METHOD AND SYSTEM FOR IMAGE DENOISING USING DISCRETE TOTAL VARIATION (TV) MINIMIZATION WITH ONE-DIRECTION CONDITION

In the CT imaging system, a one-direction condition improves noise uniformity in the denoised images and avoids over smoothing in the low noise regions in an image, assuming that the image originally has an unequal noise distribution. On the other hand, the discrete gradients of total variation (DTV) minimization results in an improved edge preserving effects in comparison to the conventional total variation (TV) minimization. Using DTV, the pixel values on an edge will not be substantially affected after a certain denoising treatment. The difference between the DTV minimization and the conventional TV minimization is substantially negligible for strong edges while the DTV minimization substantially improves in preserving the weak edges. DTV keeps the original values while TV slightly smoothes the values for the pixel near the bottom and top of edges.