Iterative image reconstruction using dynamic suppression of noise-induced artifact formation

摘要

An iterative reconstruction 20 of the imaging data is performed to generate a series of updated images 22 that end with the reconstructed image. During the iterative reconstruction, at least one of the updated image and the iterative reconstruction parameters is adjusted using an adjustment process separate from the iterative reconstruction. In some embodiments that use edge preserving regularization priors 26, the adjustment process 30 applies edge preserving when the slope steepness is high for later iterations, and slopes compared to earlier iterations. The edge preservation threshold is adjusted so as to reduce the steepness. In some embodiments, the adjustment process determines, for each pixel, voxel, or region of the current update image, whether the development prior to the current update image 22 satisfies the artifact feature criteria (36, 38). If the expansion meets the artifact feature criteria, a local noise suppression operation 40 is performed on the corresponding pixel, voxel, or region, otherwise it is not performed.