Effect Of Digital Image Stabilization On Coding Performance

A digital image stabilization system compensates the image movement caused by hand jiggles and thereby improves the perceptual quality of the image sequence captured by a hand-held digital video camera. In this paper, we apply two recently developed video coding standards, MPEG-4 and H.264, to investigate the effect of image stabilization on video coding. The coding performance of these algorithms on the stabilized image sequence is compared to that on the original image sequence. The results show that image stabilization greatly improves the viewing experience and results in a decrease of bit counts for motion vectors, but the difference in the overall coding performance is minor. Details of the video coding experiments and the image stabilization system are described.