Transformations analysis for image denoising using complex wavelet transform

This paper presents an analysis to image denoising using discrete wavelet transformation(DWT) and complex wavelet transformation(CWT). The affect of image denoising on the variation of image noise variance and the selection of threshold value is analyzed. The image denoising effect based on soft thresholding and hard thresholding is carried out and evaluated for different noise variances, and different threshold values. The quality evaluation of denoising operation is carried out using PSNR observation, and found to be higher for complex wavelet transformation than the DWT approach. The soft thresholding result in higher PSNR for higher noise variance and for the higher threshold values. The denoising performance for CWT is observed to be improved with increase in threshold value under a constant noise effect. In the variation of noise effect for a constant threshold the PSNR decreases, however the CWT is observed to have higher denoising efficiency compared to conventional DWT approach.