Prediction image generation device, moving image decoding device, moving image encoding device, and prediction image generation method
A prediction image generation method using two prediction images to generate a prediction image by a device is provided. First and second prediction image are generated. Bidirectional prediction gradient change prediction processing is performed by using a first shift value and difference values of the first and second prediction images respectively in horizontal and vertical directions to generate a first, second, third and fourth gradient images. Motion information is derived by using the first and second prediction images, the first, second, third and fourth gradient images, a second shift value, and a third shift value. Motion compensation correction value is derived by using the motion information and the first, second, third and the fourth gradient images. The prediction image is generated by using the first and second prediction images and the motion compensation correction value. The first, second and third shift values are respectively equal to 6, 4 and 1.
1. A prediction image generation device for generating prediction images, the prediction image generation device comprising:
prediction image generation circuitry generating a first prediction image and a second prediction image;
gradient image generation circuitry performing bidirectional prediction gradient change prediction processing by:
generating a first gradient image by calculating a first difference value of two horizontally neighboring samples of each of a plurality of current samples of the first prediction image and right shifting the first difference value by a first shift value of six;
generating a second gradient image by calculating a second difference value of two vertically neighboring samples of each of the plurality of current samples of the first prediction image and right shifting the second difference value by the first shift value;
generating a third gradient image by calculating a third difference value of two horizontally neighboring samples of each of a plurality of current samples of the second prediction image and right shifting the third difference value by the first shift value; and
generating a fourth gradient image by calculating a fourth difference value of two vertically neighboring samples of each of the plurality of current samples of the second prediction image and right shifting the fourth difference value by the first shift value;
motion compensation correction value derivation circuitry that:
uses the first prediction image, the second prediction image, and a second shift value of four to generate a first intermediate image;
uses the first gradient image, the third gradient image, and a third shift value of one to generate a second intermediate image;
uses the second gradient image, the fourth gradient image, and the third shift value to generate a third intermediate image;
uses the first intermediate image, the second intermediate image, and the third intermediate image to derive motion information; and
uses the motion information, the first gradient image, the second gradient image, the third gradient image, and the fourth gradient image to derive a motion compensation correction value; and
prediction image generation circuitry that uses the first prediction image, the second prediction image, and the motion compensation correction value to generate a third prediction image,
wherein the first shift value, the second shift value, and the third shift value are derived based on a pixel bit length bitDepth of eight.
2. The prediction image generation device according to claim 1 , wherein the motion compensation correction value derivation uses a first threshold of 16 to derive the motion information.
3. A moving image decoding device comprising the prediction image generation device according to claim 1 , wherein an encoding object image is restored by adding a residual image to or subtracting the residual image from the generated third prediction image.
4. A moving image encoding device comprising the prediction image generation device according to claim 1 , wherein a residual of the generated prediction image and an encoding object image is encoded.
5. A prediction image generation method for generating prediction images, the prediction image generation method comprising:
generating a first prediction image and a second prediction image; and
performing bidirectional prediction gradient change prediction processing by:
generating a first gradient image by calculating a first difference value of two horizontally neighboring samples of each of a plurality of current samples of the first prediction image and right shifting the first difference value by a first shift value of six;
generating a second gradient image by calculating a second difference value of two vertically neighboring samples of each of the plurality of current samples of the first prediction image and right shifting the second difference value by the first shift value;
generating a third gradient image by calculating a third difference value of two horizontally neighboring samples of each of a plurality of current samples of the second prediction image and right shifting the third difference value by the first shift value;
generating a fourth gradient image by calculating a fourth difference value of two vertically neighboring samples of each of the plurality of current samples of the second prediction image and right shifting the fourth difference value by the first shift value;
using the first prediction image, the second prediction image, and a second shift value of four to generate a first intermediate image;
using the first gradient image, the third gradient image, and a third shift value of one to generate a second intermediate image;
using the second gradient image, the fourth gradient image, and the third shift value to generate a third intermediate image;
using the first intermediate image, the second intermediate image, and the third intermediate image to derive motion information;
using the motion information, the first gradient image, the second gradient image, the third gradient image, and the fourth gradient image to derive a motion compensation correction value; and
using the first prediction image, the second prediction image, and the motion compensation correction value to generate a third prediction image,
wherein the first shift value, the second shift value, and the third shift value are derived based on a pixel bit length bitDepth of eight.
6. The prediction image generation method according to claim 5 , further comprising:
using a first threshold of 16 to derive the motion information.