A Robust Image Denoising Technique in the Contourlet Transform Domain

The contourlet transform can very well capture the directional geometrical structures of natural images. In this paper, by incorporating the ideas of Stein’s unbiased risk estimator (SURE) approach in nonsubsampled contourlet transform (NSCT) domain, a new image denoising technique is devised. We utilize the properties of NSCT coefficients in high and low subbands and apply SURE shrinkage and bilateral filter respectively. Moreover, SURE-LET strategy is modified to minimize the estimation of the mean square error (MSE) between the clean image and the denoised one in the NSCT domain. The simulation testing has been carried on under the different noise level, and the denoising effect has been evaluated by using the peak signal to noise ratio (PSNR). Results for different kinds of sample image show that the introduced algorithm in this paper can maintain most important details of images, remove Gaussian white noise more effectively, and get a higher PSNR value, which also has a better visual effect.

Keywords:

Denoising , Contourlet Transform , Bilateral Filter , SURE Shrink , Multiresolution Analysis , Universal Index , Image Fidelity , Mean Square Error