Motion blur estimation of handheld camera using regular- and short-exposure image pair

We present in this paper a new image stabilization technology to address motion blurs from camera shake with a regular- and short-exposure image pair taken consecutively. The proposed motion estimation process in this technology adopts a Wiener filter to estimate motion blur using few selected feature corresponding points in the image pair. To obtain a more reliable and precise result, a multi-resolution and morphological image processing technique is then used to further refine the initial estimation. This technology has been applied successfully as an image stabilization function on a consumer digital still camera. The resulting performance is 0.7-1.4 EV exposure time step gain with an average processing time of 3.1 seconds for a 6 M pixel image.