METHOD FOR ENCODING AND DECODING VIDEO USING MOTION VECTOR PREDICTION, AND APPARATUS THEREOF

摘要

A method for decoding a video using motion information prediction according to an embodiment of the present invention comprises the steps of: decoding some units constituting an image of video data; calculating a variation between peripheral pixels in a predetermined direction for a decoding target unit using reference pixel values of the decoded units; determining a reference pixel region based on the variation; making a candidate list of motion vector predictors based on a motion vector of the reference pixel region; and predicting motion information of the decoding target unit using the candidate list.