HELP : sequence data element in DICOM
budhik
specification and its encoding scheme.
Is there any one can describe a sample of the implementation of the
"SQ" data element and its encoding scheme ?
Best regards,
Budhi
Thomas Freier
Again as David told you, use the version with undefined length. it is
not so difficult.
It is an example only make sure you understand the concept what needs
to implement.
You could implement to add the Item Tag and Delimiters kind of fixed.
Assuming Transfer Syntax is Explicit Little Endian, let's say you want
to encode a Tag 0054, 0012 SQ
Start the sequence with:
54 00 12 00 53 51 00 00 FF FF FF FF
FE FF 00 E0 FF FF FF FF < 1st
Item
xxxxxxxxxxxxxxxxxx < Tags
of your first item (see posting with UID)
FE FF oD E0 00 00 00 00 < Item
Delimiter
FE FF 00 E0 FF FF FF FF < 2nd
Item
xxxxxxxxxxxxxxxxxx < Tags
of your 2nd item (see posting with UID)
FE FF 0D E0 00 00 00 00 < Item
Delimiter
FE FF DD E0 00 00 00 00 <
Sequence Delimiter
Note there are no VR's in the Delimiter "Tag".
Thomas
Thomas Freier
DICOM file and shows both DICOM and HEX interpretation at the same
time.
I am not the author of that tool. Might help you if you need more
examples.
budhik
Hi Thomas,
Yes, I'm trying to follow David's recommendation.
Let's say, I have a sequence A that contains sequence B and sequence
C. Sequence B contains 3 items, and sequence C contains 2 items.
Let's say, the structure is :
Sequence A (0054, 0012 SQ)
> Sequence B (0055, 0015 SQ)
>> Item B1 (0011, 0001 - UI - "Blood flow")
>> Item B2 (0011, 0002 - DS - 110)
>> Item B3 (0011, 0003 - AE - "Brad Pit")
> Sequence C (0056, 0020 SQ)
>> Item C1 (0022, 0001 - UI - "Blood pressure")
>> Item C2 (0022, 0002 - DS - 200)
What is the correct syntax in explicit little endian ?
Regards,
Budhi
Thomas Freier
>
> - Show quoted text -
Budhi,
honestly it's "hard work" to encode this things manually kind of
hopless. You can't do it by examples only, but rather have to program
the rules into your software.
Having said that if you do need more samples:
Get the DICOM Hex tool and have a look at some examples, you will
instantly see.
Or you follow the example I posted before and simply replace the
xxxxxxxx with another sequence starting with the Tag
55 00 15 00 53 51 00 FF FF FF FF .... to get a nested structure. It
is very straightforward.
Thomas
David Clunie
budhik wrote:
> Let's say, I have a sequence A that contains sequence B and sequence
> C. Sequence B contains 3 items, and sequence C contains 2 items.
> Let's say, the structure is :
>
> Sequence A (0054, 0012 SQ)
>> Sequence B (0055, 0015 SQ)
>>> Item B1 (0011, 0001 - UI - "Blood flow")
>>> Item B2 (0011, 0002 - DS - 110)
>>> Item B3 (0011, 0003 - AE - "Brad Pit")
>> Sequence C (0056, 0020 SQ)
>>> Item C1 (0022, 0001 - UI - "Blood pressure")
>>> Item C2 (0022, 0002 - DS - 200)
This is wrong in many ways:
- a Sequence Item is a complete data set, not a single data element,
and it begins with a Sequence Item tag (FFFE,E000), and in the
undefined length form (recommended) ends with the Item Delimitation
tag (FFFE,E00D)
- within the Sequence Item is the data set, i.e., another numerically
sorted list of data elements, which in turn may contain sequence
as well
- your incorrect example items that are data elements are also wrong,
even if they were valid items, in that they are private data elements
(odd group) but are not of the correct data element range - data
element ranges 00xx encode the private creator for the block of
private elements xx00
- things with a VR of UI are unique identifiers (1.2.nnn...), not plain
text strings like "blood flow"
- things with a VR of AE are application entity titles, not plain
text strings like "Brad Pit"
- if you are going to send things like "blood flow" and "blood pressure"
do not do this as private elements - these should be in annotations or
acquisition context in the waveform object or in separate SR objects
David
budhik
Hmm.... ok....
Btw, for the VR of "OW", should I treat them with the same way as
"SQ" ?
I mean, should I use same nested tags as sequence ?
I have an "OW" for 5400,100a and 5400,1010 that contains original
signals of 12-lead ECG.
Please advise,
Budhi
PacificPrashant
>
> - Show quoted text -
Hi Budhi,
I think you can find information about this in Section 8.3 WAVEFORM
DATA AND RELATED DATA ELEMENTS of DICOM Standard PDF No. 5. "Data
Structures & Encoding" (Experts, if I am wrong please correct! )
If I come across any other information will let you know.
All the Best!
Thanks.
Prashant
budhik
I already see it, actually.
Well... let's say I have 12-lead ECG signals (the numbers are in
decimal/float) :
2.471310 0.000000 0.028690 3.764345 3.721310 0.014345 0.027228
0.000183 0.038740 0.042029 0.000000 0.000000
2.390177 0.020284 0.130107 3.794770 3.630035 0.075195 0.123650
0.000647 0.176115 0.191053 0.015533 0.009868
2.300134 0.097415 0.297280 3.801225 3.501427 0.197348 0.282920
0.010754 0.403556 0.437369 0.077483 0.055046
2.243407 0.237025 0.493618 3.759784 3.374894 0.365321 0.470268
0.052002 0.671834 0.727533 0.196039 0.154307
2.219190 0.416453 0.697264 3.682178 3.260963 0.556858 0.664729
0.137032 0.950845 1.028743 0.356180 0.303389
Column 1 : ECG Lead I
Column 2 : ECG Lead II
Column 3 : ECG Lead III
Column 4 : ECG Lead aVR
Column 5 : ECG Lead aVL
Column 6 : ECG Lead aVF
Column 7 : ECG Lead 1
Column 8 : ECG Lead 2
Column 9 : ECG Lead 3
Column 10 : ECG Lead 4
Column 11 : ECG Lead 5
Column 12 : ECG Lead 6
For (5400,100a) I will use 1 row only, the VR is "OW".
For (5400,1010) I will use all rows, the VR is "OW" also. I can do it
by using looping of course.
If I'm right, each column will become items in (5400,100a) either
(5400,1010).
So, the structure of (5400,100a) should be :
00 54 0a 10 4F 57 00 00 FF FF FF FF
FE FF 00 E0 FF FF FF FF < 1stItem
column1
< column1
FE FF oD E0 00 00 00 00 < Item
Delimiter
FE FF 00 E0 FF FF FF FF < 2nd
Item
column2
< column2
FE FF 0D E0 00 00 00 00 < Item
Delimiter
........
nnnnnnnnnn <
continued until column12
........
FE FF DD E0 00 00 00 00 <
Sequence Delimiter
And for (5400,1010) will be looped until all signals placed as items.
The question is :
Is it true, it should be like that, or my algorithm is wrong ? What it
should be ?
Please advise,
Budhi