Bug: CZI BGR24/BGR48 RGB image pixel data not handled correctly

[rleigh]

See ​http://lists.openmicroscopy.org.uk/pipermail/ome-users/2013-July/003829.html which contains a sample image and jpeg of how it should look.

The sample image is correctly parsed as pixelType==3 (BGR_24), but this is not correctly displayed by showinf. Tested with current dev_4_4 and develop. The image x/y size is correct.

showinf does show " RGB mismatch " so it does look like there's an inconsistency in the RGB/channel count here.

[mlinkert]

This appears to be the same problem: ​http://lists.openmicroscopy.org.uk/pipermail/ome-users/2013-September/003944.html

[rleigh]

​http://qa.openmicroscopy.org.uk/qa/feedback/7601/

See also the images in /ome/team/rleigh/7601 which include the original images uploaded, plus exports of the images from zen (blue) 2012 and zen.txt containing my notes on the image size and pixel type of each image as shown in zen.

Greyscale images are handled correctly. RGB 24- and 48-bit images are not. There are obvious problems incorrectly processing the pixel data here seen when viewing with showinf; it looks like we're not using the correct pixel type.

I can attach screenshots of the bioformats rendering if needed.

[rleigh]

See also QA#7891 and attached image for RBG48 with 14(42)bpp.

[rleigh]

QA#7891; note the pixel type is shown as uint8 *not* 16, which is wrong

[mlinkert]

Fixed with ​https://github.com/openmicroscopy/bioformats/pull/1078; email forthcoming on ome-users/ome-devel with details of how to try this out prior to 5.0.2.