Search

CN-121999093-A - Image representation in a limited angle tomography method

CN121999093ACN 121999093 ACN121999093 ACN 121999093ACN-121999093-A

Abstract

For image representation in a limited angle tomography method, a plurality of projection images representing an object (8) are obtained and a three-dimensional image reconstruction is generated based on the plurality of projection images. A first slice image (7) for a first slice plane through the object (8) is generated on the basis of the image reconstruction, and a region of interest (9) of the object (8) is determined in the first slice image (7). In the first cutting plane, the intersection (10, 10a, 10b, 10c, 10 d) of a second cutting plane is determined, which extends through the region of interest (9). On the basis of the position of the intersection and/or on the basis of the geometric extension of the region of interest (9) in the first slice image (7), a calculation region (11, 11a, 11b, 11c, 11 d) is determined around the second slice plane. A second slice image for the second slice plane is generated from the calculated region based on the plurality of projection images.

Inventors

  • S. Kellenberg
  • A. Preus

Assignees

  • 西门子医疗股份公司

Dates

Publication Date
20260508
Application Date
20251106
Priority Date
20241108

Claims (15)

  1. 1. A computer-implemented method for image representation in a limited angle tomography method, wherein -Obtaining a plurality of projection images representing the object (8) and corresponding to a plurality of different projection directions, and generating a three-dimensional image reconstruction based on the plurality of projection images; -generating a first slice image (7) for a first slice plane through the object (8) based on the image reconstruction, and determining a region of interest (9) of the object (8) in the first slice image (7); -determining in the first cutting plane the intersection (10, 10a, 10b, 10c, 10 d) of a second cutting plane extending through the region of interest (9); -determining a calculation region (11, 11a, 11b, 11c, 11 d) around the second slice plane as a function of the position of the intersection line (10, 10a, 10b, 10c, 10 d) and/or as a function of the geometrical extension of the region of interest (9) in the first slice image (7), wherein the calculation region (11, 11a, 11b, 11c, 11 d) has a position and a slice thickness, and -Generating a second slice image for the second slice plane from the calculation region (11, 11a, 11b, 11c, 11 d) based on the plurality of projection images.
  2. 2. The computer-implemented method of claim 1, wherein the second cutting plane is perpendicular to the first cutting plane.
  3. 3. The computer-implemented method of claim 2, wherein -Determining a first boundary position of the region of interest (9) with maximum extension in a direction perpendicular to the intersection line (10, 10a, 10b, 10c, 10 d) and/or a second boundary position of the region of interest (9) with maximum extension in a direction perpendicular to the intersection line (10, 10a, 10b, 10c, 10 d), and -The calculation region (11, 11a, 11b, 11c, 11 d) is determined from the distance of the intersection line (10, 10a, 10b, 10c, 10 d) to the first boundary position and/or the distance of the intersection line (10, 10a, 10b, 10c, 10 d) to the second boundary position.
  4. 4. The computer-implemented method according to any of the preceding claims, wherein the position of the calculation region (11, 11a, 11b, 11c, 11 d) is determined from the position of the intersection line (10, 10a, 10b, 10c, 10 d).
  5. 5. A computer-implemented method according to claims 4 and 3, wherein the position of the intersection line (10, 10a, 10b, 10c, 10 d) corresponds to the first boundary position or the second boundary position, and the position of the calculation region (11, 11a, 11b, 11c, 11 d) is determined such that the part of the calculation region (11, 11a, 11b, 11c, 11 d) located outside the region of interest (9) is smaller than the part of the calculation region (11, 11a, 11b, 11c, 11 d) located within the region of interest (9), or such that the calculation region (11, 11a, 11b, 11c, 11 d) is located entirely within the region of interest (9).
  6. 6. The computer-implemented method according to any one of claims 1 to 4, wherein the position and slice thickness of the calculation region (11, 11a, 11b, 11c, 11 d) are determined such that the portion of the calculation region (11, 11a, 11b, 11c, 11 d) that is located outside the region of interest (9), independently of the position of the intersection line (10, 10a, 10b, 10c, 10 d), is smaller than the portion of the calculation region (11, 11a, 11b, 11c, 11 d) that is located within the region of interest (9), or such that the position of the calculation region (11, 11a, 11b, 11c, 11 d), independently of the position of the intersection line (10, 10a, 10b, 10c, 10 d), is located entirely within the region of interest (9).
  7. 7. The computer-implemented method according to any of the preceding claims, wherein the slice thickness of the calculation region (11, 11a, 11b, 11c, 11 d) is determined from the position of the intersection line (10, 10a, 10b, 10c, 10 d).
  8. 8. The computer-implemented method according to any one of the preceding claims, wherein the slice thickness of the calculation region (11, 11a, 11b, 11c, 11 d) is determined from the geometrical extension of the region of interest (9) in the first cross-sectional plane.
  9. 9. The computer-implemented method according to any one of the preceding claims, wherein data of the plurality of projection images corresponding to the calculation region (11, 11a, 11b, 11c, 11 d) are extracted and the second slice image is generated based on the extracted data.
  10. 10. The computer-implemented method according to any one of claims 1 to 8, wherein data of the image reconstruction corresponding to the calculation region (11, 11a, 11b, 11c, 11 d) are extracted and the second slice image is generated based on the extracted data.
  11. 11. The computer-implemented method according to any one of the preceding claims, wherein the plurality of different projection directions includes a projection direction perpendicular to the first cutting plane and/or the second cutting plane is parallel to all projection directions of the plurality of different projection directions.
  12. 12. A data processing system (6) arranged to perform the computer-implemented method according to any of the preceding claims.
  13. 13. An imaging system (1) for limited angle tomography, having a data processing system (6) according to claim 12 and an imaging device (3) arranged for generating the plurality of projection images.
  14. 14. Imaging system (1) according to claim 13, wherein the imaging device (3) is configured as an X-ray C-arm device.
  15. 15. A computer program product having instructions which, when executed by a data processing system (6), cause the data processing system (6) to perform the computer-implemented method according to any of claims 1 to 11.

Description

Image representation in a limited angle tomography method Technical Field The invention relates to a computer-implemented method for image representation in a limited angle tomography method, in which method a plurality of projection images representing an object and corresponding to a plurality of different projection directions are obtained and a three-dimensional image reconstruction is generated based on the plurality of projection images. The invention also relates to a data processing system for carrying out such a computer-implemented method, an imaging system having such a data processing system and a corresponding computer program product. Background The limited angle tomographic method (english: LIMITED ANGLE tomograph, LAT) is an X-ray imaging method in which a series of slice images can be generated from a limited number of two-dimensional projection images. The limited angle tomography method is also referred to herein and below as a tomosynthesis method. Tomosynthesis methods enable three-dimensional visualization of anatomical structures with lower radiation doses than, for example, computed tomography CT or cone beam CT. By recording a plurality of projection images around the patient from different angles, i.e. with different projection directions, a series of slice images can be generated which can be reconstructed as a three-dimensional image reconstruction, also called a volumetric dataset. This enables better visualization of complex structures, such as the lungs, and allows radiologists to identify and characterize lesions with lower radiation doses. In conventional tomosynthesis methods, the image quality achieved is often insufficient for image-guided interventions, such as bronchoscopy. This limitation is due to the limited angular acquisition and generally a smaller number of projection directions, which may lead to geometrical distortions and streak artifacts. Methods of compensating for missing projection using a deep learning algorithm have been disclosed, for example in y. One of the drawbacks is the large amount of annotated training data that may be problematic in a clinical setting due to data protection regulations, etc. Alternatively, a registration and reconstruction method is proposed by F.Saad et al ："Deformable 3D/3D CT-to-digital-tomosynthesis image registration in image-guided bronchoscopy interventions." Comput. Biol. Med. 171: 108199 to improve tomosynthesis image quality by using a diagnostic CT scan recorded prior to the tomosynthesis scan. In this case, the previous CT image is matched to the intraoperative slice-synthesis image and used as an initial estimate for reconstruction. Additional CT scans increase the total applied radiation dose. Disclosure of Invention The object of the present invention is to provide a possibility of image representation in a limited angle tomography, by means of which the image quality can be improved and the above-mentioned disadvantages can be overcome or reduced. The invention is based on the insight that due to the principle of limited angle data acquisition, the degree of significance of the artifacts is not the same for all line of sight directions or the cross sectional planes through the mapped object. It is therefore proposed to determine a region of interest of the object in a first slice plane through the object and to determine a calculation region from the region of interest thus determined, which calculation region is used to generate slice images from a second slice plane different from the first slice plane. According to one aspect of the present invention, a computer-implemented method for image representation in a limited angle tomography method is provided. Wherein a plurality of projection images representing the object and corresponding to a plurality of different projection directions are obtained and a three-dimensional image reconstruction is generated based on the plurality of projection images. A first slice image for a first slice plane, in particular a predetermined first slice plane, through the object is generated on the basis of the image reconstruction, and a region of interest of the object is determined in the first slice image. In the first cutting plane, a second cutting plane, in particular an intersection of a predetermined second cutting plane, is determined, which intersects the first cutting plane and extends through the region of interest, in particular in the first cutting plane. A calculation region surrounding the second plane is determined from the position of the intersection line in the first cutting plane, in particular the position of the intersection line relative to the region of interest, and/or from the geometric extension of the region of interest in the first slice image, wherein the calculation region has a position and a slice thickness. Based on the plurality of projection images, a second slice image for the second slice plane is generated from the calculation region and