Computer-implemented method for determining a subtraction image, image recording facility, computer program, and electronically readable data carrier
A computer-implemented method for determining a subtraction image from a feature image and a plurality of mask images of an imaging area of an examination object is provided. A combination of the plurality of mask images weighted by mask weights is subtracted from the feature image to determine the subtraction image. The mask weights are determined in an optimization process for optimizing an image quality function for test images calculated by subtracting a combination of the mask images determined with a test set of mask weights from the feature image. The image quality function is minimized in the form of a convex quadratic program of a dimension corresponding to the number of mask images.
1 . A method for determining a subtraction image from a feature image and a plurality of mask images of an imaging area of an examination object, the method being computer-implemented and comprising:
subtracting a combination of the plurality of mask images weighted by mask weights from the feature image, such that the subtraction image is determined,
wherein the mask weights are determined in an optimization process for optimizing an image quality function for test images calculated by subtracting a combination of the plurality of mask images determined with a test set of mask weights from the feature image, and
wherein the image quality function is minimized in the form of a convex quadratic program of a dimension corresponding to a number of mask images of the plurality of mask images.
2 . The method of claim 1 , wherein an L2 standard is used as the image quality function.
3 . The method of claim 2 , wherein the convex quadratic program is formulated by multiplying subtraction terms for the test images and separating a quadratic term and a linear term in a vector of the mask weights.
4 . The method of claim 2 , wherein the L2 standard is weighted by standard weights for each pixel,
wherein the standard weights are determined by an image analysis process of the feature image and the plurality of mask images, based on additional information, or by the image analysis process of the feature image and the plurality of mask images and based on additional information.
5 . The method of claim 4 , wherein in the image analysis process:
maximum deviations of image values of the feature image from image values of the plurality of mask images are determined for each pixel; and
lower standard weights are assigned to at least some of the pixels with the greatest maximum deviations.
6 . The method of claim 5 , wherein pixels having a maximum deviation that is above a threshold value are assigned a standard weight of zero, the standard weights are determined according to a monotonically decreasing function at least in a functional range extending around the threshold value, or a combination thereof.
7 . The method of claim 5 , wherein the standard weights in the image analysis process are determined at least partially using a trained image analysis function.
8 . The method of claim 4 , wherein the additional information describes a region of interest of higher weighting in the imaging area.
9 . The method of claim 1 , wherein the feature image and the mask images are pre-processed for the optimization process by an edge-sharpening image processing operation.
10 . The method of claim 9 , wherein in the image processing operation, a gradient image is determined for the feature image and for each of the mask images, and an edge-sharpened respective image is determined by weighted linear combination of the original respective image and the associated gradient image.
11 . The method of claim 1 , wherein the optimization process is carried out by a Goldfarb-Idnani method.
12 . The method of claim 1 , wherein the feature image is a filling image of digital subtraction angiography with a contrast medium.
13 . The method of claim 1 , wherein the image quality function includes a regularization of the weights with the aim of counteracting an uneven distribution of the weights.
14 . The method of claim 13 , wherein the regularization is performed as a function of the Euclidean norm of the weights.
15 . An image recording facility comprising:
a control facility comprising:
a recording unit configured to control a recording of a plurality of mask images and a feature image of an imaging area of an examination object;
a determination unit configured to determine a subtraction image, the determination of the subtraction image comprising subtraction of a combination of the mask images weighted by mask weights from the feature image; and
an optimization unit configured to determine the mask weights in an optimization process for optimizing an image quality function for test images calculated by subtracting a combination of the mask images determined with a test set of mask weights from the feature image,
wherein the optimization unit for optimizing the image quality function is configured in the form of a convex quadratic program of a dimension corresponding to a number of mask images.
16 . A non-transitory computer-readable storage medium that stores a computer program having instructions executable by a computing facility to determine a subtraction image from a feature image and a plurality of mask images of an imaging area of an examination object, the instructions comprising:
subtracting a combination of the plurality of mask images weighted by mask weights from the feature image, such that the subtraction image is determined,
wherein the mask weights are determined in an optimization process for optimizing an image quality function for test images calculated by subtracting a combination of the plurality of mask images determined with a test set of mask weights from the feature image, and
wherein the image quality function is minimized in the form of a convex quadratic program of a dimension corresponding to a number of mask images of the plurality of mask images.