Motion vector generation device, prediction image generation device, video decoding device, and video coding device
To improve a coding efficiency. There are included a PU level search unit configured to search for a motion vector for each prediction block by using a matching process, and a sub-block level search unit configured to search for a motion vector of each of sub-blocks in the prediction block, wherein a precision of a local search by the PU level search unit is lower than a precision of a local search by the sub-block level search unit.
1. A motion vector generation device for generating a motion vector in a current picture, the motion vector generation device comprising:
a merge candidate derivation circuitry that derives a first motion vector for a coding block by using (i) a motion vector of a neighboring block neighboring to the coding block and (ii) a merge index for the coding block; and
a matching prediction parameter derivation circuitry that derives a second motion vector of each sub-blocks for the coding block by the first motion vector,
wherein:
the second motion vector is searched by performing a second local search after performing a first local search for the sub-blocks by using the first motion vector and two reference pictures which are on different sides across the current picture, and
a second a precision of the second local search is higher precision than a precision of the first local search.
2. A prediction image generation device comprising:
the motion vector generation device according to claim 1 , and
a generator that generates a prediction image using the motion vector generated by the motion vector generation device.
3. A video decoding device comprising:
the prediction image generation device according to claim 2 , wherein
a coding target image is reconstructed by adding or subtracting a residual image to or from the prediction image.
4. A video coding device comprising:
the prediction image generation device according to claim 2 , wherein
a residual error between the prediction image and a coding target image is coded.
5. The motion vector generation device of claim 1 , wherein
the first local search is performed with an integer pixel precision and
the second local search is performed with a fractional pixel precision.
6. The motion vector generation device of claim 1 , wherein
the second motion vector is derived base on a sum of absolute difference between a first block, which is derived by using a first reference picture of the two reference pictures, and a second block, which is derived by using a second reference picture of the two reference pictures.