Spectral CT image reconstructing method and spectral CT imaging system

Patent metadata

Field	Value
Publication number	US-10489939-B2
Application number	US-201615580005-A
Country	US
Kind code	B2
Filing date	Aug 22, 2016
Priority date	Sep 9, 2015
Publication date	Nov 26, 2019
Grant date	Nov 26, 2019

Abstract

A spectral CT image reconstructing method includes: collecting incomplete original projection data in each of a plurality of energy windows; performing a projection data cross estimation using corresponding original projection data in at least one pair of energy windows constituted by different energy windows of the plurality of energy windows to obtain estimated projection data, wherein each pair of energy windows comprises a first energy window and a second energy window; combining the original projection data and the corresponding estimated projection data to obtain complete projection data; and reconstructing a spectral CT image using the complete projection data.

First claim

The invention claimed is:

1. A spectral CT image reconstructing method, comprising:
collecting incomplete original projection data in each of a plurality of energy windows;
performing a projection data cross estimation using corresponding original projection data in at least one pair of energy windows constituted by different energy windows of the plurality of energy windows to obtain estimated projection data, wherein each pair of energy windows comprises a first energy window and a second energy window;
combining the original projection data with the corresponding estimated projection data to obtain complete projection data; and
reconstructing a spectral CT image using the complete projection data,
wherein the performing a projection data cross estimation using corresponding original projection data in at least one pair of energy windows constituted by different energy windows of the plurality of energy windows to obtain estimated projection data comprises:
establishing, on an image domain, a mapping relation between image data in the first energy window and image data in the second energy window;
obtaining estimated image data using the mapping relation; and
obtaining the estimated projection data from the estimated image data, and
wherein the establishing, on an image domain, a mapping relation between image data in the first energy window and image data in the second energy window comprises:
extracting, from the corresponding original projection data of the first energy window and of the second energy window, first identical projection data in the first energy window and second identical projection data in the second energy window passing through an identical ray path;
reconstructing using the first identical projection data to obtain first training image data;
reconstructing using the second identical projection data to obtain second training image data; and
establishing, by means of a machine learning method using the first training image data and the second training image data, a first mapping relation from the image data in the first energy window to the image data in the second energy window and/or a second mapping relation from the image data in the second energy window to the image data in the first energy window.

2. The spectral CT image reconstructing method according to claim 1 , wherein obtaining estimated image data using the mapping relation comprises:
respectively obtaining first initial image data and/or second initial image data by reconstructing in advance using the corresponding original projection data in the first energy window and the second energy window; and
obtaining second estimated image data in the second energy window by mapping the first initial image data by means of the first mapping relation, and/or obtaining first estimated image data in the first energy window by mapping the second initial image data by means of the second mapping relation.

3. The spectral CT image reconstructing method according to claim 2 , wherein the obtaining the estimated projection data from the estimated image data comprises:
obtaining first estimated projection data using the first estimated image data and a projection operator used in the first energy window, and/or obtaining second estimated projection data using the second estimated image data and a projection operator used in the second energy window.

4. The spectral CT image reconstructing method according to claim 1 , wherein the collecting incomplete original projection data in each of a plurality of energy windows comprises limited angle sampling, angle sparse sampling, or sparse detector sampling.

5. A spectral CT imaging system, comprising:
a ray generating device, comprising a ray source;
a photon-counting detector or an energy-sensitive detector, comprising a plurality of pixels;
a data collecting system, configured to perform data collection on a photon passing through an imaged object from the photon-counting detector or energy-sensitive detector; and
an image reconstruction system, configured to reconstruct a spectral CT image using data acquired by the data collecting system,
wherein the data collecting system is configured to collect incomplete original projection data in each of a plurality of energy windows; and
the image reconstruction system comprises a processor configured to:
perform a projection data cross estimation using corresponding original projection data in at least one pair of energy windows constituted by different energy windows of the plurality of energy windows to obtain estimated projection data, wherein each pair of energy windows comprises a first energy window and a second energy window;
combine the original projection data with the corresponding estimated projection data to obtain complete projection data; and
reconstruct the spectral CT image using the complete projection data,
wherein the processor is further configured to:
establish, on an image domain, a mapping relation between image data in the first energy window and image data in the second energy window;
obtain estimated image data using the mapping relation; and
obtain the estimated projection data from the estimated image data, and
wherein the processor configured to establish, on an image domain, a mapping relation between image data in the first energy window and image data in the second energy window is configured to:
extract, from the corresponding original projection data of the first energy window and of the second energy window, first identical projection data in the first energy window and second identical projection data in the second energy window passing through an identical ray path;
reconstruct using the first identical projection data to obtain first training image data;
reconstruct using the second identical projection data to obtain second training image data; and
establish, by means of a machine learning method using the first training image data and the second training image data, a first mapping relation from the image data in the first energy window to the image data in the second energy window and/or a second mapping relation from the image data in the second energy window to the image data in the first energy window.

6. The spectral CT imaging system according to claim 5 , wherein the processor configured to obtain estimated image data using the mapping relation is configured to:
respectively obtain first initial image data and/or second initial image data by reconstructing in advance using the corresponding original projection data in the first energy window and the second energy window; and
obtain second estimated image data in the second energy window by mapping the first initial image data by means of the first mapping relation, and/or obtain first estimated image data in the first energy window by mapping the second initial image data by means of the second mapping relation.

7. The spectral CT imaging system according to claim 6 , wherein the processor configured to obtain the estimated projection data from the estimated image data is configured to:
obtain first estimated projection data using the first estimated image data and a projection operator used in the first energy window, and/or obtain second estimated projection data using the second estimated image data and a projection operator used in the second energy window.

8. A non-transitory computer-readable storage medium having stored therein instructions that, when executed by a processor of a spectral CT imaging system, causes the system to perform a spectral CT image reconstructing method, the method comprising:
collecting incomplete original projection data in each of a plurality of energy windows;
performing a projection data cross estimation using correspond

Assignees

• Univ Tsinghua
• Nuctech Co Ltd

Inventors

• Xing Yuxiang
• Zhang Li
• Zhang Huayu
• Shen Le

Classifications

• G06T12/10 Primary

  Image preprocessing, e.g. calibration, positioning of sources or scatter correction · CPC title

• G06T2207/20081

  Training; Learning · CPC title

• G06T2211/408

  Dual energy · CPC title

• G06T7/0012

  Biomedical image inspection · CPC title

• G06T2207/10081

  Computed x-ray tomography [CT] · CPC title