A heavy-tailed levy distribution for despeckling ultrasound image
Sima Sahu, Harsh Vikram Singh, Basant Kumar · 2017
A new homomorphic approached, Bayesian based, filter is proposed for de-noising ultrasound image. The presence of speckle noise degrades the quality and detail structure in ultrasound image. This reduces the signal-to-noise ratio in ultrasound image. A heavy-tailed Levy probability distribution function (PDF) is used to model the wavelet coefficients. The noise variance is estimated by exploiting the orthogonal property of wavelet transform. A Bayesian Minimum Mean Absolute Error (MMAE) estimator is used to estimate the noise free wavelet coefficients. The proposed method in this paper is able to reduce the speckle noise and improves the image assessment and visual evaluation. Quality and performance parameters such as Peak Signal to Noise Ratio (PSNR), Structural similarity index (SSIM) and Correlation coefficient (CoC) are used to evaluate the effectiveness of the proposed method. The improvement of PSNR, SSIM and CoC is 10.11%, 1.5% and 0.4% than the standard wavelet based technique.