image denoisig adaptive wavelet packet thresholdin
no vote
The statistical optimized adaptive wavelet packet (WP) thresholding function for image denoising is based on generalized Gaussian distribution. It is suitable for noisy images to obtain the best tree or wavelet basis, and uses information entropy to calculate efficient multilevel WP decomposition. It selects adaptive thresholds, that is, the level and subband dependence analysis are based on the statistical parameters of subband coefficients. In the used threshold function, based on the maximum a posteriori probability estimation, the optimal linear interpolation between each coefficient and the mean value of the corresponding subband estimates the modified version of the dominant coefficient. Experimental results show that several test images under different noise intensities show that the proposed algorithm, called Open Learning Institute shrinkage, produces better peak signal-to-noise ratio and superior visual image quality - a common image quality indicator used to measure - relative to standard denoising methods, especially at high noise intensities. It is also superior to some of the best state denoising techniques based on wavelet transform.