Though the demo was quite impresive, and definitely more advanced than any other I've seen before, the "shakyness" is still there, detaching the "augmented" and

Though the demo was quite impresive, and definitely more advanced than any other I've seen before, the "shakyness" is still there, detaching the "augmented" and the "regular" reality. What is it preventing a more accurate analysis of the video frames? Too low resolution? Changing lighting conditions? Lack of processing power? I can't see how these cannot be solved with current optical and computing equipment. Is it, then, a matter of not having developed sophisticate enough image analysis (vision) algorithms?