Basically what you are asking does not depend on OpenCV, but on your GUI environment that you are using, like for example winforms.
OpenCV does not provide an easy way to place the namedWindows in a container element on a GUI. Nor does it supply a system to combine multiple namedWindows into a single graphical intertface. This is because the OpenCV library is basically image processing algorithms, not visualization in his main focus.
What you could do is the following:
- Create a container Mat element that has dimensions of 4 seperate inputs
- Read in frame, apply the operations on it to receive 4 image outputs
- Set the region of interest on your container Mat and copy the outputs to it. Do this 4 times by shifting the region of interest.
- Display the result container in a namedWindow
It all depends on your operation time on the images if you can get this real time or not.