Cross-component linear model prediction image generation apparatus, video decoding apparatus, video coding apparatus, and prediction image generation method
The amount of memory required for CCLM prediction is reduced. The CCLM prediction parameter derivation unit ( 310442 ) derives a scale shift value corresponding to a luma difference value, and derives a CCLM prediction parameter, by shifting, by using the scale shift value, a value obtained by multiplying a value of a table referred to with a value obtained by performing right shift of the luma difference value by the scale shift value as an index and a chroma difference value. In addition, in a case of deriving a prediction image, a bit width is reduced by adaptively deriving a shift amount of a linear prediction parameter from the chroma difference value.
1. A prediction image generation apparatus for generating a prediction image based on an intra prediction mode, the prediction image generation apparatus comprising:
Cross-Component Linear Model (CCLM) prediction parameter circuitry that derives:
(i) a difference value for a luminance component, which is a difference between a first value and a second value;
(ii) a difference value for a chrominance component, which is a difference between a third value and a fourth value;
(iii) a CCLM prediction parameter by using the difference value for the luminance component and the difference value for the chrominance component; and
a CCLM prediction filter that derives the prediction image by using the CCLM prediction parameter, wherein
the CCLM prediction parameter circuitry derives:
(i) a first logarithmic value by using the difference value for the luminance component;
(ii) a second logarithmic value by using an absolute value of the difference value for the chrominance component; and
(iii) a first shift value by using the first logarithmic value and the second logarithmic value;
the CCLM prediction filter derives the prediction image by using the first shift value;
the CCLM prediction parameter is derived by right shifting a sixth value by the second logarithmic value; and
the sixth value is a value obtained by using a fifth value and the difference value for the chrominance component, the fifth value being a value in a reference table defined by using the difference value for the luminance component.
2. The prediction image generation apparatus of claim 1 , wherein the value in the reference table is defined by using an index derived by right shifting the difference value for the luminance component.
3. The prediction image generation apparatus according to claim 1 , wherein
the prediction image is derived by right shifting a product which is a value obtained by multiplying the CCLM prediction parameter and a luma sample value by the first shift value.
4. A video decoding apparatus configured to decode an image by adding a prediction image generated from the CCLM prediction filter according to claim 1 and a residual.
5. A video coding apparatus configured to, from a difference between a prediction image derived from the CCLM prediction filter according to claim 1 and an input image, derive and code a residual.
6. A prediction image generation method for generating a prediction image based on an intra prediction mode, the prediction image generation method comprising at least the steps of:
deriving a difference value for a luminance component, which is a difference between a first value and a second value;
deriving a difference value for a chrominance component, which is a difference between a third value and a fourth value;
deriving a first logarithmic value by using the difference value for the luminance component;
deriving a second logarithmic value by using an absolute value of the difference value for the chrominance component;
deriving a first shift value by using the first logarithmic value and the second logarithmic value;
deriving a Cross-Component Linear Model (CCLM) prediction parameter by using the difference value for the luminance component and the difference value for the chrominance component; and
deriving the prediction image by using the CCLM prediction parameter and the first shift value; wherein
the CCLM prediction parameter is derived by right shifting a sixth value by the second logarithmic value; and
the sixth value is a value obtained by using a fifth value and the difference value for the chrominance component, the fifth value being a value in a reference table defined by using the difference value for the luminance component.