Method and device for image coding using motion vector differences
According to embodiments of the present document, a prediction procedure may be performed for image/video coding, and the prediction procedure may include symmetric motion vector differences (SMVDs) and merge mode motion vector differences (MMVDs) according to an inter prediction. The inter prediction may be performed on the basis of reference pictures of a current picture, and types (e.g., a long term reference picture, a short term reference picture, etc.) of the reference pictures may be considered for the inter prediction. Therefore, performance and coding efficiency in the prediction procedure can be improved.
1. An image decoding method performed by a decoding apparatus, the method comprising:
receiving image information including prediction related information and information on a motion vector difference (MVD) through a bitstream;
deriving an inter prediction mode for a current block based on the prediction related information;
constructing a motion vector predictor (MVP) candidate list for the current block based on neighboring blocks of the current block;
deriving an MVP for the current block based on the MVP candidate list;
deriving motion information for the current block based on the information on the MVD and the MVP; and
generating predicted samples for the current block based on the motion information,
wherein the motion information includes a motion vector and a reference picture index,
wherein the motion vector is derived based on the information on the MVD and the MVP,
wherein the prediction related information includes information on symmetric motion vector differences (SMVD), and
wherein based on a value of the information on the SMVD being 1, the reference picture index is derived based on a picture order count (POC) difference between a short-term reference picture among reference pictures included in reference picture lists and a current picture including the current block.
2. The method of claim 1 , wherein an MVD is derived based on the information on the MVD, and the motion vector for the current block is derived based on a sum of the MVD and the MVP.
3. The method of claim 1 ,
wherein an MVD L 0 for L 0 prediction is derived based on the information on the MVD,
an MVD L 1 for L 1 prediction is derived from the MVD L 0 based on the value of information on the SMVD being 1, and
the motion information is derived based on the MVD L 0 and the MVD L 1 .
4. The method of claim 3 , wherein a size of the MVD L 1 is the same as that of the MVD L 0 , and
a sign of the MVD L 1 is opposite to that of the MVD L 0 .
5. The method of claim 1 , wherein the reference picture lists include reference picture list 0 ,
the reference picture list 0 includes the short-term reference picture, and
the reference picture index is derived based on picture order count (POC) differences between each of the reference pictures included in the reference picture list 0 and the current picture including the current block.
6. The method of claim 1 , wherein
the reference picture index is derived based on the information on the SMVD without being signaled directly.
7. An image encoding method performed by an encoding apparatus, the method comprising:
deriving an inter prediction mode for a current block;
constructing a motion vector predictor (MVP) candidate list for the current block based on neighboring blocks of the current block;
deriving an MVP for the current block based on the MVP candidate list;
deriving motion information including a motion vector and a reference picture index for the current block;
deriving a motion vector difference (MVD) for the current block based on the MVP and the motion vector;
generating prediction related information related to the inter prediction mode and information on the MVD for the current block;
generating predicted samples for the current block based on the motion information;
generating residual information based on the predicted samples; and
encoding image information including the prediction related information, the information on the MVD and the residual information,
wherein the prediction related information includes information on symmetric motion vector differences (SMVD), and
wherein based on a value of the information on the SMVD being 1, the reference picture index is derived based on a picture order count (POC) difference between a short-term reference picture among reference pictures included in reference picture lists and a current picture including the current block.
8. The method of claim 7 ,
wherein the MVP candidate list for the current block includes an MVP candidate list for L 0 prediction and an MVP candidate list for L 1 prediction,
wherein an MVP L 0 for L 0 prediction is derived based on the MVP candidate list for L 0 prediction, and an MVP L 1 for L 1 prediction is derived based on the MVP candidate list for L 1 prediction,
wherein an MVD L 0 for L 0 prediction is derived based on the MVP L 0 and the motion vector of the current block, and an MVD L 1 for L 1 prediction is derived based on the MVP L 1 and the motion vector of the current block, and wherein information on the MVD L 0 is encoded and information of the MVD L 1 is not encoded based on the value of information on the SMVD being 1.
9. The method of claim 8 , wherein a size of the MVD L 1 is the same as that of the MVD L 0 , and
a sign of the MVD L 1 is opposite to that of the MVD L 0 .
10. The method of claim 7 , wherein the reference picture lists include reference picture list 0 ,
the reference picture list 0 includes the short-term reference picture, and
the reference picture index is derived based on picture order count (POC) difference between each of reference pictures included in the reference picture list 0 and the current picture including the current block.
11. A non-transitory computer readable storage medium storing a bitstream generated by a method, wherein the method comprising:
deriving an inter prediction mode for a current block;
constructing a motion vector predictor (MVP) candidate list for the current block based on neighboring blocks of the current block;
deriving an MVP for the current block based on the MVP candidate list;
deriving motion information including a motion vector and a reference picture index for the current block;
deriving a motion vector difference (MVD) for the current block based on the MVP and the motion vector;
generating prediction related information related to the inter prediction mode and information on the MVD for the current block;
generating predicted samples for the current block based on the motion information;
generating residual information based on the predicted samples; and
encoding image information including the prediction related information, the information on the MVD and the residual information,
wherein the prediction related information includes information on symmetric motion vector differences (SMVD), and
wherein based on a value of the information on the SMVD being 1, the reference picture index is derived based on a picture order count (POC) difference between a short-term reference picture among reference pictures included in reference picture lists and a current picture including the current block.