| 摘要 |
FIELD: information technology.SUBSTANCE: invention relates to encoding and decoding of video data. Method of decoding video data comprises steps on which for each unit (PU) prediction in multiple PU belonging to current unit (CU) of coding video data: generating candidate list for PU, wherein generation of candidate list for PU uses only initial position outside current CU, wherein candidate list for PU includes one or more spatial candidates, wherein initial position include at least one of: initial position above current CU, initial position above and on right of current CU, initial position above and to left of current CU, initial position to left of current CU and initial position below and on left of current CU; determining, at least partially based on motion information, said selected candidate in list of candidates for PU, motion information PU; and generating at least partially on basis of reference block, said motion information PU, prognostic video unit for PU.EFFECT: efficient generation of candidate list, which is reduced number of bits required for motion information presentation unit of prediction.72 cl, 30 dwg |
