Adaptive multichannel filter for image processing

B. K. Ng, K. N. Plataniotis, A. N. Venetsanopoulos

Research output: Contribution to journalConference articlepeer-review


In this paper, an adaptive and computationally efficient filter is proposed for suppressing the unknown and non-stationary noises which exist in monochrome or colour images. This new methodology is based on estimating the unknown noise characteristics by using the generalized Gaussian probability density function (gGf). This function is used as a generic noise distribution model for various types of noise. Depending upon the shape parameter of the gGf, a specific model for the noise is assumed. It has been shown that the shape parameter, which characterizes the gGf, can be found easily by extracting the sample points from the corrupted image. Once the unknown noise distribution is modeled by the gGf, a suitable filter is utilized to smooth out the noise from the input image. Simulation results had shown that the proposed methodology is able to suppress effectively the unknown types of noise in the image.

Original languageEnglish
Pages (from-to)696-699
Number of pages4
JournalCanadian Conference on Electrical and Computer Engineering
Publication statusPublished - 1997
Externally publishedYes
EventProceedings of the 1997 Canadian Conference on Electrical and Computer Engineering. CCECE'97. Part 2 (of 2) - St.John, Can
Duration: 25 May 199728 May 1997


Dive into the research topics of 'Adaptive multichannel filter for image processing'. Together they form a unique fingerprint.

Cite this