Training data; uncorrelated constraint; within-class geometric structure
• ISSN: 1070-9908; 1558-2361
• DOI: 10.1109/LSP.2008.919841

In this letter, a new manifold learning algorithm, called uncorrelated discriminant locality preserving projections (UDLPP), is proposed. The aim of UDLPP is to preserve the within-class geometric structure, while maximizing the between-class distance. By introducing a simple uncorrelated constraint into the objective function, we show that the extracted features via UDLPP are statistically uncorrelated, which is desirable for many pattern analysis applications. Moreover, UDLPP can be performed in reproducing kernel Hilbert space, which gives rise to kernel UDLPP. Experimental results on both face recognition and radar target recognition demonstrate the effectiveness of the proposed algorithm.