banner
Home > Open Journal Systems

Performance evaluation of KPCA pre-imaging methods for speech signal denoising

Shashikant L. Sahare, Ashok R. Khedkar, Sandeep S. Musale

Abstract


Kernel principal component analysis (KPCA) has gained wider interest amongst the researchers in nonlinear dimensionality reduction, data compression, feature extraction, and denoising applications. In KPCA, data from low dimensional input space implicitly mapped to higher dimensional feature space where linear PCA is enforced. However, for denoising application data need to invert back to input space, which is impossible and known as pre-imaging problems. Over recent years, several pre-imaging methods have proposed each with its benefits and disadvantages. In this paper, we evaluated the performance for selected pre-imaging methods for denoising speech signal, whose intelligibility and quality degraded by background noises. Further, we extend our work by comparing the performance of these pre-imaging methods by objective evaluation of denoised speech signal on generated toy examples and NOIZUES database.


Keywords


denoising; projective subspace; delay embedding; kernel principal component analysis (KPCA); pre-imaging problem

Full Text:

PDF

References


1. Loizou PC. Speech Enhancement: Theory and Practice, 2nd ed. CRC Press; 2013.

2. Jolliffe IT. Principal Component Analysis. Springer Verlag; 1986.

3. Scholkopf B, Smola A, Müller KR. Nonlinear component analysis as a kernel eigenvalue problem. Neural Computation 1998; 10(5): 1299–1319.

4. Mika S, Schölkopf B, Smola A, et al. Kernel PCA and de-noising in feature spaces. In: Kearns M, Solla S, Cohn D (editors). Advances in Neural Information Processing Systems. MIT Press; 1999. pp. 536–542.

5. Kwok JTY, Tsang IWH. The pre-image problem in kernel methods. IEEE Transactions on Neural Network 2004; 15(6): 1517–1525. doi: 10.1109/TNN.2004.837781

6. Rathi Y, Dambreville S, Tannenbaum A. Statistical shape analysis using kernel PCA. Processing of SPIE-IS&T Electronic Imaging 2006; 6064: 425–432. doi: 10.1117/12.641417

7. Honeine P, Richard C. Solving the pre-image problem in kernel machines: A direct method. In: Proceedings of 19th IEEE Workshop on Machine Learning for Signal Processing (MLSP); 1–4 September 2009; Grenoble, France. pp. 1–6.

8. Abrahamsen TJ, Hansen LK. Input space regularization stabilizes pre-images for kernel PCA de-noising. In: Proceedings of the 2009 IEEE International Workshop on Machine Learning for Signal Processing; 1–4 September 2009; Grenoble, France. pp. 1–6.

9. Teixeira AR, Tome AM, Lang EW, et al. On the use of KPCA to extract artifact in one-dimensional biomedical signals. 16th IEEE Signal Processing Society Workshop on Machine Learning for Signal Processing 2006. doi: 10.1109/MLSP.2006.275580

10. Loizou PC, Kim G. Reasons why current speech-enhancement algorithms do not improve speech intelligibility and suggested solutions. IEEE Transaction on Audio, Speech, and Language Processing 2011; 19(1): 47–56. doi: 10.1109/TASL.2010.2045180.

11. Jorgensen KW, Hansen LK. Model selection for Gaussian kernel PCA denoising. IEEE Transactions on Neural Networks and Learning Systems 2012; 23(1): 163–168. doi: 10.1109/TNNLS.2011.2178325




DOI: https://doi.org/10.32629/jai.v7i1.1086

Refbacks

  • There are currently no refbacks.


Copyright (c) 2023 Shashikant L. Sahare, Ashok R. Khedkar, Sandeep S. Musale

License URL: https://creativecommons.org/licenses/by-nc/4.0/