What I was trying to do is to find
out a way to figure out where the object is in the first place. When I
take pictures I try to simplify the subject and make everything that is
in the picture but irrelevant to the idea somehow unimportant and less
visible. So, blurring it out sounds good (yes, one indeed can trick that
camera that we have into doing a lot of things it is not design to do :-)
). Now if I take the picture this way I can select everything that is in
focus (well, to a certain extent) and try to do something to amplify the
background. FFT from PS1 should come in handy. So, I will take a few images
which are focused on different things in the same scene and select what
I want to focus on with the mouse. Then I will display the image from the
set, which has the greates power in its high frequency components in that
area. So, it is like autofocus.
Another thing that I wanted to
do was to reassemble the image out of this stack which is the most focused
at every point.
And the third thing is to try to simulate
the hand coloring thing. That is, having that resynthesized image I would
only keep the original color for everything that is in focus in the image
in the stack which has the most focused part where the mouse is clicked.
Sounds confusing, but it is really pretty simple.
