| 摘要 |
Conventional inter-view shift prediction motion vector derivation has a problem in that the process to determine a reference position is complicated. Further, with a sequence parameter set there is a problem in that an ON-OFF flag of a texture extension tool and an ON-OFF flag of a depth extension tool cannot be set independently. There is also a problem in that if either an intra-SDC wedge division flag IntraSdcWedgeFlag or an intra-contour division flag IntraContourFlag is equal to 1, then a depth_intra_mode_flag selecting a wedge division mode or a contour division mode is decoded meaninglessly. This prediction vector deriving device derives reference block coordinates of an inter-view merge candidate IV from a value obtained by normalizing to a multiple of 8 or a multiple of 16 the sum of the upper-left coordinates of a target block and half the size of the target block, and a disparity vector converted to the integer precision of the target block. The prediction vector deriving device also derives reference block coordinates of an inter-view shift-merge candidate IVShift from a value obtained by normalizing to a multiple of 8 or a multiple of 16 the sum of the upper-left coordinates of the target block, the size of the target block, and a prescribed constant between 0 and 4, and the disparity vector converted to the integer precision of the target block. The prediction vector deriving device derives an inter-view merge candidate IV motion vector and an inter-view shift-merge candidate IVShift motion vector from a motion vector located at the derived reference block coordinates. |
