Performance of Wavelet, Contourlet and 2D PCA Transforms for Image Denoising

This paper uses the wavelet, Contourlet transform and the two dimensional Principle Component Analysis (2DPCA) for image denoising. The simulation results is carried out to demonstrate the performance of proposed techniques and the result shows that 2DPCA is superior over contourlet and wavelet transform in maintaining high peak signal -to-noise -ratio. The denoising algorithm is validated by numerical experiments on different standard images and roughness images. Numerical result shows that the proposed method can obtain higher PSNR and less visual artifacts compared with other methods. Also, Contourlet can provide a much more detailed representation for natural images with abundant textural information than Wavelets. The improved two dimensional PCA (2DPCA) extended the traditional principal component analysis theory to the two dimensional situation, and proved to be especially efficient in the analysis of images.