Color Image Denoising with Multi-channel Circular Spatial Filtering

Abstract

A multi-channel circular spatial filter (MCSF) in YCbCr-color space is developed for color image denoising. In [2], Bhoi and Meher proposed and demonstrated that circular spatial filter performed very well for efficient suppression of additive white Gaussian noise (AWGN) from gray images. Here multi-channel versions of the circular spatial filter are proposed and developed. The developed filter is based on three-channel processing (e.g., RGB-processing, YCbCr-processing, etc.). The YCbCr version of MCSF is observed to yield quite superior performance as compared to the other versions in RGB, CMY and CIE Lab color spaces. The proposed filter outperforms the multi-channel versions of existing filters in terms of objective and subjective evaluation measures.