1. A method for the reduction of artifacts based on an unequal representation of the same material classes in various locations in a three-dimensional image data set, reconstructed from two-dimensional x-ray projection images in which an image datum describing an attenuation value is allocated respectively to a voxel, the method performed by a computer processor and comprising:
locating at least two material class regions in a post-processing step, which receive image data which is homogeneously distributed and lies in an expected material class interval of the attenuation values;
considering a characteristic of the material class regions and determining a smooth homogenization function, which is to be applied to the image data of the entire image data set; and
applying the smooth homogenization function to the image data of the image data set,
wherein the image data set is divided into image slices along a division direction, and
wherein an image slice homogenization function is determined for each image slice and the homogenization function for the entire image data set is derived from the image slice homogenization function.
2. The method as claimed in claim 1, wherein, before carrying out the post-processing step, an algorithm for a reduction of artifacts on the projection images and the image data set has been applied and, the post-processing step is the last step of a post-processing procedure before a representation of the image data set.
3. The method as claimed in claim 1, wherein, before carrying out the post-processing step, an algorithm for a reduction of artifacts on the projection images or the image data set has been applied or the post-processing step is the last step of a post-processing procedure before a representation of the image data set.
4. The method as claimed in claim 1, wherein the three-dimensional image data set, which shows a head of a person, is used.
5. The method as claimed in claim 1, wherein the image slice homogenization function is calculated as a moving average via the image slice homogenization function.
6. The method as claimed in claim 5, wherein the homogenization function is calculated as a moving average via the image slice homogenization function by filters with a low-pass filter in the division direction.
7. The method as claimed in claim 1,
wherein first a focus of the image slice and a target value is determined for each image slice,
wherein a central region lying around the focus is determined for the calculation of the target value, for which an average attenuation value is specified as the target value, or the target value is specified as a value expected for the material class.
8. The method as claimed in claim 7, wherein for the determination of the focus, the image slice is binarized by means of an object threshold value describing an attenuation value of an object represented in the image slice.
9. The method as claimed in claim 7, wherein for the determination of the average attenuation value, only image data of the central region is considered, which lies in a rough estimate interval that is dependent on the material class.
10. The method as claimed in claim 7,
wherein for each image slice, evenly distributed candidate regions or candidate regions covering the entire image slice are defined and for each candidate region an average image datum, a maximum image datum and a minimum image datum are calculated via the pixels in the candidate region, and
wherein a candidate region is determined as a material class region, when the average image datum in which the material class interval lies, the minimum is larger than a predetermined minimum value and the maximum is smaller than a predetermined maximum value, and
wherein the minimum value and the maximum value describe acceptable deviations for the material class.
11. The method as claimed in claim 10,
wherein then, when the total number of the material class regions is smaller than the predetermined first threshold value or the total number of material class regions in a quadrant, defined by the focus, is smaller than a predetermined second threshold value, candidate regions are also used as material class regions, for which the minimum is larger than the predetermined minimum value and the maximum is smaller than the predetermined maximum value, and
wherein for these material class regions the average image datum is specified as the target value.
12. The method as claimed in claim 10,
wherein for each image slice the function parameters describing the image slice homogenization function are calculated in an optimization procedure, and
wherein a cost function is set as a sum of the deviations of the average image data of the target value, multiplied by the homogenization function at the location of the material class region.
13. The method as claimed in claim 12, wherein the image slice homogenization function is set such that it always results in the target value for the focus.
14. The method as claimed in claim 12, wherein, in a case in which the total number of the material class regions is smaller than a predetermined first threshold value or the total number of material class regions in a quadrant, defined by the focus is smaller than a predetermined second threshold value, the image slice homogenization function is specified for the entire image slice as multiplication by one.
15. The method as claimed in claim 1,
wherein first a focus of the image slice or a target value is determined for each image slice,
wherein a central region lying around the focus is determined for the calculation of the target value, for which an average attenuation value is specified as the target value, or the target value is specified as a value expected for the material class,
wherein the target value is a Hounsfield Unit value.
16. The method as claimed in claim 1,
wherein the homogenization function is restricted to an admissibility interval of one, described by a maximum distance, before its multiplicative application to the image data, and
wherein the maximum distance is chosen between 0.1 and 0.3.
17. The method as claimed in claim 16, wherein the maximum distance is 0.2.
18. The method as claimed in claim 1, wherein the homogenization function is only applied to an application interval, which contains the material class interval, of possible attenuation values or the application of the homogenization function is continued in an attenuating manner with increasing distance from the application interval by the smooth continuation outside the application interval.
19. A computing facility for the reduction of artifacts based on an unequal representation of the same material classes in various locations in a three-dimensional image data set, reconstructed from two-dimensional x-ray projection images, wherein an image datum, describing an attenuation value is allocated respectively to a voxel, the computing facility comprising:
a location unit for the location of at least two material class regions containing image data, homogeneously distributed and which lies in an expected material class interval of the attenuation values;
a homogenization function determination unit for the determination of a smooth homogenization function, which is applied to the image data of the entire image data set, considering at least one characteristic of the material class regions; and
a correcting unit for the application of the homogenization function to the image data of the image data set,
wherein the image data set is divided into image slices along a division direction, and
wherein an image slice homogenization function is determined for each image slice and the homogenization function for the entire image data set is derived from the image slice homogenization function.
The claims below are in addition to those above.
All refrences to claim(s) which appear below refer to the numbering after this setence.
1. A semiconductor light-emitting device, comprising:
a plurality of semiconductor light sources and
a driver with no galvanic isolation for operating the semiconductor light sources, wherein
the semiconductor light sources are divided into at least two carriers,
the carriers are applied on an electrically conductive substrate,
and the driver, and current-conducting regions also provided on a surface of the carriers, are electrically insulated from the substrate,
wherein the carriers are electrically connected to one another via at least one unsupported connecting line,
wherein the at least one unsupported connecting line is a connecting line with threefold insulation.
2. The semiconductor light-emitting device as claimed in claim 1, wherein at least two of the carriers are arranged so as to be angled with respect to one another.
3. The semiconductor light-emitting device as claimed in claim 1, wherein an insulating element is provided in a region between two carriers beneath at least one electrical connecting line connecting said carriers.
4. The semiconductor light-emitting device as claimed in claim 1, wherein at least one of the carriers is a ceramic carrier.
5. The semiconductor light-emitting device as claimed in claim 1, wherein at least one of the at least two carriers is applied to the substrate via an electrically insulating interface layer, and the interface layer protrudes laterally beyond the carrier.
6. The semiconductor light-emitting device as claimed in claim 1, wherein the substrate is a surface of a heat sink.
7. A semiconductor light-emitting device as claimed in claim 1, wherein the semiconductor light-emitting device is a retrofit lamp.
8. A semiconductor light-emitting device, comprising:
a plurality of semiconductor light sources and
a driver with no galvanic isolation for operating the semiconductor light sources, wherein
the semiconductor light sources are divided into at least two carriers,
the carriers are applied on an electrically conductive substrate, and the driver, and current-conducting regions also provided on a surface of the carriers, are electrically insulated from the substrate,
wherein at least one of the carriers comprises a printed circuit board basic material.
9. A semiconductor light-emitting device, comprising:
a plurality of semiconductor light sources and
a driver with no galvanic isolation for operating the semiconductor light sources, wherein
the semiconductor light sources are divided into at least two carriers,
the carriers are applied on an electrically conductive substrate,
and the driver, and current-conducting regions also provided on a surface of the carriers, are electrically insulated from the substrate,
wherein the carriers are electrically connected to one another via at least one unsupported connecting line,
wherein at least two carriers are configured as printed circuit board regions of a common flexible printed circuit board.
10. The semiconductor light-emitting device as claimed in claim 9, wherein the at least two printed circuit board regions are arranged on in each case one heat-spreading element, which is arranged on the electrically conductive substrate.
11. The semiconductor light-emitting device as claimed in claim 9, wherein the at least one unsupported connecting line is formed between two printed circuit board regions by at least one narrowed connecting region of the flexible printed circuit board, said connecting region having at least one conductor track.
12. The semiconductor light-emitting device as claimed in claim 10, wherein the at least one unsupported connecting line is formed between two printed circuit board regions by at least one narrowed connecting region of the flexible printed circuit board, said connecting region having at least one conductor track.