Warning: Can't synchronize with repository "(default)" (/home/git/ome.git does not appear to be a Git repository.). Look in the Trac log for more information.
Notice: In order to edit this ticket you need to be either: a Product Owner, The owner or the reporter of the ticket, or, in case of a Task not yet assigned, a team_member"

Task #10097 (closed)

Opened 11 years ago

Closed 8 years ago

BUG: No ROIs shown

Reported by: pwalczysko Owned by: jburel
Priority: critical Milestone: ROI
Component: Insight Version: 4.4.8
Keywords: ROI Cc: ux@…, rleigh, mlinkert, jamoore
Resources: n.a. Referenced By: n.a.
References: n.a. Remaining Time: n.a.
Sprint: n.a.

Description

Log in as user-6 on Howe.
Log in as user-6 on Gretzky.
The same problem on both.
ROIs are not displayed in images with ROIs after import
Images were imported from
from squig: data_repo/ROI-samples/zeiss-zvi/... folder

The ROIs are there though and correctly read by Bio-formats (tested using the LOCI plugin in Fiji - image F1-annotations.ome.tiff - displays many regions in ROI manager. (see screenshot)

Attachments (2)

ROIs displayed in Fiji.png (653.4 KB) - added by pwalczysko 11 years ago.
No ROIs in Insight.png (713.9 KB) - added by pwalczysko 11 years ago.

Download all attachments as: .zip

Change History (33)

Changed 11 years ago by pwalczysko

Changed 11 years ago by pwalczysko

comment:1 Changed 11 years ago by pwalczysko

  • Cc rleigh added

comment:2 Changed 11 years ago by jburel

  • Sprint changed from 2013-01-15 (4) to 2013-02-12 (5)

Moved from sprint 2013-01-15 (4)

comment:3 Changed 11 years ago by pwalczysko

New folder was created by rleigh in data_repo/test_images_roi which has images in it with ROI which should be readable.

Build OMERO.insight-4.4.5-602-2ea0777-dirty-ice33-b124-mac
Identical results were delivered by the 4.4.5 release build of Insight downloaded from the webpage (OMERO.clients-4.4.5-ice33-b20.mac).

Images were imported using Insight UI.
ROI presence was checked in ImageJ using loci_tools.jar from 4.4. build on Jenkins (15 Jan build) and the result was compared with the display in Insight and Web.

Mias plate: (folder test_images_roi/mias)
IJ and Fiji show 173 regions in EACH of the 16 wells of this plate. Insight and Web both show 527 regions in the first well only, the other wells are without ROIs.
The regions shown by Image J do not seem totally plausible either, as they are the same for all 16 wells (anybody has an idea if this is the case ?)

test_images_roi/ome-tiff
No regions were displayed on any of these images in Insight and Web - regions were displayed in ImageJ in each of these images except "roi-omero.ome.tiff" and "roi-template.ome.tiff"

test_images_roi/zvi:
No regions displayed on any of the images nor in Insight neither in web - Regions present in each of these in Image J

Last edited 11 years ago by pwalczysko (previous) (diff)

comment:4 Changed 11 years ago by pwalczysko

  • Cc mlinkert-x jmoore added
  • Priority changed from major to critical

comment:5 Changed 11 years ago by pwalczysko

The same results as with today's build for Insight UI (see above) were achieved when using the 4.4.5 release version of Insight. Thus, this is not a bug introduced during the work on 4.4.6.

comment:6 Changed 11 years ago by jburel

  • Sprint changed from 2013-02-12 (5) to 2013-03-12 (6))

Moved from sprint 2013-02-12 (5)

comment:7 Changed 11 years ago by jburel

  • Owner changed from jburel to rleigh

Roger could you have a look at that one?

comment:8 Changed 11 years ago by rleigh

This is most likely due to the bug in my open PR for the ZVI reader: an exception gets thrown during ROI parsing and all ROIs in the layer are lost. I'll check with the current merge builds if this is now fixed.

comment:9 Changed 11 years ago by rleigh

Looking at current bioformats dev_4_4, with or without my zvi-roi branch merged in, showinf will generate OME-XML descriptions of all the ROI types which map onto our ROI model. For f1-measurements.zvi, this is 18 separate ROIs. While bioformats and ImageJ show these ROIs, they are not shown with OMERO. Since bioformats is reading and generating OME-XML containing the ROIs, I think the ROIs are being lost at import time.

comment:10 Changed 11 years ago by jburel

  • Sprint changed from 2013-03-12 (6)) to 2013-04-09 (7))

Moved from sprint 2013-03-12 (6))

comment:11 Changed 11 years ago by jburel

  • Sprint changed from 2013-04-09 (7)) to 2013-05-07 (8)

Moved from sprint 2013-04-09 (7))

comment:12 Changed 11 years ago by pwalczysko

  • Milestone changed from OMERO-4.4.8 to OMERO-4.4.7

During testing of @rleigh PR https://github.com/openmicroscopy/bioformats/pull/476 found out following.

The ROIs are not displayed on any of the images from test_images_roi/zvi/ folder after import to Insight AND are invisible in Web as well.
This is true for dev_4_4 insight and Web connected to Howe as well as develop insight and Web connected to Gretzky.

The ROIs from the same images are displayed when these images are opened using ImageJ or Fiji.
Although the PR has some problems on its own, which are at present dealt with by @rleigh, this shows that there is some separate, serious bug preventing viewing of the ROIs in the OMERO clients.

Last edited 11 years ago by pwalczysko (previous) (diff)

comment:13 Changed 11 years ago by pwalczysko

  • Milestone changed from OMERO-4.4.7 to OMERO-4.4.8
  • Sprint 2013-05-07 (8) deleted

comment:14 Changed 11 years ago by rleigh

  • Owner changed from rleigh to mtbcarrol
  • Version set to 4.4.8

This does not appear to be a Bio-Formats bug, so reassigning it.

With "showinf -nopix -omexml", ROIs are shown in the OME-XML for the images in question (e.g. /ome/data_repo/test_images_roi/zvi). Likewise, import into ImageJ using the Bio-Formats importer shows all the ROIs in the ROI manager (not visible unless you change the colour, since I used a white background, but there nonetheless). When importing into OMERO, no ROIs are present at all. So it looks like they are not being imported for some reason.

comment:15 Changed 11 years ago by mlinkert

  • Owner changed from mtbcarrol to mtbcarroll

comment:16 Changed 11 years ago by mtbcarroll

  • Sprint set to Blocker 4.4.9 (1)

comment:17 Changed 11 years ago by mtbcarroll

  • Status changed from new to accepted

So, my initial impression is that there is simply no code in OMERO to add the non-well image ROIs and that a fruitful first step may be to look at populate_roi.py to see how it adds the ROIs, and then to try to find some way in MetadataStoreI.postProcess_async to get hold of images' Bio-Formats readers and to use the stuff in PrintROIs.java to figure out how to write code to get the ROIs added.

Confusion includes if I should be using methods from OMEROMetadataStore and why the ROIs of the model in OME appear to be javax.media.jai.ROI whereas various other code works with omero.model.Roi.

If this isn't the approach to take with this ticket, speak up now. (-:

comment:18 Changed 11 years ago by wmoore

Importing this dir will give you an image with 16 ROIs:
ome/data_repo/from_skyking/fv1000/kashif/OIF_V1

comment:19 Changed 11 years ago by mtbcarroll

test_images_roi/zvi/roi-rectangle-aligned.zvi claims to have no ROIs but during import ROIs and Shapes do come through OMEROMetadataStore.updateObject and the ROIs end up in the roiList field; however, the roi database table gains no rows.

Last edited 11 years ago by mtbcarroll (previous) (diff)

comment:20 Changed 11 years ago by mtbcarroll

My plan for tomorrow is to check for differences between the two images in what happens with OMEROMetadataStore.handleReference.

comment:21 Changed 11 years ago by mtbcarroll

In ImportLibrary.importMetadata only for the wrongly ROI-less image, store's reference cache's values do not include any ROIs even though the container cache does.

comment:22 Changed 11 years ago by mtbcarroll

There was a problem with the Zeiss reader that http://github.com/openmicroscopy/bioformats/pull/694 now fixes.

The next problem is that InputServerStrategy.createROIShape rejects the z = -1, t = -1 shapes coming from test_images_roi/zvi/roi-rectangle-aligned.zvi. Is this a problem with the Bio-Formats reader or with Insight?

comment:23 Changed 11 years ago by mtbcarroll

Will helped me to find that there are no z or t set in the response from the server. (Is that acceptable in the response?) The -1's in Insight are probably coming from the,

if (value == null) return -1;

in getZ() and getT() in ShapeData.

comment:24 Changed 11 years ago by jamoore

a null response means "All Z, T, or C".

comment:25 Changed 11 years ago by mtbcarroll

Thanks.

comment:26 Changed 11 years ago by mtbcarroll

Having Insight interpret no z or t correctly looks like a rather large job, and agents.measurement.view is due for some kind of overhaul anyway. I am happy to fix this issue for 4.4.9 but I expect it would take me a few days; should I push this out to 4.4.x?

comment:27 Changed 11 years ago by jburel

Modify the measurement tool now is dangerous.The entire structure needs to be reviewed and library upgrade (not possible as it stands!)
Please push

comment:28 Changed 11 years ago by mtbcarroll

  • Milestone changed from OMERO-4.4.9 to OMERO-4.4.x
  • Owner changed from mtbcarroll to jburel
  • Sprint Blocker 4.4.9 (1) deleted

comment:29 Changed 10 years ago by jburel

see #9738

comment:30 Changed 8 years ago by jamoore

  • Milestone changed from 5.x to ROI

comment:31 Changed 8 years ago by dlindner

  • Resolution set to fixed
  • Status changed from accepted to closed

Fixed, not reproduceable with latest build.

Note: See TracTickets for help on using tickets. You may also have a look at Agilo extensions to the ticket.

1.3.13-PRO © 2008-2011 Agilo Software all rights reserved (this page was served in: 0.70339 sec.)

We're Hiring!