Posts by marcelvanherk

    Here is a list of resolved issues between 1.4.12 and 1.4.12b.


    20070105. Archiving options: dgate manual incorrect, dgate --restoremagflags: does not work.


    20070107. Web access: when using the built-in dbaseIII driver, the CGI interface does not work correctly. Even though a series list is asked for a particular patient, all series of all patients are shown. This is a side effect of the critical bug fix in the dbaseIII driver.


    20070108. Timestamps missing in PacsUser and PacsTrouble under Linux (sorry Mark Pearson, I forgot).


    20070113. It should not be possible to store in files with .dcm file extension (set through FileNameSyntax) images with NKI compression.


    20070114. De- or re-compression of jpeg images sometimes fails in dcmc/djpeg because invalid transfer syntax passed.


    20070122. The free space in MIRROR devices (MIRRORDevice0) is not printed.


    20070124. PatientID with ' (quote) causes big trouble (loading images, checking consistency, forward filtering, deleting).


    20070126. PatientID with " " (multiple spaces) causes trouble - fails to store correctly in dbf database: duplicate entries. Also fixed in conquest_service_patches20070127.zip for older versions


    Marcel

    1. Importconverters that can changes VR's in incoming images were added


    2. Requested extensions to flexible filename syntax (%year, %month, %day, %centernumber) were implement as e.g., %studydate[0,3] ([] is a substring operator)


    3. Added multiple forwards in a rule. Implemented as multiple actions per rule separated by ;


    4. Redistributed mysql 5.0.22 client DLL as seperate zip


    5. Added preferred MAG device to store images. (14.12c) Use "storage MAG1" importconverter.

    What: Version 1.4.12c is partial update of the conquest dicom server, with scriptable importconverters (image modifier) as main new feature.


    Why: Release 1.4.12 fixed an important problem in the sql parser in the DbaseIII driver - but introduced a bug for database items containing sequences of spaces. These would not query correctly due to a simple coding mistake. Such items occur rarely but can cause a lot of junk in the database. The bug was found a few days ago, and fixed in release 1.4.12b. The fix is relevant only for users using the DBASEIII WITHOUT ODBC driver, i.e, the built in database driver. Another fix introduced a small bug disabling for all database drivers most of the GUI browser functions (such as delete series). This is fixed in 1.4.12c.


    How: For 1.4.12 and 1.4.12b users, just replace dgate.exe and conquestdicomserver.exe with the 1.4.12c counterparts. If you want minimal change, you can also download conquest_service_patches20070127.zip, and just replace the dgate.exe for your version: 1.4.10, 1.4.11, 1.4.12alpha, or 1.4.12. These would show up as sr2 versions


    Where: http://www.xs4all.nl/~ingenium/dicom.html

    Hi,


    All image are always stored unless identical UIDs are reused, e.g., for different patients - which would be a gross dicom violation. The *** warning message is to detect and log instances of double patient ID's you mentioned.


    Marcel

    From the manual:


    The following consistency checking is performed on the data entered in the database (without changing the image information that is stored):
    ? Inconsistent link information (e.g., two images of the same series belong to different patients), lead to a reject to store the new image with reported failure to the sending client.
    ? Filled items will not be overwritten by empty items.
    ? Known sex (M or F) in the patient database will not be overwritten with any other value than M or F.
    ? A known date of birth in the database will not be overwritten with an empty date or with a date on the 1st of January (which has a high probability to be wrong). When the original date of birth is empty, any value will be accepted.
    ? In case of any other inconsistency, the newer values will be written in the database, and the change will be logged as a warning.
    ? The (series) Modality field is appended to the Study Modality field in the database if it does not already contain this Modality.
    ? The PatientName, PatientBirthDate and PatientSex items are duplicated in the study table (database rev8 and up), to allow detection of patient ID mix-ups.


    Not from the manual: patient birthdate chaged gives a more serious warning. Efilm shows other names because it queries the study table not the patient table.


    Marcel

    Did you try the existing CGI interface? I guess some functions that are useful (for an administrator) are "send image/series/study to", "delete this series/etc", "merge", "split", i.e., all functions of the current windows GUI browser, but that are unavailable for linux users. Also maintaining acrnema.map would be useful. PhP on apache seems ok to me, although I do not know how to make the link to dgate -- commands (now used in the cgi app).


    Marcel

    Hi, Maurak,


    very useful data. However, the results indicate that you did not enable JPEG support (i.e., tell acuson that you support jpeg) in dgatesop.lst. Otherwise "as" should be small. It may also be that the acuson transfer syntax is never enabled. You can try to find what it is (the log of the server should list it) and enable it in dgatesop.lst. Normally, with enabled jpeg support. it should show these lines:


    #BigEndianExplicit 1.2.840.10008.1.2.2 transfer
    JPEGBaseLine1 1.2.840.10008.1.2.4.50 transfer LittleEndianExplicit
    JPEGExtended2and4 1.2.840.10008.1.2.4.51 transfer LittleEndianExplicit
    #JPEGExtended3and5 1.2.840.10008.1.2.4.52 transfer LittleEndianExplicit
    JPEGSpectralNH6and8 1.2.840.10008.1.2.4.53 transfer LittleEndianExplicit
    #JPEGSpectralNH7and9 1.2.840.10008.1.2.4.54 transfer LittleEndianExplicit
    JPEGFulllNH10and12 1.2.840.10008.1.2.4.55 transfer LittleEndianExplicit
    #JPEGFulllNH11and13 1.2.840.10008.1.2.4.56 transfer LittleEndianExplicit
    JPEGLosslessNH14 1.2.840.10008.1.2.4.57 transfer LittleEndianExplicit


    Marcel

    So, the problem is probably the v2 format (raw vr dump) not compression (dgate -nd detects that). Set filenamesyntax to a value (like 4) that uses .dcm extension (see manual). Newly stored images are then dicom compliant. Make sure incomingcompression is 'un'. To convert all images you have to send then to another (conquest) server.


    Marcel

    Hi,


    by disabling JPEG support you force Acuson to send images uncompressed. So decompression is probably the culprit of the problem. You could try enabling JPEG support again and setting UseBuiltInDecompressor in dicom.ini to 0: this will force use of the Offis libraries for decompression - see if they can handle these images (maybe also you need a fixed dgate1412). If this works, you can also try setting IncomingCompression to "as": this will leave incoming JPEG images as is and only decompress them when transferring to another server.


    Also, if someone can reproduce the problem by dragging and dropping a compressed acuson object that would be great - and/or get us a copy of the object (before send to conquest) and the broken one (in conquest).


    Marcel