Detecting haemorrhagic stroke in CT image data

• US 9,159,127 B2
• Filed: 06/17/2008
• Issued: 10/13/2015
• Est. Priority Date: 06/20/2007
• Status: Active Grant

• Current Assignee: Koninklijke Philips Electronics N.V. (Koninklijke Philips N.V.)
• Original Assignee: Koninklijke Philips N.V.
• Inventors: Meetz, Kirsten, Buelow, Thomas, Young, Stewart
• Primary Examiner(s): Makiya, David J
• Assistant Examiner(s): CORBETT, JOHN M
• Application Number:
  US12/665,422
• Publication Number:
  US 20100183211A1
• Time in Patent Office: 2,674 Days
• Field of Search: 382/131
• US Class Current: 1/1
• CPC Class Codes: A61B 2576/026 for the brain
• CPC Class Codes: A61B 6/032 Transmission computed tomography [CT]
• CPC Class Codes: A61B 6/12 Arrangements for detecting or locating foreign bodies
• CPC Class Codes: A61B 6/501 for diagnosis of the head, e.g. neuroimaging or craniography
• CPC Class Codes: G06T 2207/10081 Computed x-ray tomography [CT]
• CPC Class Codes: G06T 2207/20156 Automatic seed setting
• CPC Class Codes: G06T 2207/30016 Brain
• CPC Class Codes: G06T 2207/30096 Tumor; Lesion
• CPC Class Codes: G06T 2210/41 Medical
• CPC Class Codes: G06T 7/0012 Biomedical image inspection
• CPC Class Codes: G06T 7/44 using image operators, e.g. filters, edge density metrics or local histograms

Abstract

The invention relates to a system (100) arranged to delineate the acute intracerebral haematoma in non-contrasted CT images in two stages. The first stage, performed by the extraction unit (110), employs an analysis of gray values of the image data in order to extract the candidate region. The candidate region may comprise both an acute haematoma and other regions having similar gray values, e.g., regions resulting from partial volume effects at the interface of the bony structures of the skull and the brain. The novel second stage, performed by the classification unit (120), analyzes spatial features of the candidate region such as, for example, the size, shape, and connectedness to the skull bone of the candidate region. Using spatial features of the candidate region improves the correctness of classification of the candidate region as a true or false acute haematoma.

Claims

1. A system for identifying an acute haematoma in non-contrasted computed tomography image data, the system comprising a computer programmed to:

perform a first analysis of gray values of non-contrasted computer tomography image data, the first analysis comprising region growing and thresholding;

extract a candidate region suspected to be the acute haematoma, based on the first analysis of the gray values;

perform a second analysis of spatial features of the candidate region, including a size of the candidate region, a shape of the candidate region, and a connectedness of the candidate region to a skull; and

classify, via a distance-based histogram of a mean gray value of the candidate region, the candidate region as a positive or negative acute haematoma, based on the analysis of spatial features.

2. The system as claimed in claim 1, wherein extracting the candidate region comprises:

extracting a skull region;

extracting a brain region, based on the extracted skull region; and

extracting the candidate region from within the brain region.

3. The system as claimed in claim 1, wherein classifying the candidate region comprises:

computing a topological feature of the candidate region; and

/or

computing a geometrical feature of the candidate region; and

classifying the candidate region on the basis of the topological and/or geometrical feature of the candidate region.

4. The system as claimed in claim 3, wherein the topological and/or geometrical feature of the candidate region is computed, based on the distance-based histogram of the mean gray value of the candidate region.

5. The system as claimed in claim 4, wherein the computer is programmed to:

compute the topology feature by computing connectedness of the candidate region to the skull in accordance with a distance between the candidate region and the skull;

compute the geometry feature by computing dimensions of the candidate region; and

classify the candidate region by classifying the candidate region in accordance with the computed connectedness and the dimensions of the candidate region.

6. The system as claimed in claim 1, wherein the classifying includes classifying the candidate region as one of an acute intracranial haematoma, acute subdural haematoma, acute epidural haematoma, or partial volume effect.

7. The system as claimed in claim 1, further including image acquisition apparatus which generates the non-contrasted computed tomography image data and communicates the computed tomography data to the computer.

8. A method of identifying an acute haematoma in non-contrasted computed tomography image data, the method comprising:

extracting a candidate region suspected to be the acute haematoma, based on a first analysis of gray values of the image data including region growing and thresholding, the region growing and thresholding comprising;

casting rays from a border of the image data,

responsive to a ray contacting an object having a gray value greater than associated skull threshold value and with an extension greater than a preselected distance, setting a seed within the object, and

determining a binary mask of a skull defining a skull region for extraction; and

classifying the candidate region as a positive or negative acute haematoma, based on a second analysis of spatial features of the candidate region, wherein the spatial features include a size of the candidate region, a shape of the candidate region, and a connectedness of the candidate region to the skull.

9. The method as claimed in claim 8, wherein the computer is programmed to classify the candidate region by:

computing a connectedness of the candidate region to the skull in accordance with a distance between the candidate region and the skull;

computing dimensions of the candidate region; and

classifying the candidate region in accordance with the computed connectedness and the dimensions of the candidate region.

10. The method as claim in claim 9, wherein the computer is further programmed to:

calculate a Euclidean distance map of the brain so as to assign to each of a plurality of points of the brain, a Euclidean distance from the point to the skull;

bin each of the plurality of points in accordance with the calculated Euclidean distance; and

calculate a distance-based histogram including each candidate region and each bin in accordance with a mean of gray values for each candidate region and each bin, wherein classification of the candidate region is performed in accordance with the distance-based histogram.

11. The method as claimed in claim 8, further comprising:

setting a seed within a mass center or a geometrical center of the skull region;

determining a connected component of the seed, the connected component comprising voxels having a gray value greater than an associated brain threshold value and less than the associated skull threshold value; and

extracting the determined connected component of the seed as the brain region.

12. The method as claimed in claim 11, further comprising:

identifying a voxel within the brain region having a gray value greater than an associated haematoma threshold value; and

determining a connectivity component of the identified voxel as the candidate region.

13. The method as claimed in claim 12, further comprising:

determining a value of each of the skull threshold, the brain threshold and the haematoma threshold in accordance with a set of training images.

14. A non-transitory computer-readable storage medium storing a computer program which, when executed on a computer configures the computer to perform a method of identifying an acute haematoma in non-contrasted computed tomography image data, including:

extracting a candidate region suspected to be the acute haematoma, based on a first analysis of gray values of the image data, the first analysis including region growing and thresholding; and

classifying the candidate region as a positive or negative acute haematoma, based on a second analysis of spatial features of the candidate region via a distance-based histogram, wherein the spatial features include a size of the candidate region, a shape of the candidate region, and a connectedness of the candidate region to a skull.

15. A system for identifying an acute haematoma in non-contrasted computed tomography image data, the system comprising a computer programmed to:

perform a first analysis of gray values of the non-contrasted computer tomography image data, the first analysis comprising region growing and thresholding;

extract a candidate region suspected to be the acute haematoma, based on the first analysis of the gray values, the extracting comprising;

extracting a skull region by;

casting rays from a border of the image data,

responsive to a ray contacting an object having a gray value greater than an associated skull threshold value and with an extension greater than a preselected distance, setting a seed within the object, and

determining a binary mask of the skull defining the skull region for extraction;

extracting a brain region, based on the extracted skull region; and

extracting the candidate region from within the brain region;

perform a second analysis of spatial features of the candidate region, including a size of the candidate region, a shape of the candidate region, and a connectedness of the candidate region to a skull; and

classify the candidate region as a positive or negative acute haematoma, based on the analysis of the spatial features.

16. The system as claimed in claim 15, wherein the computer is further programmed to extract the brain region, based on the extracted skull region by the brain unit by:

setting a seed within a mass center or a geometrical center of the skull region;

determining a connected component of the seed, the connected component comprising voxels having a gray value than an associated brain threshold value and less than the associated skull threshold value; and

extracting the determined connected component of the seed as the brain region.

17. The system as claimed in claim 16, wherein the computer is programmed to extract the candidate region within the brain region by:

identifying a voxel within the brain region having a gray value greater than an associated haematoma threshold value; and

determining a connectivity component of the identified voxel as the candidate region.