A Novel Adaptive Probabilistic Nonlinear Denoising Approach for Enhancing PET Data Sinogram

Figure 2

Chi-square prior versus Laplacian prior. (a) Histogram of the noise-free gradient. (b) Histogram of the noisy gradient magnitude. (c) Laplacian and Chi-square priors for the noisy-free gradient, estimated from the noisy data.

(a) Noise-free histogram

(b) Noisy gradient histogram

(c) Chi-square and Laplacian prior for the noise free histogram