11 September 2013 Local minimum squared error for face and handwritten character recognition
Author Affiliations +
Abstract
The minimum squared error (MSE) for classification is a linear discriminant function-based method that has been used in many applications such as face and handwritten character recognition. Nevertheless, MSE may not deal well with nonlinearly separable data sets. To address this problem, we improve the MSE and propose a new MSE-based algorithm, local MSE (LMSE), which is a local learning algorithm. For a test sample, we first determine its nearest neighbors from the training set. By using the determined neighbors, we construct a local MSE model to predict the class label of the test sample. LMSE can effectively capture the nonlinear structure of the data. It generally outperforms MSE, particularly when the data distribution is nonlinearly separable. Extensive experiments on many nonlinearly separable data sets show that LMSE achieves desirable recognition results.
© 2013 SPIE and IS&T
Zizhu Fan, Jinghua Wang, Qi Zhu, Xiaozhao Fang, Jinrong Cui, Chunhua Li, "Local minimum squared error for face and handwritten character recognition," Journal of Electronic Imaging 22(3), 033027 (11 September 2013). https://doi.org/10.1117/1.JEI.22.3.033027 . Submission:
JOURNAL ARTICLE
10 PAGES


SHARE
RELATED CONTENT

Toward quantifying the amount of style in a dataset
Proceedings of SPIE (January 16 2006)
Object classification using local subspace projection
Proceedings of SPIE (May 19 2011)
Sampling design for face recognition
Proceedings of SPIE (April 17 2006)
Multiple-agent adaptation in whole-book recognition
Proceedings of SPIE (January 24 2011)
Face recognition experiments with random projection
Proceedings of SPIE (March 28 2005)

Back to Top