Motion vector derivation apparatus, video decoding apparatus, and video coding apparatus
A load in processing of searching for a motion vector is reduced. In order to solve the problem described above, a motion vector derivation apparatus ( 3032, 3036 to 3038 ) according to one aspect of the present invention that derives a motion vector to be referred to for generating a prediction image to be used for coding or decoding of a video includes a motion vector search unit ( 30373 ) configured to search for a motion vector on a prediction unit basis through matching processing. The motion vector search unit ( 30373 ) is configured to stop search of the motion vector, depending on whether or not a conditional expression according to a pixel bit-depth is satisfied.
1. A motion vector derivation apparatus for deriving a motion vector for each of subblocks, the motion vector derivation apparatus comprising:
a matching prediction processing circuit that derives a first prediction block from a first reference picture and a second prediction block from a second reference picture,
that derives a sum of absolute difference between the first prediction block, which is defined by locations, and the second prediction block, which is defined by the locations,
that derives an initial motion vector by using the sum of absolute difference, and
that derives a motion vector of a subblock by performing a local search by using the initial motion vector, wherein:
the sum of absolute difference is derived by using (i) the first prediction block derived by right shifting a first value of pixel bits by a shift value and (ii) the second prediction block derived by right shifting a second value of pixel bits by the shift value, and
the shift value is derived by subtracting a predetermined value from a bit depth value.
2. The motion vector derivation apparatus according to claim 1 , wherein the initial motion vector is derived by a comparison between the sum of absolute difference and a prescribed threshold.
3. A motion vector derivation method for deriving a motion vector for each of subblocks, the motion vector derivation method comprising:
deriving a first prediction block from a first reference picture and a second prediction block from a second reference picture,
deriving a sum of absolute difference between the first prediction block, which is defined by locations, and the second prediction block, which is defined by the locations,
deriving an initial motion vector by using the sum of absolute difference and
deriving a motion vector of a subblock by performing a local search by using the initial motion vector, wherein:
the sum of absolute difference is derived by using (i) the first prediction block derived by right shifting a first value of pixel bits by a shift value and (ii) the second prediction block derived by right shifting a second value of pixel bits by the shift value, and
the shift value is derived by subtracting a predetermined value from a bit depth value.
4. A video decoding apparatus comprising:
a motion vector derivation apparatus according to claim 1 ; and
a prediction image generation unit configured to generate a prediction image with reference to a motion vector.
5. A video coding apparatus comprising:
a motion vector derivation apparatus according to claim 1 ; and
a prediction image generation unit configured to generate a prediction image with reference to a motion vector.
6. The motion vector derivation apparatus according to claim 1 , wherein the shift value is a value for luminance.