| 摘要 |
A method and an apparatus for extracting a motion vector using a multi-resolution search algorithm are provided to extract a correct motion vector by converting a current image and a reference image into multi-resolution images, giving weights to the multi-resolution images and controlling a motion vector, which is extracted from a low-resolution image, in a high-resolution image to determine a final motion vector. A current image and a reference image are converted into multi-resolution images. Predetermined weights are given to a low-resolution current image and a low-resolution reference image(110). The entire region of the low-resolution reference image is searched with the low-resolution current image to extract a motion vector. An intermediate-resolution image(120) is searched to a predetermined region to control the extracted motion vector. A high-resolution image(130) is searched for a predetermined region to control the controlled motion vector again.
|
