Saving fixed pixel locations in presets and subsequently comparing those pixels.

Question

Hello world.
I am abolutely new to plugin development and have been ploughing through the SDK and sample projects thus far, but I had a few questions about the approach with respect to what I wish to achieve and thought it would be best to pick the brains of everyone here.
I need to compare frames from a source video to a sample image which I shall call the target. I found how to do this using the Checkout example as well as the tutorial over at Mactech. However, the problem gets complex where I need to compare only fixed pixels from the source frame to certain fixed pixels in the target image. To give a better insight, I believe I should describe the problem statement a little better.
Scenario:
A user saves a target image along with the fixed pixels for comparison as a preset. The user also chooses the fixed pixels in a test source frame and saves this as well. The plugin takes the source video and the preset as an input and compares the fixed pixels to the target pixels. By 'compares' I mean some complex math would be performed in the CIE Lab space, but that is a discussion for later, and I'm taking this one step at a time.
Can someone please guide me as to how to set up this system to start with?
In short:
1. How do I store the x and y's of the fixed pixels in the source test frame and target as 2 presets? Possibly a custom data structure that would be bundled with the image?
2. How do I then access these and iterate only over those fixed pixels in a source frame and target image?
I am sorry if these are ridiculous questions, but I've been reading for a while and don't seem to be able to find these answers myself.
Thanks,
Chirag

Answer 1

hello chirag_raman!
welcome to our little forum, where everyone has an opinion, but no one knows for sure. :-)
what you're asking is not complex, so let's tackle the issues:
1. how to save a set of x,y coordinates:
i'd say it will mostly depend on how the user let's you know what these coordinates are.
i would imagine the simplest way would be to use a point param.
you can keep the coordinates in a point param and make it hidden in the case where you don't want the user to see it.
but there are a couple of other ways of keeping data with a plug-in.
you can either use an "arbitrary" type param, or use the sequence_data.
both of the above are means to store a structure of data that you create. you can put anything you want in it.
both methods allow you to modify the size of the stored data at (almost) any time.
in both cases the data is stored with the project file. (unless you don't want it to)
the differences are:
sequence_data -
you only have one of these per instance of your plug-in.
you handle the changing of the data, and it is not cached in the undo stack.
a change in the data doesn't cause a re-render.
arbitrary param-
you can have as many as you like in each plug-in.
a change of data in such param is undoable by the user.
a change causes a re-render.
look at the "ColorGrid" sample to see how to handle an arbitrary param, and the "Supervisor" sample for how to handle the sequence_data.
2. how to access specific pixels.
a couple of ways.
you can use the sample suite to grab the color of a pixel/subpixel/area from a buffer,
or you can access the position in ram directly and grab the data yourself.
look at the "shifter" sample to see the sample suite in action, and the "CCU" sample to see how to access the position in ram directly.
i hope this helps.
cheers!