I want to know how can I find Field of View of each patient in dataset, because I need it to compute the voxel volume
Typically, the voxel dimensions are specified as metadata together with the actual image data. If you have DICOM data, then the voxel size can be obtained from the fields "Pixel Spacing"(0028,0030) and either "Spacing between Slices" (0018,0088) or "Slice Thickness" (0018,0050) or the differences of the "Slice Location" (0020,1041). (If the slices are acquired with a gap in between, you may or may not add the gap to the slice thickness depending on your application).
Other frequently used data formats such as the Analyze format, the NIfTI format, NRRD format etc. also provide the voxel dimensions as metadata. (So, probably you need not look for the FOV, but directly for the voxel dimensions.)
It can be reversely calculated or Check DICOM header. However, scanner settings do not always match up with values stored in the DICOM header. For example, FoV scanner may be set to 240mmx240mm and DICOM header metadata might report FoV DICOM=1440x1440. Often due to storage rounding errors.
Many thanks Olaf and Hugues for your comments but my question if I segment tumour how can I determine the tumour volume, what value that I should multiply each pixel in the segmented image?
If you have segmented a tumor, you can multiply the number of segmented voxels by the voxel volume dx × dy × dz, which can be obtained, e. g., from the DICOM tags "Pixel Spacing"(0028,0030) (for dx and dy) and "Spacing between Slices" (0018,0088) (for dz).
You have not told us what kind of image dataset you start with in the beginning (i. e., which file format/image format). You will need some kind of metadata associated with your dataset to find out the voxel dimensions. If you should have pure image data without these metadata, it will not be easily possible to find out the original field of view and voxel dimensions.
Dear Olaf
Thank you for your comment, I have real dataset in dicom format and I have all associated information with each patient. I implemented my segmentation algorithm on BMP images after converting dicom to BMP. I want to know at first what is the relationship between pixels and voxels in order to be able to compute volume. I think it is not possible to multiply immediately the segmented pixels by pixel spacing and depth
regards
Ali
I implemented the segmentation on BMP in order to reduce the complexity as well as the BMP image have the same resolution of Voxel, is it possible to suppose that the ratio of pixel and voxel is equivalent
- Badges
- Science topic