Method and apparatus for encoding/decoding motion vector
Provided are methods and apparatuses for encoding and decoding a motion vector including a method of decoding that includes obtaining a current coding unit hierarchically split from a maximum coding unit according to a current depth, obtaining a prediction mode information of a current prediction unit in the current coding unit from bitstream, when a prediction mode of the current prediction unit is inter-prediction mode, determining motion vector predictor candidates from among motion vectors of neighboring prediction units adjacent to the current prediction unit, and determining a motion vector predictor of the current prediction unit from among the motion vector predictor candidates, wherein the neighboring prediction unit comprises a first block directly to the left side of a leftmost block among blocks located directly to a bottom side of the current prediction unit and a second block located directly to a upper side of the first block.
1. A method of decoding an image, the method comprising:
obtaining a current coding unit hierarchically split from a maximum coding unit according to a current depth;
obtaining prediction mode information of a current prediction unit in the current coding unit from a bitstream;
when a prediction mode of the current prediction unit indicated by the prediction mode information is inter-prediction mode, determining motion vector predictor candidates from among motion vectors of neighboring prediction units adjacent to the current prediction unit;
determining a motion vector predictor of the current prediction unit from among the motion vector predictor candidates; and
generating a decoded video by using the determined motion vector predictor,
wherein the neighboring prediction units comprise a first block directly to a left side of a leftmost block among blocks located directly below a bottom side of the current prediction unit and a second block located directly above an upper side of the first block.
2. The method of claim 1 , wherein the motion vector predictor candidates further comprise a motion vector of a co-located prediction unit at a position in a reference picture co-located with a position of the current prediction unit in a current picture.
3. The method of claim 2 , wherein the motion vector of the co-located prediction unit is scaled based on a temporal distance between the reference picture and the current picture.