Abstract
In this paper, we present a supervised manifold learning based dimensionality reduction method, which is titled local uncorrelated subspace learning (LUSL). In the proposed LUSL, a local margin based on point to feature space (P2S) distance metric is recommended for discriminant feature extraction. What’s more, it has been validated that the locally statistical uncorrelation is an important property to reduce the feature redundancy. Hence, a novel locally uncorrelated criterion using P2S distance metric is also put forward, which is taken to constrain the local margin. Finally, by solving both the orthogonality and the local uncorrelation constrained objective function using an iterative way, a low dimensional subspace will be explored for pattern recognition. Compared to some related subspace learning methods, the effectiveness of the proposed LUSL have been shown from experimental results on some benchmark face data sets as AR and FERET.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Disclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.