Decide on a way to do it

[michal2229]

Problem 1:
when video is shaking, there is different part of scene visible, so when stabilized, there will be missing part of scene on one side and on onther side of a frame there will be image cut.

[michal2229]

Options:

• we can determine transform matrices for each frame and then encode video using input frames and these matrices
• we can expand to a full scene of video footage, so that it will not cut any data out of frame
• we can stack frames to have better background where image is missing, first frame will have background of last, fully stacked frame