Medical image processing device, medical image processing method, program
Provided is a medical image processing device capable of notifying the diagnosis personnel that a segmentation error has occurred or may have occurred during tissue segmentation processing. This medical image processing device specifies a gray matter image of a subject, smoothes the gray matter image, and, in accordance with an elevation function for calculating an absolute Z score, calculates an elevation value. Next, the medical image processing device compares the evaluation value with a pre-defined threshold value and determines the segmentation result, and, if the separation result is determined to be abnormal, warns that segmentation result is abnormal and displays a segmentation result display screen showing the segmentation result.
1. A medical image processing device, comprising:
an input means for inputting a brain image of a test subject, the brain image being made up of voxels, each of the voxels having a respective value;
a segmentation means for segmenting gray matter tissue by performing a tissue segmentation process on the brain image of the test subject and creating a resultant tissue segmented image;
a memorizing means for saving an image group of gray matter tissues of healthy subjects;
a memorizing means for saving normal tissue segmentation images;
an output means for outputting diagnosis support information based on statistical comparison between the resultant tissue segmentation image of the test subject obtained by the segmentation means and the image group of gray matter tissues of healthy subjects; and
a distinguishing means for distinguishing between normal and abnormal results or segmentation by the segmentation means, based on the voxel value of the resultant tissue segmented image for the brain image of the test subject and per-voxel statistical value of the normal tissue segmentation images, wherein
a per-voxel statistical value of the normal tissue segmentation images is represented, at every position, by an average value and a standard deviation of respective distribution of the voxel values according to a position to which each image of the normal tissue segmentation images corresponds, and then
the distinguishing means distinguishes between normal and abnormal results of segmentation by the segmentation means by comparing, against a predefined threshold value, an average value over the whole of the resultant tissue segmented images of an absolute Z-score determined from the voxel value of the resultant tissue segmented image for the brain image of the test subject, and the average value and the standard deviation at a position to which the normal tissue segmentation images corresponds.
2. The medical image processing device of claim 1 , further comprising:
a memorizing means for memorizing the average value and the standard deviation at every voxel to which each image corresponds substituting for the memorizing means for memorizing the normal tissue segmentation images, and
a reflection means for reflecting the voxel value of the gray matter tissue for the brain image of the test subject, which is distinguished as normal by the distinguishing means, to the memorizing means.
3. The medical image processing device of claim 1 , further comprising:
a warning means for outputting a warning information when the distinguishing means recognizes the result of the segmentation as an abnormal result.
4. The medical image processing device of claim 1 , further comprising:
a display means for displaying the segmentation result by the segmentation means when the distinguishing means recognizes the result of the segmentation as an abnormal result.
5. A medical image processing method, comprising:
an input step of inputting brain image of a test subject;
a segmentation step of segmenting gray matter tissue by performing a tissue segmentation process on the brain image of the test subject;
an output step of outputting diagnosis support information based on statistical comparison between the gray matter tissue image of the test subject obtained by the segmentation step and the image group of gray matter tissues of healthy subjects; and
a distinguishing step of distinguishing between normal and abnormal result of segmentation in the segmentation step, based on the voxel value of the gray matter tissue for the brain image of the test subject and the per-voxel statistical value of the gray matter tissue for the brain image group, for which the tissue segmentation process has been performed normally, wherein
the distinguishing step distinguishes between normal and abnormal result of segmentation by the segmentation means by comparing the absolute Z-score of the voxel value of the gray matter tissue for the brain image of the test subject against a predefined threshold value.
6. A non-transitory computer-readable medium containing a program for causing a processing device to perform:
inputting a brain image of a test subject, the brain image being made up of voxels, each of the voxels having a respective value;
segmenting gray matter tissue by performing a tissue segmentation process on the brain image of the test subject and creating a resultant tissue segmented image;
saving an image group of gray matter tissues of healthy subjects;
saving normal tissue segmentation images:
outputting diagnosis support information based on statistical comparison between the resultant tissue segmentation image of the test subject obtained by the segmentation means and the image group of gray matter tissues of healthy subjects; and
distinguishing between normal and abnormal results of segmentation by the segmentation means, based on the voxel value of the resultant tissue segmented image for the brain image of the test subject and the per-voxel statistical value of the normal tissue segmentation images, wherein
a per-voxel statistical value of the normal tissue segmentation images is represented, at every position, by an average value and a standard deviation of respective distribution of the voxel values according to a position to which each image of the normal tissue segmentation images corresponds, and then
distinguishing between normal and abnormal results of segmentation by comparing, against a predefined threshold value, an average value over the whole of the resultant tissue segmented images of an absolute Z-score determined from the voxel value of the resultant tissue segmented image for the brain image of the test subject, and the average value and the standard deviation at a position to which the normal tissue segmentation images corresponds.
7. The medical image processing device of claim 2 , further comprising:
a warning means for outputting a warning information when the distinguishing means recognizes the result of the segmentation as an abnormal result.
8. The medical image processing device of claim 2 , further comprising:
a display means for displaying the segmentation result by the segmentation means when the distinguishing means recognizes the result of the segmentation as an abnormal result.
9. The medical image processing device of claim 3 , further comprising:
a display means for displaying the segmentation result by the segmentation means when the distinguishing means recognizes the result of the segmentation as an abnormal result.
10. The medical image processing device of claim 7 , further comprising:
a display means for displaying the segmentation result by the segmentation means when the distinguishing means recognizes the result of the segmentation as an abnormal result.