Mixed Noise Image De-Noising Based on EM Algorithm

Abstract:

Obtaining high quality images is very important in many areas of applied sciences, but images are usually polluted by noise in the process of generation, transmission and acquisition. In recent years, wavelet analysis achieves significant results in the field of image de-noising. However, most of the studies of noise-induced phenomena assume that the noise source is Gaussian. The use of mixed Gaussian and impulse noise is rare, mainly because of the difficulties in handling them. In the process of image de-noising, the noise model’s parameter estimation is a key issue, because the accuracy of the noise model’s parameters could affect the de-noising quality. In the case of mixed Gaussian noises, EM algorithm is an iterative algorithm, which simplifies the maximum likelihood equation. This thesis takes wavelet analysis and statistics theory as tools, studies on mixed noise image de-noising, provides two classes of algorithms for dealing with a special type of non-Gaussian noise, mixed Gaussian and Pepper & Salt noise.

• Periodical: Applied Mechanics and Materials (Volumes 556-562)
• Pages: 4734-4741
• DOI: https://doi.org/10.4028/www.scientific.net/AMM.556-562.4734
• Online since: May 2014
• Authors: Gui Cun Shi*, Fei Xing Wang
• Keywords: EM Algorithm, Gaussian Noise, Variance Estimation, Wavelet Transform (WT)

References

[1] Lei Tian-hu, Udupa Jayaram K. Performance evaluation of finite normal mixture model-based image segmentation techniques [J]. IEEE Transactions on Image Processing, 2003, 12(10): 1153-1169.

DOI: 10.1109/tip.2003.817251

[2] R.C. Gonzalez, R.E. Woods, S.L. Edins, Digital Image Processing Using MATLAB, Publishing House of Electronics Industry. Beijing, (2004).

[3] Jutten C, Herault J. Blind separation of sources, Part I: An adaptive algorithm based on neuromimetic architecture. Signal Processing, 1991, 24(1): 1-10.

DOI: 10.1016/0165-1684(91)90079-x

[4] Jeff A. Bilmes. A Gentle Tutorial of the EM Algorithm and its Application to Parameter Estimation for Gaussian Mixture and Hidden Markov Models, Department of Electrical Engineering and Computer Science, U.C. Berkeley. (1998).

[5] Vidakovic. B, Lozoya C.B. On Time-dependent Wavelet De-noising [J]. IEEE Transactions on Signal Processing (S1053-587X), 1998, 46(9): 2549-2551.

DOI: 10.1109/78.709544

[6] GONG Sheng Rong, Liu Chunping, Wang Qiang. Digital image processing and analysis [M]. Beijing: Tsinghua University Press, 2006, (24): 88-90.

[7] Guo Haitao, Tian Tan,　Wang Lianyu,　Zhang Chuntian. Image Segmentation Using the Maximum Entropy of the TwoDimensional Bound Histogram [J]. ACTA OPTICA SINICA, 2006, 26(4): 507-508.