Image processing device, bit stream generation method, coefficient data generation method, and quantization coefficient generation method
In encoding an image, a transform skip flag that is flag information indicating, for each component, whether or not to skip transform processing of transforming a residual between an image and a predicted image of the image into coefficient data is generated, the transform skip flag generated is encoded, coded data of the transform skip flag is generated, and a bit stream including the generated coded data of the transform skip flag is generated. The present encoding/decoding can be applied to, for example, an image processing device, an image encoding device, an image decoding device, a transmission device, a reception device, a transmission-reception device, an information processing device, an imaging device, a reproduction device, a bit stream generation method, a coefficient data generation method, a quantization coefficient generation method, or the like.
1. An image processing device comprising:
a flag decoder that decodes coded data of a transform skip flag corresponding to a component identifier and obtains the transform skip flag corresponding to the component identifier, wherein the component identifier identifies luminance and chrominance;
a mode controller that controls, on a basis of the transform skip flag corresponding to the component identifier obtained by the flag decoder, whether a decoding mode of coded data of coefficient data corresponding to the component identifier is set to a transform skip (TS) residual decoding mode that is a mode for a case of skipping inverse transform processing of transforming coefficient data into a residual between an image and a predicted image, or set to a non-TS residual decoding mode that is a mode for a case of not skipping the inverse transform processing, wherein based on the transform skip flag, data of the residual is decoded according to a syntax structure corresponding to each decoding mode set by the mode controller; and
a coefficient data decoder that decodes the coded data of the coefficient data corresponding to the component identifier according to the decoding mode set by the mode controller, and generates the coefficient data corresponding to the component identifier,
wherein in a case that the TS residual decoding mode is set, when a number of residual context coding bins is determined to be equal to or greater than a threshold, context decoding is applied to decoding of a sign code, and when the number of residual context coding bins is determined to be not equal to or greater than the threshold, bypass decoding is applied to the decoding of the sign code, and
wherein in a case where the inverse transform processing is skipped, the coefficient data decoder shares a context variable between decoding of coded data of a luminance component of the coefficient data and decoding of coded data of a chrominance component.
2. The image processing device according to claim 1 , wherein
the coefficient data decoder applies a sign code decoding method according to the transform skip flag corresponding to the component identifier obtained by the flag decoder.
3. The image processing device according to claim 1 , wherein
the flag decoder further decodes coded data of a transform skip residual encoding use flag that is flag information indicating whether to apply the TS residual decoding mode or the non-TS residual decoding mode, and obtains the transform skip residual encoding use flag corresponding to the component identifier in the case where the inverse transform processing is skipped, and
the mode controller further controls whether a decoding mode of the coded data of the coefficient data corresponding to the component identifier is set to the TS residual decoding mode or the non-TS residual decoding mode on a basis of the transform skip residual encoding use flag corresponding to the component identifier generated by the flag decoder.
4. The image processing device according to claim 1 , wherein
the flag decoder further decodes coded data of a transform skip residual encoding use specifying mode flag that is flag information indicating whether to apply the TS residual decoding mode or the non-TS residual decoding mode in a specific mode, and obtains the transform skip residual encoding use specifying mode flag corresponding to the component identifier, and
the mode controller further controls whether a decoding mode of the coded data of the coefficient data corresponding to the component identifier is set to the TS residual decoding mode or the non-TS residual decoding mode on a basis of the transform skip residual encoding use specifying mode flag corresponding to the component identifier generated by the flag decoder.
5. A coefficient data generation method comprising:
decoding coded data of a transform skip flag corresponding to a component identifier and obtaining the transform skip flag corresponding to the component identifier, wherein the component identifier identifies luminance and chrominance;
controlling, on a basis of the transform skip flag corresponding to the component identifier obtained, whether a decoding mode of coded data of coefficient data corresponding to the component identifier is set to a TS residual decoding mode that is a mode for a case of skipping inverse transform processing of transforming coefficient data into a residual between an image and a predicted image, or set to a non-TS residual decoding mode that is a mode for a case of not skipping the inverse transform processing, wherein based on the transform skip flag, data of the residual is decoded according to a syntax structure corresponding to each decoding mode set by the mode controller; and
decoding the coded data of the coefficient data corresponding to the component identifier according to the decoding mode set, and generating the coefficient data corresponding to the component identifier,
in a case that the TS residual decoding mode is set, when a number of residual context coding bins is determined to be equal to or greater than a threshold, applying context decoding to decoding of a sign code, and when the number of residual context coding bins is determined to be not equal to or greater than the threshold, applying bypass decoding to the decoding of the sign code, and
wherein in a case where the inverse transform processing is skipped, sharing a context variable between decoding of coded data of a luminance component of the coefficient data and decoding of coded data of a chrominance component.
6. An image processing device comprising:
a quantization parameter corrector that, in a case where a transform skip flag corresponding to a component identifier indicates a transform skip for skipping inverse transform processing of transforming coefficient data into a residual between an image and a predicted image, corrects a quantization parameter to be applied to a processing target transform block corresponding to the component identifier, wherein the component identifier identifies luminance and chrominance; and
inverse quantization circuitry that performs inverse quantization of a processing target transform block corresponding to the component identifier by using the quantization parameter corrected by the quantization parameter corrector,
wherein in a case that a transform skip (TS) residual decoding mode that is a mode for a case of the skipping of the inverse transform processing of the transforming coefficient data into the residual between the image and the predicated image is set, when a number of residual context coding bins is determined to be equal to or greater than a threshold, context decoding is applied to decoding of a sign code, and when the number of residual context coding bins is determined to be not equal to or greater than the threshold, bypass decoding is applied to the decoding of the sign code, and
wherein in a case where the inverse transform processing is skipped, a coefficient data decoder shares a context variable between decoding of coded data of a luminance component of the coefficient data and decoding of coded data of a chrominance component.
7. The image processing device according to claim 6 , wherein
the quantization parameter corrector
sets, in a case where the transform skip flag corresponding to the component identifier indicates the transform skip, one having a larger value of a minimum quantization parameter for the transform skip and a quantization parameter corresponding to the component identifier as the quantization parameter to be applied to the processing target transform block corresponding to the component identifier, and
sets, in a case where the transform skip flag corresponding to the component identifier indicates non-transform skip in which the transform skip is not performed, the quantization parameter corresponding to the component identifier as the quantization parameter to be applied to the processing target transform block corresponding to the component identifier.
8. The image processing device according to claim 6 , further comprising
a flag decoder that decodes coded data of the transform skip flag corresponding to the component identifier and obtains the transform skip flag corresponding to the component identifier,
wherein in a case where the transform skip flag corresponding to the component identifier obtained by the flag decoder indicates the transform skip, the quantization parameter corrector corrects the quantization parameter to be applied to the processing target transform block corresponding to the component identifier.
9. A coefficient data generation method comprising:
correcting, in a case where a transform skip flag corresponding to a component identifier indicates a transform skip for skipping inverse transform processing of transforming coefficient data into a residual between an image and a predicted image, a quantization parameter to be applied to a processing target transform block corresponding to the component identifier, wherein the component identifier identifies luminance and chrominance; and
performing inverse quantization of a processing target transform block corresponding to the component identifier by using the quantization parameter corrected, and generates the coefficient data corresponding to the component identifier,
wherein in a case that a transform skip (TS) residual decoding mode that is a mode for a case of the skipping of the inverse transform processing of the transforming coefficient data into the residual between the image and the predicated image is set, when a number of residual context coding bins is determined to be equal to or greater than a threshold, context decoding is applied to decoding of a sign code, and when the number of residual context coding bins is determined to be not equal to or greater than the threshold, bypass decoding is applied to the decoding of the sign code, and
wherein in a case where the inverse transform processing is skipped, sharing a context variable between decoding of coded data of a luminance component of the coefficient data and decoding of coded data of a chrominance component.
10. The image processing device according to claim 1 , wherein
the flag decoder further decodes coded data of a transform skip residual encoding use flag that is flag information indicating whether to apply the TS residual decoding mode or the non-TS residual decoding mode, and obtains the transform skip residual encoding use flag corresponding to the component identifier in the case where the inverse transform processing is skipped.
11. The image processing device according to claim 1 , wherein
the flag decoder further decodes coded data of a transform skip residual encoding use specifying mode flag that is flag information indicating whether to apply the TS residual decoding mode or the non-TS residual decoding mode in a specific mode, and obtains the transform skip residual encoding use specifying mode flag corresponding to the component identifier.