Bio-Format processing Imagej




I have a question about the Bio-Format processing plug-in. This plugin is used in our group for visualizing tiled Tiffs. However, when I imread the same images in MATLAB the images differ. The MATLAB images show much more background/noise compared to the Bio-Format images. It seems like that in the Bio-Format images some sort of background correction or noise reduction is performed. However, I could not find clear information if this is the case or not.

Does some have more information about this?



Hard to say. What kind of images do you import? 16-bit, 32-bit, etc.?
It might be that you have to adjust the display range in ImageJ.

Have a look at the histogram of the image if the values are correct

Try the Bightness/Contrast action to adjust the display range, see:

“Click on Auto, and ImageJ will automatically optimizes brightness and contrast based on an analysis of the image’s histogram”


Good day,

I don’t think that Bio-Format does any processing behind the scene, i.e. processing that is not explicitly intended by the user.

In fact, things may be much more complicated and depend on the exact kind and formats of the images. Image display may differ although the data stays the same …




It is a 16-bit image. The difference is not in the brightness/contrast. The values of the pixels actually differ between the images. Where the background has high values in the MATLAB image, these same pixels have low values in the ImageJ image compared to the rest of the signal.


Are you sure the values differ, not the visual appearance, i.e. the displayed values?




how are you checking it?
could you please post some example images/screenshots of this behavior?
thank you
Emanuele Martini



In het MATLAB image the background can have values up to 36070, where the circles in the middle have a value of around 13030. These circles is what I am interested in. In the Bio-Format images the background 0,00089 and the circle 0.1954. The ratio between the background and signal differs thus between the two images. Hopefully this makes it a bit more clear.


Could it be that the second image is loaded and displayed as a RGB image in Matlab?

The display (noisy, second image, below the first image) indicates this whereas the first image seems to display the correct range of values.

Where does the original data come from?

What is the expected values range of the original data?


First thanks for you time and advice.

The images are marked as grayscale images in MATLAB.

Originally these images are fluorescence intensity images and these are saved by the system in a tiled TIF file. I’m not sure what the expected value range is of the data.


thanks, could you also upload the tif so we can also do some test?


I agree with @emartini. We need to inspect the original data. I suspect the Matlab display is wrong


This is one of the images. Hopefully it works this way.


With Bio-Formats importer your sample image opens as 32-bit image in ImageJ. Its values are bipolar, reaching from -0.0126 to + 0.0385. The image shows a spatial scale which maps the 1300x964 pixels to 110500x81940 µm.

Here is the central part of the image with full range display:

The image looks like a CT image of a phantom which would explain the bipolar values that may result from scaled Hounsfield units. But I may be wrong.

The image looks plausible.




In addition to the infos of @Herbie the file gives you some information about the imaging software
which was used (Image Studio 5.2.5) and the instrument (Pearl 1.3.4 from LI-COR). Look at the metadata!

So you can control for yourself if the values are correct in ImageJ.

If you have any doubts I would contact LI-COR to verify that the values are reasonable.


Thanks for your information and advice.

Maybe I misunderstood, but my main problem was the difference in visualization between MATLAB and ImageJ. Do you maybe have any advice about that?


From your first images I guess that the display of your sample image in MATLAB is at least inadequate. Because this forum is an ImageJ forum and because I’m lacking experience with MATLAB I can’t help your any further. If however you have further questions about ImageJ and image processing in general you will find great expertise on this forum.




Oké! Thank you for all your help!