Method and apparatus for processing video signal on basis of inter prediction
A method and device derive a motion vector difference of a current block based on layer information and index information, where the layer information indicates a current layer to which a motion vector difference used in inter prediction of the current block belongs and index information indicates a specific combination in the current layer, and thereby the MVD coding efficiency may be significantly increased by coding the horizontal and vertical components of the MVD together based on layer information and index information.
1. A method of processing a video signal based on inter prediction by an apparatus, the method comprising:
deriving a motion vector predictor based on motion information of a neighboring block of a current block based on a motion vector prediction (MVP) mode applied to the current block;
obtaining layer information indicating a current layer to which a motion vector difference used in inter prediction of the current block belongs, in a pre-defined layer structure in which at least one combination of horizontal and vertical components of motion vector differences is divided into a plurality of layers;
obtaining index information indicating a specific combination in the current layer;
deriving the motion vector difference of the current block based on the layer information and the index information;
deriving a motion vector of the current block based on the motion vector predictor and the motion vector difference; and
generating a prediction block of the current block based on the motion vector of the current block,
wherein obtaining the layer information includes:
obtaining a first syntax element representing whether an identification (ID) of the current layer is greater than 0; and
obtaining ID information indicating the ID of the current layer based on the ID of the current layer is greater than 0.
2. The method of claim 1 , wherein the ID information is binarized based on an exponential Golomb code with order 1.
3. The method of claim 1 , wherein the index information is binarized based on a truncated binarization scheme.
4. The method of claim 1 , wherein obtaining the layer information further comprises:
obtaining a second syntax element indicating whether the ID of the current layer belongs to a first group based on the ID of the current layer is greater than 0; and
obtaining a third syntax element indicating whether the ID of the current layer is 1 or 2 based on the ID of the current layer belonging to the first group.
5. The method of claim 4 , wherein when the ID of the current layer does not belong to the first group, the ID of the current layer is determined as 3.
6. A method of encoding a video signal based on inter prediction by an apparatus, the method comprising:
deriving a motion vector of a current block based on the inter prediction;
deriving a motion vector predictor based on motion information for a neighboring block of the current block;
generating a motion vector difference of the current block based on the motion vector and the motion vector predictor of the current block;
generating layer information indicating a current layer to which the motion vector difference belongs, in a predefined layer structure in which combinations of at least one horizontal and vertical components of motion vector differences are divided into a plurality of layers; and
generating index information indicating a specific combination in the current layer,
wherein generating the layer information includes:
generating a first syntax element representing whether an identification (ID) of the current layer is greater than 0; and
generating ID information indicating the ID of the current layer based on the ID of the current layer is greater than 0.
7. A non-transitory computer-readable storage medium for encoded image information generated by performing steps of:
deriving a motion vector of a current block based on the inter prediction;
deriving a motion vector predictor based on motion information for a neighboring block of the current block;
generating a motion vector difference of the current block based on the motion vector and the motion vector predictor of the current block;
generating layer information indicating a current layer to which the motion vector difference belongs, in a predefined layer structure in which combinations of at least one horizontal and vertical components of motion vector differences are divided into a plurality of layers; and
generating index information indicating a specific combination in the current layer,
wherein generating the layer information includes:
generating a first syntax element representing whether an identification (ID) of the current layer is greater than 0; and
generating ID information indicating the ID of the current layer based on the ID of the current layer is greater than 0.