I was wondering if there is a way to find out in what channels the image has been splitted. I have a normal 2D image that I split over channels, so retaining the X,Y plane. Now I can't figure out what colour space the image was, or in what color space my channels are. RGB, HSB etc? Is there a way to find out? The thing is that I use IJ1 macros, but their settings I have to find out in FIJI. Therefore I first need to know how my KNIME worklfow creates the binary image.

The channels are split in the order you put them in. Depending on the fileformat, the color-space may be attached in the metadata of the image (image reader can read the metdata as XML).

Does this help?

Thanks. I found the metadata, but nothing that mentions the colourspace.

Here is a screenshot of most of the text in the xml column. Not everything fits to my screen, but the part that is lacking has information about the time channel.



mhm. maybe the information is not provided with the image at all?

Could be, as this image is created with paint from another. I'll have a try on the original, it would probably be the same.

For the original image the KNIME output is the same. However the colour space is mentioned in the metadata. I found it by using the properties tab of the image in the file browser. So for some reason it doesn't load all metadata.

I think I have the same problem. if I use Fiji I can easily split the channels in R-G-B but here is uncler to me what is happening. I need to do a subtraction between R and G channels and I cannot get my head around it. Any help would be fantastic.

In the image viewer I see the original image uploaded but no rendering of the 3 channels and seems I can only add one column and there is no way to see 3 colums with the 3 channel thumbnails...  Thanks!


Hi Avallepa,

here is some explanation which maybe help you:

Your image is most likely (X,Y,Channel) which means that you have a 3D image. In this image we can intepret the last dimension as (R,G,B) if there are only three elements in the channel available (this means number of channels = 3). It's a way more complicated to think like this, but it provides you with a lot of flexibility.

Ok, this image we render as RGB image in the thumbnail. In the Image Viewer you can browse through the channels  and inspect the individual intensity values of each channel or you can select the RGB renderer to have a colored image. If you want to split the image now into it's three channels, you can use the `Splitter` node and you get three output columns: First Channel, Second Channel and third Channel (they have mystical names, but you can use the Column Renamer to give nicer names). The Image Calculator can then be used to subtract two of channels e.g. by simply doing $FirstChannelName$ - $SecondChannelName$. I'd suggest to choose a image output type which supports negative values (result can be lower than 0). For example FloatType, ByteType, ShortType etc. All the UnsignedXYTypes only support values starting from 0.

If you need any help we can quickly post an example workflow which covers all of the mentoined techniques and aspects.

I hope this helps. If not, let us know!


If you need any help we can quickly post an example workflow which covers all of the mentoined techniques and aspects.

Please find a suitable example workflow attached.


Thanks Christian, an example workflow would be great, actually. I am using the Splitter Uncropper (I must keep images the same size) and I do not understand how you can add 3 columns in the image viewer and see separathe thumbnails... Thanks so much!

Hi Avallepa,

you can see the image in RGB only before the Splitter! Afterwards you have separated three individual channels (try mark them all in the image viewer, then they are displayed next to each other).

I don't understand what you mean by "see separate thumbnails"? After the Splitting, you should have three images, right?