Wavelet Based Speckle Noise Reduction by Using Statistical Method

In image processing, image de-noising has become an essential task for many researchers. Mainly ultrasound images contain speckle noise which degrades the quality of the images. Eliminating such kind of noise is an important preprocessing task. A wavelet based thresholding scheme for noise suppression in images by using statistical method is purposed in this paper. Our method helps to find out the optimum threshold value by analyzing the statistical parameters of the wavelet sub-band coefficients like standard deviation, arithmetic mean and geometrical mean. The noisy image is decomposed into 2 levels to obtain the different frequency bands. Then consequently soft thresholding method is applied to remove the noisy coefficients, by calculating the optimum threshold by the purposed method. Experimental results on several test images by using the proposed method show that, the proposed method yields significantly better image quality and better Peak Signal to Noise Ratio (PSNR). Quantitative and qualitative comparisons of the results obtained by the proposed method with the results achieved from the other speckle noise reduction techniques demonstrate its better performance for speckle reduction.