Method for distinguishing between gray matter and white matter and CT system for carrying out the method
A method for distinguishing between gray matter and white matter starting from a time-dependent computed tomography image data record from a perfusion CT examination is disclosed. In at least one embodiment, a plurality of time-independent images are calculated from the time-dependent image data record, a plurality of threshold histogram analyses are performed in order to determine regions of the brain which can be assigned to one or more types of cerebral matter, and subsequently the region of gray matter is determined from the information obtained in respect of type and region of the cerebral matter using at least one logical combination and at least one exclusion method. A control and computational unit is also disclosed with a storage medium in which a computer program or program module is stored, which executes the described method during operation.
1. A method for determining a region of gray matter in a brain starting from a time-dependent computed tomography image data record from a perfusion CT examination, the method comprising:
defining a work region in the time-dependent computed tomography image data record which only comprises healthy gray matter, healthy white matter and ischemic gray matter;
calculating a plurality of three-dimensional time-independent images from the time-dependent computed tomography image data record, the plurality of three-dimensional time-independent images including an average image with and without a contrast agent being present, a maximum value image over time and a baseline image without a contrast agent being present;
determining a normalized difference image from the image data of a difference image, the difference image being determined from the average image and the maximum value image;
generating a high-contrast image with increased contrast between white matter and gray matter from the average image, the maximum value image and the normalized difference image;
determining a first grayscale value range, which only comprises gray matter, and a second grayscale value range, which only comprises white matter, by threshold histogram analysis of the high-contrast image, wherein only the image data of an undamaged hemisphere of the brain is considered;
determining a third grayscale value range which comprises both the healthy white matter and the ischemic gray matter by threshold histogram analysis of the normalized difference image;
determining a fourth grayscale value range which comprises both the healthy gray matter and the ischemic gray matter by threshold histogram analysis of the baseline image;
generating a fifth grayscale value range which lies in the damaged hemisphere and only comprises the ischemic gray matter by a logical combination of the third grayscale value range and the fourth grayscale value range;
determining seed points for respectively the first grayscale value range, the second grayscale value range and the fifth grayscale value range;
performing a region growing method respectively starting at the previously determined seed points and generating a first image region, a second image region, and a fifth image region;
generating an expanded first image region by adding cerebral regions which constitute anatomically assured gray matter to the first image region; and
determining a resultant image region which comprises both the healthy gray matter and the ischemic gray matter and all regions of the brain which have not yet been acquired by way of a logical combination of the changed first image region, the second image region and the fifth image region.
2. The method as claimed in claim 1 , wherein the time-independent images are registered to the time-dependent computed tomography image data record.
3. The method as claimed in claim 1 , wherein a region which, starting from a pia mater, reaches one centimeter into the cerebral matter is used as the anatomically assured gray matter.
4. The method as claimed in claim 1 , wherein a formula is used to calculate the average image A, the formula being:
A
(
x
,
y
,
z
)
=
1
N
∑
t
=
t
1
t
=
t
N
I
(
x
,
y
,
z
,
t
)
,
wherein t is a time of a recording, N is a total number of records and I is an intensity at the time of the respective recording.
5. The method as claimed in claim 1 , wherein a formula is used to calculate the maximum value image T, the formula being:
T(x,y,z)=max I(x,y,z,t),
wherein t is a time of a recording and I is an intensity at the time of the respective recording.
6. The method as claimed in claim 1 , wherein a formula is used to calculate the baseline image B, the formula being:
B
(
x
,
y
,
z
)
=
1
M
∑
t
=
t
1
t
=
t
M
I
(
x
,
y
,
z
,
t
)
,
wherein t is a time of the recording, M is a number of records without the contrast agent and I is an intensity at the time of the respective recording.
7. The method as claimed in claim 1 , wherein the image data of the average image is subtracted from the image data of the maximum value image in order to determine the difference image.
8. The method as claimed in claim 1 , wherein a lower bound and an upper bound as well as an average maximum grayscale value are determined in the difference image.
9. The method as claimed in claim 8 , wherein the grayscale values are divided by the previously determined average maximum grayscale value in order to normalize the difference image.
10. The method as claimed in claim 1 , wherein a formula is used to generate the high-contrast image K, the formula being:
K
(
x
,
y
,
z
)
=
{
D
*
(
x
,
y
,
z
)
≤
D
min
;
A
(
x
,
y
,
z
)
D
*
(
x
,
y
,
z
)
≥
D
max
;
T
(
x
,
y
,
z
)
otherwise
;
[
T
(
x
,
y
,
z
)
·
D
*
(
x
,
y
,
z
)
]
+
[
A
(
x
,
y
,
z
)
·
(
1
-
D
*
(
x
,
y
,
z
)
)
]
.
11. The method as claimed in claim 1 , wherein the threshold histogram analysis in order to determine the first and the second grayscale value range is performed under an assumption that the histogram corresponds to a unimodal Gaussian distribution.
12. The method as claimed in claim 11 , wherein a maximum likelihood method is applied in order to determine parameters of the unimodal Gaussian distribution.
13. The method as claimed in claim 12 , wherein an upper and a lower threshold are determined starting from the parameters, and
the image data lies above the upper threshold within the first grayscale value range and the image data lies below the lower threshold within the second grayscale value range.
14. The method as claimed in claim 12 , wherein the parameters of the unimodal Gaussian distribution include the maximum value and a standard deviation.
15. The method as claimed in claim 1 , wherein only one threshold which corresponds to an average grayscale value of the histogram is used in the threshold histogram analysis in order to determine the third grayscale value range, and
the image data lies below the threshold within the third grayscale value range.
16. The method as claimed in claim 1 , wherein the threshold histogram analysis, in order to determine the fourth grayscale value range, is performed under an assumption that the histogram corresponds to a unimodal Gaussian distribution.
17. The method as claimed in claim 16 , wherein a maximum likelihood method is applied in order to determine parameters of the unimodal Gaussian distribution.
18. The method as claimed in claim 17 , wherein an upper threshold is determined starting from the parameters, and
the image data of the fourth image region lies above the threshold.
19. The method as claimed in claim 17 wherein the parameters of the unimodal Gaussian distribution include the maximum value and a standard deviation.
20. The method as claimed in claim 1 , wherein a logical AND operator is utilized in the logical combination for extracting the fifth grayscale value range.
21. The method as claimed in claim 1 , wherein two different termination criteria are used in the region growing method.
22. The method as claimed in claim 21 , wherein a static termination criterion is used.
23. The method as claimed in claim 21 , wherein a dynamic termination criterion is used.
24. The method as claimed in claim 23 , wherein the dynamic termination criterion is used in the region growing method in order to generate the fifth image region.
25. The method as claimed in claim 1 , wherein the resultant image region is determined using an exclusion operator.
26. The method as claimed in claim 25 , wherein the exclusion operator corresponds to a formula:
A=RG ig +RG* gg +R G with R G =G−└RG ig +RG* gg +RG gw ┘,
wherein G corresponds to an entire cerebral tissue and R G corresponds to all regions of the brain that have not previously been acquired.
27. A control and computational unit for determining a region of gray matter in a brain starting from a time-dependent computed tomography image data record from a perfusion CT examination, the unit comprising:
a processor configured to
define a work region in the time-dependent computed tomography image data record which only comprises healthy gray matter, healthy white matter and ischemic gray matter;
calculate a plurality of three-dimensional time-independent images from the time-dependent computed tomography image data record, the plurality of three-dimensional time-independent images including an average image with and without a contrast agent being present, a maximum value image over time and a baseline image without a contrast agent being present;
determine a normalized difference image from the image data of a difference image, the difference image being determined from the average image and the maximum value image;
generate a high-contrast image with increased contrast between white matter and gray matter from the average image, the maximum value image and the normalized difference image;
determine a first grayscale value range, which only comprises gray matter, and a second grayscale value range, which only comprises white matter, by threshold histogram analysis of the high-contrast image, wherein only the image data of an undamaged hemisphere of the brain is considered;
determine a third grayscale value range which comprises both the healthy white matter and the ischemic gray matter by threshold histogram analysis of the normalized difference image;
determine a fourth grayscale value range which comprises both the healthy gray matter and the ischemic gray matter by threshold histogram analysis of the baseline image;
generate a fifth grayscale value range which lies in the damaged hemisphere and only comprises the ischemic gray matter by a logical combination of the third grayscale value range and the fourth grayscale value range;
determine seed points for respectively the first grayscale value range, the second grayscale value range and the fifth grayscale value range;
perform a region growing method respectively starting at the previously determined seed points and generating a first image region, a second image region, and a fifth image region;
generate an expanded first image region by adding cerebral regions which constitute anatomically assured gray matter to the first image region; and
determine a resultant image region which comprises both the healthy gray matter and the ischemic gray matter and all regions of the brain which have not yet been acquired by way of a logical combination of the changed first image region, the second image region and the fifth image region.
28. A computer readable medium including a computer program product, the computer program product comprising instructions, which when executed on a computer device, causes the computer device to perform functions for determining a region of gray matter in a brain starting from a time-dependent computed tomography image data record from a perfusion CT examination, the functions including:
defining a work region in the time-dependent computed tomography image data record which only comprises healthy gray matter, healthy white matter and ischemic gray matter;
calculating a plurality of three-dimensional time-independent images from the time-dependent computed tomography image data record, the plurality of three-dimensional time-independent images including an average image with and without a contrast agent being present, a maximum value image over time and a baseline image without a contrast agent being present;
determining a normalized difference image from the image data of a difference image, the difference image being determined from the average image and the maximum value image;
generating a high-contrast image with increased contrast between white matter and gray matter from the average image, the maximum value image and the normalized difference image;
determining a first grayscale value range, which only comprises gray matter, and a second grayscale value range, which only comprises white matter, by threshold histogram analysis of the high-contrast image, wherein only the image data of an undamaged hemisphere of the brain is considered;
determining a third grayscale value range which comprises both the healthy white matter and the ischemic gray matter by threshold histogram analysis of the normalized difference image;
determining a fourth grayscale value range which comprises both the healthy gray matter and the ischemic gray matter by threshold histogram analysis of the baseline image;
generating a fifth grayscale value range which lies in the damaged hemisphere and only comprises the ischemic gray matter by a logical combination of the third grayscale value range and the fourth grayscale value range;
determining seed points for respectively the first grayscale value range, the second grayscale value range and the fifth grayscale value range,
performing a region growing method respectively starting at the previously determined seed points and generating a first image region, a second image region, and a fifth image region;
generating an expanded first image region by adding cerebral regions which constitute anatomically assured gray matter to the first image region; and
determining a resultant image region which comprises both healthy gray matter and ischemic gray matter and all regions of the brain which have not yet been acquired by way of a logical combination of the changed first image region, the second image region and the fifth image region.