I created a program (and I have seen one before) that reads individual frames from 2 avi's recorded from a front view DV camera and a side view DV camera. With a suit on and colors on the joints you can point to these joints and mark them frame by frame (time consuming but works better than keyframe). What I need to know is how to scale my pixel points that were recorded from the program into blender 3d points. Then how to apply the 3d points to empties in blender for an armature skeleton to move. Is python the way to do this? If so, can python read a text file inside blender and automatically send the points in the text file to an empties xyz coordinates?

theres a sequencer plugin for blender called the poor mans motion tracker it used to be at http://www-users.cs.umn.edu/~mein/blender/plugins/sequence.html but i cant find it anymore

i think blender python can read text files from inside its text window, if not you could just generate them to the HD and read them from there

It sucked so I took it off of the main site

You can still grab it from http://www.cs.umn.edu/~mein/blender/plugins/mein/pmmotion
If you really want it.