Apparatus for generating assignments between image regions of an image and element classes

Patent metadata

Field	Value
Publication number	US-9336613-B2
Application number	US-201214117698-A
Country	US
Kind code	B2
Filing date	May 23, 2012
Priority date	May 24, 2011
Publication date	May 10, 2016
Grant date	May 10, 2016

Abstract

An apparatus for generating assignments between image regions of an image of an object and element classes includes an assigning unit ( 13 ) for assigning element classes to image regions of an element image of the object, which is indicative of a distribution of the element classes, depending on region and/or boundary features, which are determined depending on image values of a provided object image and provided first preliminary assignments. Thus, the resulting element image with the assignments to the element classes is not necessarily based on the provided object image only, but can also be based on the provided preliminary assignments. If the quality of the assignments defined by the element image would be restricted due to restrictions of the provided object image, these restrictions of the provided image can therefore be compensated by the preliminary assignments such that the quality of the resulting element image can be improved.

First claim

The invention claimed is:

1. A nuclear imaging apparatus comprising:
a processor configured to:
receive a time-of-flight non-attenuation-corrected positron emission tomography image and a magnetic resonance image,
differentiate the time-of-flight non-attenuation-corrected positron emission tomography image to generate a gradient time-of-flight non-attenuation-corrected positron emission tomography image,
differentiate the magnetic resonance image to generate a gradient magnetic resonance image,
segment the gradient time-of-flight non-attenuation-corrected positron emission tomography image using a watershed algorithm to generate boundaries which distinguish between image regions and applying the boundaries to the time-of-flight non-attenuation-corrected positron emission tomography image, each region representing a class of tissue,
segment the gradient magnetic resonance image using a watershed algorithm to generate boundaries which distinguish between image regions and applying the boundaries to the magnetic resonance image, each region representing a class of tissue,
assign tissue classes to the image regions of the magnetic resonance image,
assign tissue classes and probabilities to the image regions of the segmented time-of-flight non-attenuation-corrected positron emission tomography image based on the tissue classes of the image regions of the time-of-flight non-attenuation-corrected positron emission tomography image and the tissue classes and probabilities of the segmented time-of-flight non-attenuation-corrected positron emission tomography image,
determine attenuation values for each image region and generating and storing an attenuation map,
reconstruct non-attenuation-corrected position emission data using the attenuation map to generate an attenuation-corrected position emission tomographic image.

2. The apparatus as defined in claim 1 , wherein the tissue values assigned to the image regions of the PET image include tissue, air, and probability of tissue or air.

3. The apparatus as defined in claim 1 , wherein the attenuation values are determined for radiation of 511 keV and the non-attenuation-corrected PET image data is time-of-flight non-attenuation corrected positron emission data.

4. An apparatus for generating an attenuation map, the apparatus comprising:
one or more processors configured to:
reconstruct a time-of-flight non-attenuation-corrected positron emission tomography (TOF-NAC-PET) image of a subject from time-of-flight positron emission data,
generate a gradient image by differentiating the TOF-NAC-PET image such that each voxel for the image represents a slope,
segment the gradient image using a watershed algorithm to generate boundaries to distinguish between image regions with the TOF-NAC-PET image, each region representing a class of tissue,
receive a preliminary tissue class image which has been segmented to generate preliminary boundaries and preliminary image regions, the preliminary image regions each identifying a preliminary tissue class,
assigning attenuation values to the image regions of the TOF-NAC-PET image based on tissue classification rates and the preliminary tissue classes for producing the attenuation map.

5. The apparatus as defined in claim 4 , wherein the processor is further configured to:
determine second preliminary assignments between the image regions and tissue classes depending on the determined features of the image regions,
combine the first preliminary assignments and the second preliminary assignments for producing combined assignments indicative of a distribution of the tissue classes.

6. The apparatus as defined in claim 5 , wherein the assigning unit is adapted to apply at least one of a) predefined assignment rules to the determined features for determining the preliminary second assignments, the assignment rules defining the assignment of element classes depending on the features, and b) predefined combination rules to the first and second preliminary assignments for producing the combined assignments, the combination rules defining assignments of tissue classes depending on the first and second preliminary assignments.

7. The apparatus as defined in claim 6 , wherein the combination rules are adapted to be applicable to the first and second preliminary assignments and provided anatomical information of the object, in order to produce the combined assignments depending on the first and second preliminary assignments and the provided anatomical information.

8. The apparatus as defined in claim 4 , wherein the the processor is further configured to:
reconstruct a time-of-flight attenuation-corrected positron emission tomography image based on the time-of-flight positron emission data and the attenuation map.

9. A nuclear imaging system for generating an image of a subject, the nuclear imaging system comprising:
a time-of-flight PET scanner configured to acquire the time-of-flight positron emission data,
the apparatus as defined in claim 8 configured to produce the attenuation map,
a reconstruction processor configured to use the attenuation map in reconstructing the acquired time-of-flight positron emission data into an attenuation-corrected time-of-flight positron emission tomographic image,
a display device configured to display the time-of-flight attenuation-corrected positron emission tomographic image.

10. The apparatus as defined in claim 4 , wherein determining the preliminary tissue class includes determining average image values of the image regions, minimal image values of the image regions, maximal image values of the image regions, a standard deviation of the image values of the image regions, average image values along the boundaries, minimal image values along the boundaries, maximal image values along the boundaries, and a standard deviation of image values along the boundaries.

11. The apparatus as defined in claim 4 , wherein the image regions of the segmented time-of-flight non-attenuation-corrected PET image differentiates between body tissue and air.

12. The apparatus as defined in claim 4 , wherein the class of tissues represented by the image regions of the time-of-flight non-attenuation-corrected PET image include tissue, possibility of tissue, air, and possibility of air.

13. A method for generating assignments between image regions of an object image of an object and element classes, the method comprising:
providing a time-of-flight non-attenuation corrected positron emission tomography image as an object image of the object,
differentiating the object image to generate a gradient image,
segmenting the gradient image into image regions by applying a watershed segmentation to define boundaries between the image regions,
determining features of a) the image regions and b) the boundaries between the image regions based on image values of the object image,
providing a preliminary element image comprising first preliminary assignments of preliminary image regions and tissue classes of each preliminary image region,
assigning tissue classes to image regions of an object image, which is indicative of a distribution of the tissue classes, based on the determined features and the first preliminary assignments,
assigning attenuation values to the image regions based on the assigned tissue classes to generate an attenuation map,
reconstructing positron emission data into an attenuation-corrected positron emission tomographic image using the attenuation map.

14. A nuclear imaging method for generating a nuclear image of an object, the nuclear imaging method comprising:
acquiring nuclear data by a nuclear data scanner,

Assignees

• Blaffert Thomas
• Renisch Steffen
• Schweizer Bernd
• Hu Zhiqiang
• Koninkl Philips Nv

Inventors

• Blaffert Thomas
• Renisch Steffen
• Schweizer Bernd
• Hu Zhiqiang

Classifications

• G06T12/30 Primary

  Image post-processing, e.g. metal artefact correction · CPC title

• G06T11/008 Primary

  Physics · mapped topic

• G06T7/0097

  Physics · mapped topic

• G06T7/0012

  Biomedical image inspection · CPC title

• G06T2207/10088

  Magnetic resonance imaging [MRI] · CPC title