Wavelet based adaptive Bayesian despeckling for medical ultrasound images

This paper presents a new wavelet domain adaptive despeckling technique for contrast enhancement of medical ultrasound (US) images for improving the clinical diagnosis. This method uses the Rayleigh distribution for modeling the speckle wavelet coefficients and the signal wavelet coefficients are approximated using the Gaussian distribution. Combining these as statistical priors, we have developed a context modeling based adaptive Bayesian shrinkage estimator for processing the wavelet coefficients of detail subbands. The experimental results using real ultrasound images demonstrate the superiority of the proposed technique both quantitatively and qualitatively as compared to other competitive schemes reported in the ultrasound image denoising literature.