Oviyam2.1 + Tomcat 7 - dark images - Help!

[Docjay]

Suresh,

   I'm hoping you will respond, or anyone else who might have experienced this issue.

My current setup:

dcm4chee 2.17.2
MSSQL 2008 R2
Server 2008 R2 (x64)

   I am using Chrome to view Oviyam 2.1.

   I have the latest Oviyam 2.1 installed on my dcm4chee & JBOSS 4.2.3 setup and everything is working except images load and are very dark.  I have to hit the 'window/level' button on the top right of the tool bar to view the image or else it is windowed almost 'black'.  I also have an install of Oviyam 2.1 + Tomcat 7 and it still behaves the same way.  The thumbnails (preview images) are dark as well as the main view port.  I have both installs setup to pull images from DCM4CHEE and studies are stored JPEG2K Lossless (1.2.840.10008.1.2.4.90). 

   If I use weasis, the images are displayed correctly.  The tomcat logs show no error messages at all.  I have tried Oviyam 2.0 and I get the same results.

If there is anything I can provide to help correct this issue, please let me know.  Oviyam 2 can be a great answer to a problem we have around here with outside CDs coming in from other hospitals!

p.s. I have attached a copy of my server.log file to help show what is going on.  It looks like its only accepting '1.2.840.10008.1.2/Implicit VR Little Endian'.  Is there a way to force it to accept JP2K (1.2.840.10008.1.2.4.90) so that it won't have to decompress?

thanks

Jamie

EDIT:

Just to clarify some things:   CR images are always dark.  US images always appear to load normally (must already be compressed I think).  SC or screen save images load normally as well.  CT & MR images always load dark or black.

[Devishree.V]

 Hi,
   Can you attach a sample DICOM file to troubleshoot?

[Docjay]

Here is a link to a sample DICOM image for you.  It was too big to upload for this example. This is JP2K lossless compressed.

https://drive.google.com/file/d/0B-bV90lIXJRwNGFSbDZqOXpRdWc/edit?usp=sharing

thanks for your help

[Docjay]

Devishree,

    Just wondering if you had any luck testing out the sample DICOM I supplied?

thanks

[Chris Hafey]

I just tried out your sample image and it displays fine for me (although it does invert when I try to WW/WL it).  Are you sure you are using the 32 bit JDK?  I ask because 64 bit JDK on Windows does not support JPEG2000 right now so that might be causing the black images.

Chris

[Jamie Henning]

I am using  64bit jdk,  but I'm using a codec (jar) from  aware.Com that works with it.   Is there a way for me to use that codec with oviyam?

--
You received this message because you are subscribed to a topic in the Google Groups "dcm4che" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/dcm4che/TbD0vyxzcQ0/unsubscribe.
To unsubscribe from this group and all its topics, send an email to dcm4che+u...@googlegroups.com.
To post to this group, send email to dcm...@googlegroups.com.
Visit this group at http://groups.google.com/group/dcm4che.
For more options, visit https://groups.google.com/d/optout.

[Jamie Henning]

I'm aware that  the x64 jdk doesn't play nice with compression on x64 windows.

[Jamie Henning]

... But then I read that Oviyam only does jpk lossy, so then dcm4chee would have to decompress the image 1st.  My log shows that it's doing the decompression. Wonder why it still won't display correctly?

[Andrew Crosby]

I experience the same darkening problem on Oviyam 2.0 and 2.1, but I am using 32-bit JDK.

When I click the Windowing icon:
DX always turns black but has an outline
US always turn inverted and shows a double image (see screenshot for both of these)
CT and MR work ok

The CR from Docjay worked ok on my setup, oddly enough. But I did see this warning in the log, just for Docjays's study:
2014-08-28 22:39:09,555 INFO  -> (http-0.0.0.0-8080-99) [org.dcm4chex.wado.mbean.WADOSupport] Get WADO object for 1.2.840.113564.10.1.2493372104589062020615111312110117210619817
2014-08-28 22:39:09,570 WARN  -> (http-0.0.0.0-8080-99) [org.dcm4chex.wado.mbean.WADOSupport] WADO parameter transferSyntax should neither Implicit VR, nor Big Endian! Use Explicit VR little endian instead! transferSyntax:1.2.840.10008.1.2

I see Docjay and I are both using Windows Server, is it recommended to use Linux or another database to be more consistent in displaying?

my setup:
dcm4chee 2.17.1
Java 6 update33 x86
Postgres 9.3
Server 2012 (x64)

[Suresh Viswanathan]

Our development environment is Linux so it tends to get tested more on that compared Windows or Mac. It would help if you could check and see if you can replicate the issue under Linux. We will also test it out under Windows.

Suresh.

You received this message because you are subscribed to the Google Groups "dcm4che" group.
To unsubscribe from this group and stop receiving emails from it, send an email to dcm4che+u...@googlegroups.com.

To post to this group, send email to dcm...@googlegroups.com.
Visit this group at http://groups.google.com/group/dcm4che.
For more options, visit https://groups.google.com/d/optout.

<DX-and-US-dark.png>

[Chris Hafey]

I tracked down the problem - it is a bug in Oviyam.  The file windowing.js line 383 and 387 need a parseFloat() when getting the windowWidth and windowCenter from the DICOM file.  Oviyam also doesn't handle PhotometricInterpretation for MONOCHROME1 (which this image has) so the image shows up inverted.  I'll see if I can send in a patch for these issues today.

Chris

[Docjay]

Wow, Excellent work.  I'll test it out for sure to help squash this bug.

[Andrew Crosby]

That is great. I noticed my images are MONOCHROME2, will this patch support that?

Example dicom: https://drive.google.com/file/d/0B5Yi1Ti2edApSGpWMDBwSmNHTXc/edit?usp=sharing

On Aug 29, 2014 9:17 AM, "Docjay" <jlhe...@gmail.com> wrote:

Wow, Excellent work

On Friday, August 29, 2014 7:33:43 AM UTC-5, Chris Hafey wrote:

[Chris Hafey]

Yes, MONOCHROME2 doesn't require inverting so it is already supported by Oviyam and my patch didn't break that.  You can try out the fix here:

https://drive.google.com/file/d/0BzovUgmP3BOKZExGZkxPSklfMUU/edit?usp=sharing

[Docjay]

FYI - already spoke to Chris about this, but it didn't correct the issue I'm experiencing.  I really do appreciate his efforts though, so thanks Chris!

[Devishree.V]

 Hi Docjay,
    Are you able to view the images correctly using DCM4CHEE web interface?

[Devishree.V]

   Hi Andrew,
       While applying WW/WL in this sample DICOM image, DCM4CHEE web interface produces the same image for any set of windowing parameters. Kindly check the following sample WADO URL with windowing parameters,

  /wado?requestType=WADO&studyUID=1.3.6.1.4.1.19179.1.110452023442174.1.8152.5528&seriesUID=1.3.6.1.4.1.19179.1.110452023442174.2.9393.5529&objectUID=1.3.6.1.4.1.19179.1.110452023442174.3.9393.5530&rows=1024&windowCenter=8257&windowWidth=16460

  Most of the pixels are zero and we are facing unexpected results on windowing, when viewing in some other DICOM viewers also.

 Thanks&Regards,
  Devishree.V

[Jamie Henning]

Yes, they view correctly using dcm4chee Web interface.

[Andrew Crosby]

Devishree,

I checked the WADO URL and it displayed correctly, no darkening. I could adjust the 2 windowing parameters and each image was different and displayed correctly. Is that what you expected?

Thanks,
Andrew

[Andrew Crosby]

Chris - I tried your oviyam2.war, I had to upgrade from Java 6u33 to 7u60 to deploy (it did not resolve the issue as Docjay stated). But when I went back to the 2.0 war from sourceforge, the thumbnails in the left pane are full size... I cleared my chrome cache, wadoservice cache, tried 2.1 from sourceforge, reverted back to Java 6u33 but no luck - are there any other property files I need to clear when switching war files?

[Devishree.V]

 Hi Andrew,
      Processing 8,12,16 bit variations caused the darkening problem. The issue has been almost solved except for 8 bit images.
      Once its fixed for all type of images, a commit will be made in sourceforge along with JIRA entry.

[Docjay]

Devishree,

Will this solve my issue as well?  I provided sample DICOM above.

thanks

[Devishree.V]

 Docjay,
    The sample DICOM you attached works fine for us. There is no darkening issues.
    So please try some uncompressed studies to know whether there is any problem in compression. Although screenshots may help to troubleshoot further.

On Tuesday, September 2, 2014 6:54:20 PM UTC+5:30, Docjay wrote:

Devishree,

Will this solve my issue as well?  I provided sample DICOM above.

thanks

On Monday, September 1, 2014 8:51:48 AM UTC-5, Devishree.V wrote:

[Docjay]

Devishree,

  I have some questions for you.

1)  Did you view the sample while running Oviyam on a tomcat install? 
2)  Was the sample DICOM loaded into a version of DCM4CHEE?  if so, what version was used?
3)  What version of JAVA are you running if the study was loaded into DCM4CHEE?  v6 or v7?

thanks

[Devishree.V]

 Docjay,
     My replies are inline.

    1)  Did you view the sample while running Oviyam on a tomcat install? 

        Yes, i checked the sample in apache-tomcat-7.0.54

    2)  Was the sample DICOM loaded into a version of DCM4CHEE?  if so, what version was used?

         Yes, dcm4chee-2.17.3

    3)  What version of JAVA are you running if the study was loaded into DCM4CHEE?  v6 or v7?

         v6

[Docjay]

Devishree,

    One more question.  Was Java 32bit or 64bit?

thanks

[Devishree.V]

  Docjay,

       I've checked in both 32 and 64 bit.

[Andrew Crosby]

Docjay - In case this helps, I experience my issue on the following:
- DX Uncompressed, Implicit VR Little-endian
- US JPEG Lossless First Order
- US JPEG 2000 (Lossless) - same compression as your sample dicom image, but as I mentioned above I did not experience the issue on your sample dicom image
Since others cannot recreate, have you tried creating a new environment with the most basic dcm4chee installation (maybe with java 32bit) and see if it still occurs?

Devishree - when do you think the sourceforge commit may occur? Any benefit of getting a development war once 8 bit is resolved?

Thanks,
Andrew

[Devishree.V]

 
 Andrew,
     The issue was resolved and in testing level. We will make the commit ASAP.

[Andrew Crosby]

Thanks Devishree!

[Devishree.V]

 Andrew,
      We have committed the latest source code in sourceforge (revision 18333).

[Andrew Crosby]

This has resolved my issue, thank you for working this Devishree.

[rodrigo nascimento]

Hello, I installed Oviyam 2.1 on a Tomcat Server on my Mac, my intention is to work with the Oviyam in cases yesterday is necessary to use without the OsiriX.

The installation was a success and already can access the studies remotely, but can not work with the windowing of studies using the Wado protocol. Does anyone have this problem too?

Thanks.