Image processing device and method
The present disclosure relates to an image processing device and method that enable suppression of an increase in the amount of coding of a quantization matrix. An image processing device of the present disclosure includes an up-conversion unit configured to up-convert a quantization matrix limited to a size less than or equal to a transmission size that is a maximum size allowed for transmission, from the transmission size to a size that is identical to a block size that is a processing unit of quantization or dequantization. The present disclosure is applicable to, for example, an image processing device for processing image data.
1. An image processing device, comprising:
a memory; and
processing circuitry configured to
set a 16×16 quantization matrix corresponding to a 16×16 transform unit by performing a nearest neighbor interpolation process in which each element of a first 8×8 quantization matrix is copied as a value for its 3 neighboring elements in a case where the 16×16 transform unit is selected when image data is to be orthogonally transformed and set a 32×32 quantization matrix, which corresponds to a 32×32 transform unit, by performing a nearest neighbor interpolation process in which each element of a second 8×8 quantization matrix is copied as a value for its 15 neighboring elements in a case where the 32×32 transform unit is selected when the image data is to be orthogonally transformed;
quantize transform coefficient data generated by orthogonally transforming the image data in 16×16 transform units using the 16×16 quantization matrix and quantize coefficient data generated by orthogonally transforming the image data in 32×32 transform units using the 32×32 quantization matrix to generate quantized data; and
encode the quantized data to generate a bitstream which includes the first 8×8 quantization matrix and the second 8×8 quantization matrix.
2. The image processing device according to claim 1 , wherein the processing circuitry sets the first 8×8 quantization matrix and the second 8×8 quantization matrix as a picture parameter set of the bitstream.
3. An image processing method, comprising:
setting a 16×16 quantization matrix corresponding to a 16×16 transform unit by performing a nearest neighbor interpolation process in which each element of a first 8×8 quantization matrix is copied as a value for its 3 neighboring elements in a case where the 16×16 transform unit is selected when image data is to be orthogonally transformed and setting a 32×32 quantization matrix, which corresponds to a 32×32 transform unit, by performing a nearest neighbor interpolation process in which each element of a second 8×8 quantization matrix is copied as a value for its 15 neighboring elements in a case where the 32×32 transform unit is selected when the image data is to be orthogonally transformed;
quantizing transform coefficient data generated by orthogonally transforming the image data in 16×16 transform units using the set 16×16 quantization matrix and quantizing coefficient data gathered by orthogonally transforming the image data in the 32×32 transform units using the 32×32 quantization matrix to generate quantized data; and
encoding the quantized data to generate a bitstream which includes the first 8×8 quantization matrix and the second 8×8 quantization matrix.
4. The image processing method according to claim 3 , wherein the 8×8 quantization matrix and the second 8×8 quantization matrix are set as a picture parameter set of the bitstream.