Figure 3a. Image Comparison for Noise Coefficients Modeled as Laplacian

The clean image was obtained by averaging 25 images of the carotid. The test image is shown in the upper-right. The 2 bottom images are the results of denoising using the Wiener filter with a 3x3 block, and denoising using the Bayesian wavelet coring technique, with the noise coefficients modeled as a Laplacian distribution. There is a large amount of blurring in the cored image. Contrast this result with that shown in Figure 3a, which assumes the noise coefficients are distributed as a Gaussian.

Figure 3a. Image Comparison for Noise Coefficients Modeled as Gaussian

The clean image was obtained by averaging 25 images of the carotid. The test image is shown in the upper-right. The 2 bottom images are the results of denoising using the Wiener filter with a 3x3 block, and denoising using the Bayesian wavelet coring technique, with the noise coefficients modeled as a Gaussian distribution.