How to calculate Temporal Resolution without DICOM header?

1
votes

How can I calculate the Temporal Resolution of MRI sequence if it's not written in DICOM header?

What other DICOM information can I use to calculate out the temporal resolution?

It might worth mentioning that it's a 3D sequence. I have 20 slices and each slice 99 time points.

Dump

1
image-processingdicom
Short answer: not at all. Long answer: If you could post a DICOM dump of the MRI dataset (one slice is sufficient), we could have a closer look at it. – kritzel_sw
Your did not post the whole dump. If the slices are not temporarily related, the time attributes are not mandatory. So if you do not find any attributes like Content Time (0008,0033) or Acquisition Time (0008, 0022), I do not see a chance to determine the temporal resolution. Unfortunately your screenshots do not include the group 0008 elements so I cannot tell from the information provided. – kritzel_sw
I attached the information you asked for. Is temporal resolution = acquisition time / number of slices – user161260

1 Answers

1
votes

First: The encoding of the time values looks more than weird. It does not conform to DICOM encoding rules. See: PS3.5, Table Table 6.2-1. DICOM Value Representations This may either be a wrong encoding in the DICOM file or by an issue in the dump tool you are using. The values should read:

HH.MM.SS.FFFFFF where HH is hours, MM is minutes, SS is seconds and FFFFFF the microsecond fraction. Parts of it may be omitted, so HH.MM would also be valid but a value of "69481.000" does not match this pattern at all.

In general, the Acquisition Time (0008, 0032) is the attribute I would recommend to consider since it is "The time the acquisition of data that resulted in this image started".

How the "temporal resolution" should be calculated very much depends on what you would define the "temporal resolution" as. If you want to know the temporal distance between two adjacent slices, I would recommend to sort the images along their Slice Location (0020,1041) and calculate the difference between their Acquisition Times. If you want to average out rounding errors, it might be worth consideration to calculate the difference of Acquisition Time between the first and the last slice and divide it by the number of slices.