Fluorescence images in .VSI forrmat

[H. Dale]

Hi. 

I'm just starting to investigate QuPath and so far it seems very useful. We have a Olympus Slide scanner, file format .vsi, which also acquires fluorescent images. I have a few questions related to this. I totally understand that the main purpose of QuPath is handling brightfield color images, but it would be very useful to do simple operations on fluorescence as well. 

1) When I open a 3 channel vsi-file in QuPath the blue and red LUT is mixed up, green is correct. I get red nuclei which should be blue. Is there a way to change the LUT back to the original? 

2) Vsi-files are not possible to open in Fiji (ImageJ), and the idea of directly sending them to ImageJ from QuPath is very appealing and saves extra work. At first glance it seemed all well, the image comes up as a 3 channel image in grey scale. But when changing the LUT it's not possible to do that individually on the different channels. Changing Ch1 to blue changes all channels to blue. Another approach would be to split the channels before applying a new LUT, but then the message "Multichannell image required" comes up, so it seems like it's is not registered fully as a multichannel image, even though the image properties seem correct and show 3 channels. Is this a known problem and a possible solution to?

Hope to hear from you!

Best regards 

Hege

[micros...@gmail.com]

If you are opening the files through BioFormats, and have the newest BioFormats extension and plugin, you *should be able to double click on a channel in the Brightness/contrast (half moon button) settings to change the display color. If your images are being opened with OpenSlide (the server type, BioFormats or Openslide is listed in the Image tab on the left), there is no way to adjust the channels since it is locked as an RGB image.  VSI images should be visible with BioFormats, though, and an upcoming version of QuPath may have a fix included for channel mixups.

If all you need is data from your images, none of that should be a problem since you can look for nuclei in any channel. If you want to make pictures for a presentation, that might be more problematic using OpenSlide.

[H. Dale]

Thank you for the answer, changing the color by double clicking worked fine!

It would have been of great help if anyone have a comment to questions 2 as well, regarding sending fluorescence images via QuPath to ImageJ.

BR, Hege

[Pete]

For 2. within ImageJ run Image -> Hyperstacks -> Stack to Hyperstack

[H. Dale]

Great!