Ticket #1312 (closed defect: fixed)

Opened 7 years ago

Last modified 5 years ago

Process > FFT > Bandpass Filter fails on some images

Reported by: bdezonia Owned by: bdezonia
Priority: major Milestone: imagej2-b8-analysis
Component: Legacy Compatibility Version:
Severity: serious Keywords:
Cc: Blocked By:
Blocking: #1459

Description (last modified by bdezonia) (diff)

Bandpass Filter works for some images (such as Boats). But running with Organ Of Corti you can see differences in all Z planes of the IJ1 and IJ2 versions.

Change History

comment:1 Changed 7 years ago by bdezonia

  • Blocked By 1331 added

comment:2 Changed 7 years ago by bdezonia

  • Blocking 1331 added
  • Blocked By 1331 removed

comment:3 Changed 7 years ago by bdezonia

In IJ1 the current channel color is highlighted in the current slice with all colors present and the highlighted color does not appear in any other slices when moving the slider.

In IJ2 the current channel color is SUPER highlighted in the current slice along with the other colors but the highlighted color also still appears in all other slices when moving the slider.

Last edited 7 years ago by bdezonia (previous) (diff)

comment:4 Changed 7 years ago by bdezonia

  • Description modified (diff)

comment:5 Changed 7 years ago by bdezonia

Can hack things in DefaultDatasetView::setupProjector() to see that we can do projector.setComposite(0,false) which makes all slices look like the values in IJ1 for the nonactive slices. We need to manipulate the setting of composite channel booleans in the legacy layer somehow. It also may be the case that this code was not completed. See what IJ1 can do and see how we can replicate.

comment:6 Changed 6 years ago by bdezonia

  • Blocking 1459 added; 1331 removed
  • Milestone changed from imagej-2.0.0-beta4 to imagej-2.0.0-beta5

comment:7 Changed 6 years ago by bdezonia

  • Milestone changed from imagej2-b7-ndim-data to imagej2-b8-analysis

comment:8 Changed 5 years ago by bdezonia

After further looking it seems like this might just be a channel display range setting issue. The lut min/max values are all set to 40/4095 and yet channel 0 ranges from 0 to 60000. So in IJ2 the red pixels show very brightly. If the lut min/max represented actual color values this may look better.

comment:9 Changed 5 years ago by bdezonia

Indeed in the color table harmonizer if I set the display ranges to the actual min max values of the data per channel the image displays completely correctly. This test approach is too slow for actual use but uncovers the problem. Somehow the passed CompositeImage has the wrong values for channel mins/maxes in its luts. Maybe our legacy layer approach of capturing imp.show() is keeping a reset min/max kind of call from being performed.

comment:10 Changed 5 years ago by bdezonia

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

Fixed with commit 15ce640dd832edad3453156383214b19411dfc20

Note: See TracTickets for help on using tickets.