Question about concatenations

[John Stevenson-Hoare]

I have recently needed to create a concatenation due to the number of
frames in my multi-frame image. I am somewhat confused by the
definitions in PS 3.3.

Table C.7.6.16-1 states for the Per-frame Functional Groups Sequence
that "The number of items shall be the same as the number of frames in
the Multi-frame image".

Section C.7.6.16.2.2.4 (Concatenations and Stacks) states that the
Number of Frames attribute shall be present in each SOP Instance of
the concatenation but its value may be different for each.

So my question is, should the number of items in each SOP Instance be
the same as the Number Of Frames in each SOP Instance or the total
Number of Frames from all SOP Instances?

Thanks.

[Kees Verduin]

On 29 mei, 17:06, John Stevenson-Hoare <john.stevenson-

Hi Steve,
Where in the beginning authors assumed that all concatenation parts
should be re-united before a multi-frame image could be used, the
latest discussions in which I was partitioning cleared that vendors
were planning to create and store concatenation parts in their
database, without ever combining them.
Others were of different opinion.

This difference of opinion has lead to a situation in which
concatenations as such could not be allowed for IHE profiles, until
this is resolved.

The current text of the DICOM standard explicitely allows Number of
Frames (0028,0008) to be different per concatenation part, which makes
each part a valid and complete Multi-Frame Image.

It should be clear that when splitting an image into concatenation
parts, that the Per-Frame Functional Groups Sequence only describes
those frames that are present in that part. Therefore there are as
many items as frames in that part.


This also means that when combining the concatenation parts into one
Multi-Frame Image, the values of the individual Number of Frames
attributes must be added-up, to get the correct value, and that all
the Per Frame Functional Groups Sequences must be joined into a new
one before you get the resulting MF image.

Kees Verduin
ke...@kaveco.nl

[John Stevenson-Hoare]

Hi Kees,

Thanks for your reply. I will update my application to only include the items in the Per-frame Functional Groups Sequence that are relevant to the actual frames stored in the SOP Instance.

John