Data Directory Structure

    Is there a way to get ConQuest to structure it's data directory such as:


    • Patient Name

      • Exam Date

        • Name of Study (ex DTI_medium_iso_709)

          • DICOM files


    The reason for this being, we want to share the data directory via SAMBA or Windows File Sharing as READ ONLY so people can just copy the DICOM data as needed as well as be able to do QUERY/RETRIEVE from a PACS Workstation... rather than rely on only QUERY/RETRIEVE.


    Or this would be second best... just something that gives a human-readable structure:


    • Patient Name

      • Exam Date

        • DICOM Files from Name of Study (ex DTI_medium_iso_709.dcm)


    Also, if ConQuest won't do it, do you know something that will?


    Thanks!

    Hi,


    you can set FileNameSyntax to a string that will do just that:


    FileNameSyntax = %name\%studydate\%studyid\%sopuid.dcm


    Marcel


    From manual:


    Note: since 1.4.11, FileNameSyntax may also be string containing % that is treated as flexible
    filenamesyntax. e.g., %id\%studyid\%seriesid\%sopuid.dcm.
    This string may contain: %name=(0010,0010), %id=(0010,0020), %modality=(0008,0060),
    %studyid=(0020,0010), %studyuid=(0020,000D), %seriesid=(0020,0011), %series=(0020,0011) with 4
    digits, %seriesuid=(0020,000E), %sopuid=(0008,0018), %imagenum=(0020,0013),
    %image=(0020,0013) as 6 digit integer, %imageid=(0054,0400), %time, %counter = (4 digit hex),
    %calledae, %callingae, %studydate. Any of these items can be followed by e.g., [0,3] which is a
    substring operator, e.g., %studydate[0,3] gives the year, %studydate[4,5] gives the month. Any other
    text is treaded literally – be careful to use only characters allowed in filenames plus the correct path
    separator: \ for Windows, and / for Linux.


    Thank you very much!!! What do the 0010,0020 numbers mean after the %values in the manual?

    That would be great.


    I'm currently trying to find the best combination of tags to sort with. I'm finding that the wrong combination can result in overwritten files, so that's not good.


    I'm currently using:

    Code
    %name\%studydate\%series\%modality%sopuid[46,56].dcm


    I don't like the long file name %sopuid.dcm produces. I'm trying to make sure this is compliant for both regular and enhanced DICOM, but I am not a DICOM expert (yet). :) Any input?

    Hi,


    in dicom, short filenames will not work, whil ethe entire SOP is guaranteed unique, any substring is not guaranteed so (the variable part is typically at the end, but does not need to be).


    What could work is to use the %counter option, possible combined with some series information.


    Marcel

    Hello,


    I am having problems with changing the DICOM save structure. No matter what I change the string to, it saves in the default format. I have tried removing and reinstalling the service, stopping and starting the server, and quiting and reloading the gui via the command prompt (all after changing the file structure string).


    I'm trying to set it to the string you suggested to give a pt name, date,etc. directory structure. I tried pasting it into the gui and changing the ini file directly, but neither seem to do anything. Folders I drag into the gui still just get put in a folder with the Pt ID.


    Any help is appreciated.

    Hi,


    Stop the server.


    Edit dicom.ini, as such, replacing the FileNameSyntax = 3 line with, e.g.,


    FileNameSyntax = %name\%studydate\%series\%modality%sopuid[46,56].dcm


    Start the server.


    This should work!. Which version are you using?


    Marcel

    Thank you for the response. I think I found my problem, but don't completely understand why it was the problem.


    Doing what you suggest did not work, but in doing it I noticed I was getting the "Failed to Listen () - bind error".


    So I changed the port number to 5679, restarted, and Voila, it started using the correct data structure.


    Does this make sense? It would have thought using the wrong port would have only caused network problems. Would it affect how local files are handled as well?


    Anyway, now for the network testing...so far this is the most flexible DICOM server I've come across. I love how you can specify how the images are saved. Very nice work.

    Hi,


    this means there eas a dgate.exe still running not controlled by the GUI app, so you were unable to restart is as the port was blocked. Kill the left-over dgate.exe with the task manager. The 5678 port is then free.


    Marcel

    FWIW, after consulting with a Philips DICOM expert, I wound up using this:


    Code
    %name\%studydate\%V0008,103e_%seriesuid[33,51]\%modality%series%counter.dcm


    According to him, this will guarantee a unique structure. The %V0008,103e DICOM attribute is very useful because it prints the Series Description.

    I've been working on the best way to store my cardiac MRI research images. My issue was wanting a way to store cines to allow easy conversion into correctly sorted bmp files, which could then be converted into avi or animated gif's for presentations. The issue is that if you use the default sopuid for the file name, it will not always get sorted correctly such that the resulting avi or gif looks scrambled.


    At first I was using %name\%studydate\%series\%imagenum.dcm forFileNameSyntax, but for acquisitions where multiple cine slices are stored in one sequence (at least with the way Siemens stores them) results in only one slice getting stored (the others are overwritten) because each slice has the same imagenums.


    So my solution has been to use this:


    Code
    %name\%studydate\%series_%V0008,103e\%V0020,1041[0,5]_%image.dcm


    which tacks on the first 6 characters of the image location tag to the file name. I believe this always gives me a unique file name within the directory such that files don't get overwritten. Then it also makes it easy to find and select the desired sequence to do the conversion to bmp files using something like IrfanView. It's also then easy to strip off the first 6 characters (location tag) from the filename after conversion by using, for instance, a program called "Renamer NG".


    Just figured I would document what worked for me, since it was a bit of a trial and error process. I love the flexibility of the naming, but with that flexibility requires careful file naming to make sure they aren't overwritten. Also, if anyone has any better way to do this...

Participate now!

Don’t have an account yet? Register yourself now and be a part of our community!

Register Yourself Login