Method for processing a multispectral image dataset
AssigneeSiemens Healthineers AG
InventorRainer RAUPACH
Life of this patent
Estimating the term…
Abstract
One or more example embodiments relates to a method for artifact correction, the method comprising transforming primary spectral data; reducing noise from the transformed spectral data; determining artifact images based on the noise-reduced transformed spectral data; transforming the artifact images; and correcting the primary spectral data based on the transformed artifact images.
Claims20
A computer-implemented method for artifact correction, the method comprising:
- transforming primary spectral data;
- reducing noise from the transformed spectral data;
- determining artifact images based on the noise-reduced transformed spectral data;
- transforming the artifact images;
- and
- correcting the primary spectral data based on the transformed artifact images.
A computer-implemented method for processing an image dataset which is based on multispectral computed tomography imaging, the image dataset having a first image value tuple and a second image value tuple, wherein the first image value tuple is assigned to a first volume element of an area to be imaged, wherein the second image value tuple is assigned to a second volume element of the area to be imaged, the method comprising:
- calculating a first portion of a base material by applying a mapping to the first image value tuple, wherein the first portion of the base material is assigned to the first volume element of the area to be imaged, wherein the mapping to the first image value tuple has a first gradient and the first gradient is non-zero;
- and
- calculating a second portion of the base material by applying the mapping to the second image value tuple, wherein the second portion of the base material is assigned to the second volume element of the area to be imaged, wherein the mapping at the second image value tuple has a second gradient, the second gradient is non-zero and not equal to the first gradient.
Dependencies are inferred from the claim text.