I am working on a project to classify lung CT images using CNN, and
as I understand, the pixel spacing is distance between the center of each pixel, specified by a numeric pair - adjacent row spacing and adjacent column spacing in mm (this is in 2D space), but because the CT images are 3D, the pixel spacing shows a triple value, and I thought that the second and third value are row pixel spacing and column pixel spacing, but the first value is Z dimension which means slice thickness, I am right or not?
the another question is: as it's known, CT scans has a different slice thickness, if I work on 2D slice of CT, I need to make all slice the same thickness or its only necessary when working on 3D scans?
In general CT images are in the DICOM format. Slice thickness can also be said to equal the collimator width. Slice thickness and slice spacing are in general not the same in MDCT. Slice spacing is the corresponding value to pixel spacing (in x,y plane) in z direction (scan direction) and it means the spacing between adjacent slices. It can be calculated by the difference in the slice locations of two adjacent slices. There is a DICOM tag called SliceLocation that can be used but it is saver to calculate SliceLocation by your own using the ImagePositionPatient tag. Spacing in x,y can be found in the DICOM tag PixelSpacing. In general pixel spacing and slice spacing is fix for one CT scan.
- Badges
- Science topic
- Similar topics
- Analytical Chemistry
- Column