Mixture of Bilateral-Projection Two-Dimensional Probabilistic Principal Component Analysis

Fujiao Ju, Yanfeng Sun, Junbin Gao, Simeng Liu, Yongli Hu, Baocai Yin

CVPR 2016

doi:10.1109/CVPR.2016.483 /cvpr/2016/ju2016cvpr-mixture/

Abstract

The probabilistic principal component analysis (PPCA) is built upon a global linear mapping, with which it is insufficient to model complex data variation. This paper proposes a mixture of bilateral-projection probabilistic principal component analysis model (mixB2DPPCA) on 2D data. With multi-components in the mixture, this model can be seen as a `soft' cluster algorithm and has capability of modeling data with complex structures. A Bayesian inference scheme has been proposed based on the variational EM (Expectation-Maximization) approach for learning model parameters. Experiments on some publicly available databases show that the performance of mixB2DPPCA has been largely improved, resulting in more accurate reconstruction errors and recognition rates than the existing PCA-based algorithms.

PDF CVPR Semantic Scholar

Cite

Text

Ju et al. "Mixture of Bilateral-Projection Two-Dimensional Probabilistic Principal Component Analysis." Conference on Computer Vision and Pattern Recognition, 2016. doi:10.1109/CVPR.2016.483

Markdown

[Ju et al. "Mixture of Bilateral-Projection Two-Dimensional Probabilistic Principal Component Analysis." Conference on Computer Vision and Pattern Recognition, 2016.](https://mlanthology.org/cvpr/2016/ju2016cvpr-mixture/) doi:10.1109/CVPR.2016.483

BibTeX

@inproceedings{ju2016cvpr-mixture,
  title     = {{Mixture of Bilateral-Projection Two-Dimensional Probabilistic Principal Component Analysis}},
  author    = {Ju, Fujiao and Sun, Yanfeng and Gao, Junbin and Liu, Simeng and Hu, Yongli and Yin, Baocai},
  booktitle = {Conference on Computer Vision and Pattern Recognition},
  year      = {2016},
  doi       = {10.1109/CVPR.2016.483},
  url       = {https://mlanthology.org/cvpr/2016/ju2016cvpr-mixture/}
}