Differences betwen DICOM files of the same series - DICOM

This is a discussion on Differences betwen DICOM files of the same series - DICOM ; Let me first thank you for taking the time to read this. I am developing alghoritm that would compress DICOM non-image data. Since most of the things are the same in DICOM files of the same series, I was thinking ...

+ Reply to Thread
Results 1 to 3 of 3

Thread: Differences betwen DICOM files of the same series

  1. Differences betwen DICOM files of the same series

    Let me first thank you for taking the time to read this. I am
    developing alghoritm that would compress DICOM non-image data. Since
    most of the things are the same in DICOM files of the same series, I
    was thinking of saving just the lines (or just the values) that are
    different (only the first file in the series would be saved fully). Let
    me explain further; below is part of the output for the first two files
    in the series.

    File1:
    0002,0000,File Meta Elements Group Len=188
    0002,0001,File Meta Info Version=1
    0002,0002,Media Storage SOP Class UID=1.2.840.10008.5.1.4.1.1.2.
    0002,0003,Media Storage SOP Inst
    UID=1.2.840.113674.2003080413213305132212912671.
    0002,0010,Transfer Syntax UID=1.2.840.10008.1.2.
    0002,0012,Implementation Class UID=1.2.804.114118.3
    ....

    File2:
    0002,0000,File Meta Elements Group Len=190
    0002,0001,File Meta Info Version=1
    0002,0002,Media Storage SOP Class UID=1.2.840.10008.5.1.4.1.1.2.
    0002,0003,Media Storage SOP Inst
    UID=1.2.840.113674.2003080413213405432212912671.
    0002,0010,Transfer Syntax UID=1.2.840.10008.1.2.
    0002,0012,Implementation Class UID=1.2.804.114118.3
    ....

    As we can see, only the first lines are partialy different (their
    values in particular) and there is a minor change in SOP Inst UID. If I
    would save just the value of the first line in second file, compression
    achieved would be substantial. This seems to be the case with every
    DICOM files I currently have, but I am concerned that there are DICOM
    series where files don't share this similarity betwen them.

    My question to you is: can I assume that for every same line in files,
    only their values can change, but their Item Tag stays the same?
    Meaning that if line 10 in File1 has Value Tag of 0002,0012, then line
    10 of File2 will also have the Value Tag of 0002,0012, but with
    different value.
    My second question, relating to the first, is: is it possible, that
    File2 has Item Tag/Value that is not present in File1?

    Thank you for all your anwsers,
    Dennis


  2. Re: Differences betwen DICOM files of the same series

    Why not just use the deflate transfer syntax ?

    For that matter, why bother, since the bulk data vastly
    overwhelms the size of the other attributes most times
    anyway ?

    David

    detra83@gmail.com wrote:
    > Let me first thank you for taking the time to read this. I am
    > developing alghoritm that would compress DICOM non-image data. Since
    > most of the things are the same in DICOM files of the same series, I
    > was thinking of saving just the lines (or just the values) that are
    > different (only the first file in the series would be saved fully). Let
    > me explain further; below is part of the output for the first two files
    > in the series.
    >
    > File1:
    > 0002,0000,File Meta Elements Group Len=188
    > 0002,0001,File Meta Info Version=1
    > 0002,0002,Media Storage SOP Class UID=1.2.840.10008.5.1.4.1.1.2.
    > 0002,0003,Media Storage SOP Inst
    > UID=1.2.840.113674.2003080413213305132212912671.
    > 0002,0010,Transfer Syntax UID=1.2.840.10008.1.2.
    > 0002,0012,Implementation Class UID=1.2.804.114118.3
    > ...
    >
    > File2:
    > 0002,0000,File Meta Elements Group Len=190
    > 0002,0001,File Meta Info Version=1
    > 0002,0002,Media Storage SOP Class UID=1.2.840.10008.5.1.4.1.1.2.
    > 0002,0003,Media Storage SOP Inst
    > UID=1.2.840.113674.2003080413213405432212912671.
    > 0002,0010,Transfer Syntax UID=1.2.840.10008.1.2.
    > 0002,0012,Implementation Class UID=1.2.804.114118.3
    > ...
    >
    > As we can see, only the first lines are partialy different (their
    > values in particular) and there is a minor change in SOP Inst UID. If I
    > would save just the value of the first line in second file, compression
    > achieved would be substantial. This seems to be the case with every
    > DICOM files I currently have, but I am concerned that there are DICOM
    > series where files don't share this similarity betwen them.
    >
    > My question to you is: can I assume that for every same line in files,
    > only their values can change, but their Item Tag stays the same?
    > Meaning that if line 10 in File1 has Value Tag of 0002,0012, then line
    > 10 of File2 will also have the Value Tag of 0002,0012, but with
    > different value.
    > My second question, relating to the first, is: is it possible, that
    > File2 has Item Tag/Value that is not present in File1?
    >
    > Thank you for all your anwsers,
    > Dennis
    >


  3. Re: Differences betwen DICOM files of the same series

    David, thank you for your reply (I was hoping you would offer your
    adivce). The reason for compression lies elsewhere.

    The person who I am working with has developed voxel compression
    alghoritm that can typical achive at least 40% compression rate. He
    takes all the slices from one series and produces 3D image, as well as
    above mentioned compression. My part in this is to compress all the
    other, non-image data, so we could reproduce/restore all the DICOM
    files in the same series if needed. Since we are developing this
    application for a hospital, where they have thousands of series and
    probably hundred times as more DICOM files, compressing non-image data
    would probably save a lot of space as well.

    I have already made and tested basic DICOM Viewer. My compression idea
    was to save just the first file in its original form. I would then save
    only differences for all the other files and hopefully, achieve good
    compression rate. My only concern is how different files are in the
    same series. If they can be substantially different, compression would
    probably be very poor and, as you said, not worth the trouble. But if
    all the files in the series are pretty much the same, compression would
    be substantial, especially if we consider the amount of data they have
    in hospital.

    Again, thank you for your reply. I hope I have explained my problem
    clearly and the need for compression. I will take a look at deflate
    transfer syntax, as you have suggested, as I don't really know what it
    is or why is it used (except for some kind of transfer, as the name
    implies).

    Dennis


+ Reply to Thread