Idea – Could Recognition Improve Image Stacking Processing To Reduce Noise From Non Aligned Images?

My idea is simple

Maybe the reason why we can see so good with two eyes is that the brain uses recognition.

So the idea is that if one eye image is blurred then it can still provide recognition data.

I wonder if recognition and machine learning then can improve image stacking. That is. Reduce noise and improve the image from several images that are not aligned.

Sounds simple enough.