The name of the invention: an image processing apparatus and method
Technical field
[0001]
　The present disclosure relates to an image processing apparatus and method, and particularly relates to an image processing apparatus and method which make it possible to suppress the decrease in coding efficiency.
Background technique
[0002]
　Recently, generated in an image sensor or the like, there is an increasing demand for compression of the RAW data which is the image data before demosaic processing and the like are performed.
[0003]
　The image data encoding method as MPEG-4 Part10 (Advanced Video Coding, hereinafter referred to as AVC) have. In recent years, in order to further improve the encoding efficiency, ITU-T (International Telecommunication Union Telecommunication Standardization Sector) and, ISO / IEC (International Organization for Standardization / International Electrotechnical Commission) is a standards body of the joint JCTVC (Joint Collaboration Team - the Video coding), HEVC (standardization of coding method called High Efficiency Video coding) has been advanced (for example, see non-Patent Document 1).
[0004]
　In such a coding system, a method of performing prediction between components for the residual signal is considered (e.g., see Non-Patent Document 2).
CITATION
Non-Patent Document
[0005]
非特許文献1 : Benjamin Bross, Woo-Jin Han, Jens-Rainer Ohm, Gary J. Sullivan, Ye-Kui Wang, Thomas Wiegand, " High Efficiency Video Coding (HEVC) text specification draft 10 (for FDIS & Last Call)", JCTVC-L1003_version 34, 2013-03-19
非特許文献2 : Wei Pu, Woo-Shik Kim, Jianle Chen, Joel Sole, Marta Karczewicz, "RCE1: Descriptions and Results for Experiments 1, 2, 3, and 4", JCTVC-O0202,2013-11-11
Summary of the invention
Problems that the Invention is to Solve
[0006]
　Incidentally, the bit depth of the data for each component, since generally independent of each other, there is a possibility that different values ​​are set to each other. However, in the method described in Non-Patent Document 2, so prediction of the bit depth between different components have not been assumed, the coding efficiency is likely to be reduced can not be performed prediction correctly .
[0007]
　The present disclosure has been made in view of such circumstances, it is to be able to suppress the decrease in coding efficiency.
Means for Solving the Problems
[0008]
　The image processing apparatus according to an embodiment of the present technology, when performing prediction between the components with respect to residual data between the input image and the predicted image comprising a plurality components, the bit depth of the residual data among the components a residual prediction unit which performs the prediction aligned, the prediction residual data generated by the prediction by the residual prediction unit is an image processing apparatus having an encoding unit for encoding.
[0009]
　The residual prediction portion can align the bit depth of the residual data by a bit shift.
[0010]
　The residual prediction unit, when the difference of the bit depth between the two components to perform the prediction is not 0, performing the prediction aligned between components the bit depth of the residual data by the bit shift can.
[0011]
　The residual prediction unit, when the difference of the bit depth is positive, performs the prediction aligned between components the bit depth of the residual data by the bit shift, if the difference of the bit depth is negative, it can be omitted the prediction.
[0012]
　The residual prediction unit, when the color space of the input image is not RGB space, performs the prediction aligned between components the bit depth of the residual data by the bit shift, the color space of the input image is RGB for spatial, it can be omitted the prediction.
[0013]
　A color space YUV space of the input image, the residual prediction unit, between the luminance component and the chrominance component, is possible to perform the prediction by aligning the bit depth of the residual data by the bit shift it can.
[0014]
　A color space RGB space of the input image, the residual prediction unit, between the G component and the R component or B component, the prediction by aligning the bit depth of the residual data by the bit shift it can be carried out.
[0015]
　The residual prediction unit obtains a difference between the bit depth of between two components performing the prediction, according to a difference of the bit depth with respect to the residual data of one component of said two components It performs the bit shift, the bit-shifted the residual data is multiplied by a predetermined weighting factor, performs a bit shift of the predetermined number of bits for the multiplication result, and the residual data of the other component by obtaining a difference between the multiplication result of the bit shift, it is possible to perform the prediction.
[0016]
　The residual prediction unit may set a common said weighting factors of several components.
[0017]
　According to another embodiment of the present technology, when performing prediction between the components with respect to residual data between the input image and the predicted image comprising a plurality components, the bit depth of the residual data among the components align performs the prediction, the prediction residual data generated by said prediction is an image processing method for encoding.
[0018]
　The image processing device according to another aspect of the present technique, prediction residual data is the prediction results between the components of the residual data of the image comprising a plurality components and their predictive picture to decode the coded data a decoding unit, when performing restoration of the residual data using the prediction residual data to which the encoded data obtained by decoding by the decoder, the bit depth of the residual data among the components an image processing apparatus and a residual restoration unit which performs the restoration aligned.
[0019]
　The residual restoring portion can align the bit depth of the residual data by a bit shift.
[0020]
　Further comprising a receiving unit that receives information on the bit depth, the residual restoring unit, based on information on the bit depth received by said receiving portion, of the bit depth between the two components performing the prediction obtains a difference by performing the bit-shift based on a difference of the bit depth obtained, it is possible to align the bit depth of the residual data.
[0021]
　The residual restoring unit, when the difference of the bit depth obtained is not zero, it is possible to align the bit depth of the residual data by said bit shift.
[0022]
　The residual restoring unit, when the difference of the bit depth obtained is positive, performs the restoration by aligning the bit depth of the residual data by the bit shift, if the difference of the bit depth obtained is negative , it can be omitted the restoration.
[0023]
　Said receiving unit further receives information relating to the color space of the image, the residual restoring unit, based on the information regarding the color space of the image received by said receiving unit, the color space of the image is not RGB space case, performs the restoration aligned between components the bit depth of the residual data by the bit shift, the color space of the image is a case of RGB space, it is possible to omit the restoration.
[0024]
　A color space YUV space of the image, the residual restoring unit may be performed with the luminance component and the color difference component, the restored by aligning the bit depth of the residual data by the bit shift .
[0025]
　Color space of the image is an RGB space, the residual restoration unit, between the G and R components or B components, performs the restoration by aligning the bit depth of the residual data by the bit shift be able to.
[0026]
　The residual restoring unit obtains a difference between the bit depth of between two components performing the restoration, the difference of the bit depth with respect to the residual data Restored of one component of said two components performs the bit shifting in response to the bit-shifted the residual data is multiplied by a predetermined weighting factor, it performs a bit shift of the predetermined number of bits for the multiplication result, which is the bit-shifted the by adding the multiplication result and the prediction residual data, it is possible to perform the restoration of the residual data of the other components.
[0027]
　The image processing method of another aspect of the present technology image and the prediction residual data is the prediction results between the components of the residual data and the predicted image by decoding the encoded data composed of a plurality of components , when performing restoration of the residual data using the prediction residual data obtained by decoding the coded data, performs the restore bit depth of the residual data by aligning between the component images it is a processing method.
[0028]
　In one aspect of the present technology, when performing prediction between components for the residual data between the input image and the predicted image comprising a plurality components, lines predicted bit depth of residual data is aligned between components We, prediction residual data generated by the prediction is coded.
[0029]
　In another aspect of the present technique, the prediction residual data is a prediction result of the inter-residual data of the image comprising a plurality components and the prediction image component is decoded coded data, the coded data There when performing restoration of the residual data using the prediction residual data obtained by the decoding, the bit depth of residual data restoration is performed aligned between components.
Effect of the invention
[0030]
　According to the present disclosure, images can be encoded and decoding. In particular, it is possible to suppress the decrease in coding efficiency.
Brief description of the drawings
[0031]
Is a diagram illustrating a configuration example of FIG. 1-coding unit.
Is a diagram illustrating an example of FIG. 2 semantics.
3 is a block diagram showing a main configuration example of an image encoding apparatus.
It is a block diagram illustrating a main configuration example of FIG. 4 header processing unit and the residual estimator.
5 is a flowchart illustrating an example of the encoding process flow.
6 is a flowchart illustrating an example of a flow of the residual prediction process.
7 is a flowchart illustrating an example of a flow of predictive residual data generation processing.
8 is a flowchart illustrating an example of a flow of predictive residual data generation processing.
9 is a flowchart illustrating an example of a flow of predictive residual data generation processing.
It is a block diagram illustrating a main configuration example of FIG. 10 the image decoding apparatus.
11 is a block diagram illustrating a main configuration example of a header acquisition unit and the residual restoring unit.
12 is a flowchart for explaining an example of the decoding process flow.
13 is a flowchart illustrating an example of a flow of the residual restoration process.
14 is a flowchart for explaining an example of the flow of residual data restoration process.
It is a flowchart illustrating an example of the flow of FIG. 15 residual data restoration process.
16 is a flowchart for explaining an example of the flow of residual data restoration process.
Is a diagram illustrating an example of FIG. 17 semantics.
18 is a block diagram illustrating a main configuration example of a header processing unit and the residual estimator.
19 is a flowchart illustrating an example of a flow of the residual prediction process.
It is a block diagram illustrating a main configuration example of FIG. 20 header acquisition unit and the residual restoring unit.
21 is a flowchart illustrating an example of a flow of residual restoration process.
Is a diagram illustrating an example of FIG. 22 semantics.
23 is a block diagram illustrating a main configuration example of a header processing unit and the residual estimator.
It is a flowchart illustrating an example of the flow of FIG. 24 residual prediction process.
It is a block diagram illustrating a main configuration example of FIG. 25 header acquisition unit and the residual restoring unit.
FIG. 26 is a flowchart illustrating an example of a flow of residual restoration process.
Is a diagram illustrating an example of FIG. 27 semantics.
[FIG. 28] is a block diagram illustrating a main configuration example of a header processing unit and the residual estimator.
FIG. 29 is a flowchart illustrating an example of a flow of the residual prediction process.
It is a block diagram illustrating a main configuration example of FIG. 30 header acquisition unit and the residual restoring unit.
[FIG. 31] is a flowchart illustrating an example of a flow of the residual restoration process.
Is a diagram illustrating an example of FIG. 32 syntax.
Is a diagram illustrating an example of FIG. 33 semantics.
Is a diagram illustrating an example of FIG. 34 semantics.
Is a diagram illustrating an example of FIG. 35 semantics.
Is a diagram illustrating an example of FIG. 36 syntax.
Is a diagram illustrating an example of FIG. 37 semantics.
Is a diagram illustrating an example of FIG. 38 semantics.
Is a diagram illustrating an example of FIG. 39 semantics.
It is a diagram illustrating an example of FIG. 40 multi-view image encoding method.
[FIG. 41] is a diagram showing a main configuration example of the multi-view image encoding apparatus according to the present technique.
[FIG. 42] is a diagram showing a main configuration example of the multi-view image decoding apparatus according to the present technique.
[43] is a diagram showing an example of a hierarchical image coding scheme.
It is a diagram illustrating an example of FIG. 44 Spatial scalable coding.
It is a diagram illustrating an example of FIG. 45-temporal scalable coding.
It is a diagram illustrating an example of a scalable encoding of FIG. 46 signal-to-noise ratio.
[FIG. 47] is a diagram showing a main configuration example of the applied layer image encoding device to which the present technology.
It is a diagram illustrating a main configuration example of FIG. 48 hierarchical image decoding apparatus according to the present technique.
[FIG. 49] is a block diagram illustrating a main configuration example of a computer.
[FIG. 50] is a block diagram illustrating an exemplary configuration of a television set.
[FIG. 51] is a block diagram illustrating an exemplary configuration of a mobile phone.
[FIG. 52] is a block diagram illustrating an exemplary configuration of a recording and reproducing apparatus.
[FIG. 53] is a block diagram illustrating an exemplary configuration of an imaging device.
[FIG. 54] is a block diagram illustrating an exemplary configuration of a video set.
[FIG. 55] is a block diagram illustrating an exemplary configuration of the video processor.
[FIG. 56] is a block diagram showing another example of a schematic configuration of a video processor.
DESCRIPTION OF THE INVENTION
[0032]
　The following describes embodiments of the present disclosure (hereinafter referred to as embodiments). The description will be made in the following order.
　1. First Embodiment (image encoding apparatus, image decoding apparatus)
　2. Second Embodiment (image encoding apparatus, image decoding apparatus)
　3. Third Embodiment (image encoding apparatus, image decoding apparatus)
　4. Fourth Embodiment (image encoding apparatus, image decoding apparatus)
　5. Fifth Embodiment (common weighting factor)
　6. Sixth Embodiment (the multi-view image encoding device and the multi-view image decoding apparatus)
　7. Seventh embodiment (hierarchical image encoding device and the hierarchical image decoding apparatus)
　8. Eighth Embodiment
　(Computer) 9. The ninth embodiment (Application Example)
　10. Tenth embodiment of (set-unit module processor)
[0033]
　<1. First Embodiment>
　　
　In recent years, handling image information as digital, time, transmission of a high efficiency information, storage for the purpose of, by using the image information inherent redundancy , orthogonal transform and apparatus for compressing code an image by employing the encoding method for compressing the motion compensation such as a discrete cosine transform is spreading. The encoding method include, for example, MPEG (Moving Picture Experts Group).
[0034]
　In particular, MPEG2 (ISO / IEC 13818-2) is defined as a general-purpose image encoding scheme, both the interlaced scanning images and progressive scan images as well as standard to cover standard resolution images and high-definition images. For example, MPEG2 is currently widely used in a wide range of application of professional applications and consumer applications. By using the MPEG2 compression system, it is possible to assign, if the standard resolution interlaced scanned picture 4 to the code amount of 8 Mbps (bit rate) with for example, 720x480 pixels. Further, it is possible to assign by using the MPEG2 compression method, for example, the code amount of 18 to 22-Mbps if high-resolution interlaced scan image having 1920x1088 pixels (bit rate). Thus, it is possible to realize a high compression ratio excellent image quality.
[0035]
　MPEG2 had mainly targeted conforming quality coding for broadcasting, but less code amount than MPEG 1 (bit rate) did not correspond to the encoding system of higher clogging compression ratio. With the spread of portable terminals appear to the growing needs of such encoding system future standardization of the MPEG4 encoding system it has been performed correspondingly thereto. For the image encoding method, the standard has been approved by the international standard in December 1998 as ISO / IEC 14496-2.
[0036]
　In addition, in recent years, for the purpose of the initial image encoding for television conference, H.26L (ITU-T (International Telecommunication Union Telecommunication Standardization Sector) Q6 / 16 VCEG (Video Coding Expert Group)) standard of standardization that is advanced It was. H.26L compared to conventional encoding methods such as MPEG2 and MPEG4, the encoded, although many calculation amount required by the decoding, it is known that higher encoding efficiency can be realized. In addition, currently, as part of the MPEG4 of activity, row this H.26L to base, also incorporates features that are not supported by H.26L, standardization to achieve a higher coding efficiency as the Joint Model of Enhanced-Compression Video Coding We were.
[0037]
　The standardization of the schedule, in March 2003, became the international standard in the name of H.264 and MPEG-4 Part10 (Advanced Video Coding, hereinafter referred to as AVC).
[0038]
　In addition, this H. 264 / AVC as an extension of, RGB or 4: 2: 2, 4: 4: 4, such as, and coding tools necessary for business, including also the 8x8DCT and quantization matrix is ​​specified by the MPEG-2 FRExt (Fidelity Range Extension) standardization was completed in February 2005. As a result, H. With 264 / AVC, it becomes capable encoding method to satisfactorily express even the film noise contained in the movie, was the carry to be used for a wide range of applications of such Blu-Ray Disc (trademark).
[0039]
　However, in recent years, four times that of high-definition image, you want to compress the image of about 4000x2000 pixels, or, such as the Internet, in a limited transmission capacity of the environment, but would like to deliver a high-definition image, further high compression rate coding there is a growing need for. Accordingly, the foregoing, in VCEG affiliated ITU-T, Study on improvement of the coding efficiency has been performed is continued.
[0040]
　So, now, for the purpose of further improvement of the coding efficiency than AVC, and the ITU-T, is an ISO / IEC joint standardization organization of the (International Organization for Standardization / International Electrotechnical Commission) JCTVC (Joint Collaboration Team - Video Coding ) by the standardization of coding method called HEVC (High Efficiency Video coding) is underway. For HEVC standard, Committee draft has been issued a draft specification in January 2013 (for example, see Non-Patent Document 1).
[0041]
　　
　In the following, the example of the case of applying the image encoding and decoding of HEVC (High Efficiency Video Coding) scheme, explaining the technology.
[0042]
　　
　In AVC (Advanced Video Coding) scheme, the hierarchical structure of a macroblock and sub-macroblock is defined. However, the macroblock 16x16 pixels, such that the target of the next generation coding method, UHD; not optimal for large picture frame such as (Ultra High Definition 4000 pixels x2000 pixels).
[0043]
　In contrast, in HEVC method, as shown in FIG. 1, the coding unit (CU (Coding Unit)) is defined.
[0044]
　CU, also referred to as Coding Tree Block (CTB), it plays the same role and the macro block in the AVC method, a partial region of the image in units of pictures. The latter, while being fixed to a size of 16x16 pixels, the former size is not fixed, in each sequence will be specified in the image compression information.
[0045]
　For example, in the sequence parameter set to be included in the output coded data (SPS (Sequence Parameter Set)), the minimum size and maximum size of the CU (LCU (Largest Coding Unit)) ((SCU (Smallest Coding Unit)) provisions It is.
[0046]
　Within each LCU, in a range not less than the size of the SCU, by a split-flag = 1, can be divided into smaller sizes CU. In the example of FIG. 1, the magnitude of the LCU 128, a maximum hierarchical depth is 5. The size of the CU of 2Nx2N, when the value of split_flag is "1", and one lower level, is divided into CU of the size of NxN.
[0047]
　Additionally, CU is partitioned into the intra or inter a processing unit and a region of the predicted (partial region of the image in units of pictures) is prediction unit (Prediction Unit (PU)), also serving as a processing unit of the orthogonal transform domain is a (partial area of ​​the image of each picture), is divided into transform unit (transform unit (TU)). Currently, in HEVC scheme, in addition to the 4x4 and 8x8, it is possible to use orthogonal transform 16x16 and 32x32.
[0048]
　As described above HEVC scheme defines a CU, when the coding scheme as performing various processes the CU units, macroblock in AVC format corresponds to LCU, block (sub-block) is equivalent to the CU Then it can be considered. The motion compensation block in the AVC method can be considered to correspond to PU. However, CU, since having a hierarchical structure, the size of the LCU of the highest layer, for example 128x128, as the pixel, it is generally set to be larger than macroblocks AVC method.
[0049]
　Therefore, hereinafter, LCU is intended to include macroblocks in AVC format, CU is intended to include a block (sub-block) in the AVC method. That is, "blocks" used in the following description refers to any partial area in the picture, its size, shape, and characteristics, etc. are not limited. In other words, to "block", for example, TU, PU, ​​SCU, CU, LCU, the sub-block, include macro block or slice or any of the area, (processing units). Of course, other than these partial regions (processing units) are also included. If you need to limit the size and process unit or the like, appropriately described.
[0050]
　In the present specification, CTU (Coding Tree Unit) is a CTB of LCU (maximum number of CU) (Coding Tree Block), and is a unit that contains parameters when processing at the LCU base (level) . In addition, CU that make up the CTU (Coding Unit) is, and the CB (Coding Block), and is a unit that includes the parameters of the time to be processed by the CU-based (level).
[0051]
　　
　Incidentally, in the AVC and HEVC encoding scheme, to achieve a higher coding efficiency, selection of an appropriate prediction mode is important.
[0052]
　Examples of such selection method, JM (Joint Model) and H.264 / MPEG-4 AVC reference software called (published in http://iphome.hhi.de/suehring/tml/index.htm) it can be mentioned a method that is implemented in.
[0053]
　In JM, described below, it is possible to select and the High Complexity Mode, the mode determination two ways of Low Complexity Mode. Both calculate the cost function value for each of the prediction modes Mode, which selects a prediction mode that minimizes the optimum mode for the block to macro block.
[0054]
　The cost function in the High Complexity Mode is shown by the following equation (1).
[0055]
[Formula 1]

[0056]
　Here, omega, the entire set D of candidate modes for encoding the block to macro block, when coding in the prediction mode, which is a difference of energy of the decoded image and the input image. λ is a Lagrange undetermined multiplier given as a function of the quantization parameter. R contained the orthogonal transform coefficients, the total code amount in the case of encoding in the mode.
[0057]
　That is, to do the encoding in the High Complexity Mode, for calculating the parameters D and R, by all candidate modes, once, it is necessary to perform a tentative encoding process requires a higher amount of calculation.
[0058]
　The cost function in the Low Complexity Mode is shown by the following equation (2).
[0059]
[Number 2]

[0060]
　Here, D is, unlike the the High Complexity Mode, the difference between the energy of the input image and the prediction image. QP2Quant (QP) is given as a function of the quantization parameter QP, HeaderBit does not include the orthogonal transformation coefficients, and the motion vectors, such as mode, a code amount relating to information belonging to the Header.
[0061]
　That is, in the Low Complexity Mode, for each candidate mode, it is necessary to perform prediction processing, since it is not necessary to decode the image, need not be performed until the coding process. Therefore, it is possible to realize a lower amount of computation than the High Complexity Mode.
[0062]
　　
　Incidentally, in HEVC, when 444 encoding method for performing predictive (also referred to as residual prediction) between components for the residual signal is considered (e.g., see Non-Patent Document 2) .
[0063]
　In the method described in Non-Patent Document 2, the color difference component (Cb / Cr) (or, R component and B component) for the luminance component (Y-) (or G component) using the following equation (3) prediction is performed so.
[0064]
[Number 3]

[0065]
　Here, r c (x, y) represents the color difference component residual data (Cb or Cr) (difference between the input image and the prediction image). Further, r L (x, y) indicates the residual data of the luminance component (Y-) (difference between the input image and the prediction image). Furthermore, .DELTA.r c (x, y) is the prediction result of the residual prediction (which residual data of the color difference component (Cb or Cr) is predicted by the residual data of the luminance component (Y-)) (prediction residual data for both referred to) show the. Incidentally, shown in these, x, y, the position in the image (coordinate).
[0066]
　Also, alpha denotes a weighting factor, taking a value of either ± (0,1,2,4,8). This value is set by the TU units (i.e., are transmitted to the decoding side in TU units). Further, >> denotes a bit shift to the right (right shift). For example, >> n indicates the right shift of n bits.
[0067]
　Incidentally, whether or not to perform such a residual prediction is controlled by the picture parameter set (PPS (Picture Parameter Set)) units. That, On / Off flag for controlling whether or not to perform such a residual prediction is transmitted to the decoding side in the PPS.
[0068]
　Incidentally, the bit depth of the data for each component, since generally independent of each other, there is a possibility that different values ​​are set to each other. However, as shown in equation (3) described above, the prediction method described in Non-Patent Document 2, it is the premise bit depth of each component of the residual data (luminance component and color difference components) are identical to each other There, the bit depth of residual data has not been envisaged for different cases between components. That is, in the prediction of Equation (3) can not make the correct prediction if the bit depth of the residual data is different between the components, the coding efficiency is likely to decrease.
[0069]
　　
　Therefore, when performing prediction (residual prediction) between components for the residual data between the input image and the predicted image comprising a plurality components, the bit depth of the residual data between components align (scaling) so as to. By doing so, even the bit depth of the residual data is a vary between components, it is possible to make predictions by aligning bit depth of the residual data for each component performs properly residual prediction be able to. Accordingly, it is possible to suppress the decrease in coding efficiency.
[0070]
　A method of aligning the bit depth of the residual data is arbitrary between components, for example, by bit shifting the residual data, it may be to align the bit depth of residual data between components. By doing so, it is possible to align the bit depth of residual data between readily components. That is, it is possible to reduce the load and processing time of the processing.
[0071]
　Further, the color space of the input image in this way to align the bit depth of the residual data is arbitrary. For example, it may be a YUV space (YUV 444, etc.) may be an RGB space (RGB444, etc.). For example, if the input image is a YUV space, between the luminance component and the color difference component, it may be performed predicted by aligning the bit depth of residual data by a bit shift. More specifically, for example, may be performed predicted from aligning the bit depth of the color difference component by bit shifting the bit depth of the luminance component. Also for example, if the input image is a RGB space, between the G component and the R component or the B component may be performed prediction align the bit depth of residual data by a bit shift. More specifically, for example, by bit shifting the bit depth of the G component may be performed predicted from aligning the bit depth of the R component and B component.
[0072]
　　
　a description of a specific example of the shift bit to align the above-described bit depth (shift operation). This shift calculation method is arbitrary. For example, when the difference in bit depth between the two components to make predictions is not 0, it may be performed prediction aligned between components the bit depth of residual data by a bit shift. For example, as shown in the following expression (4), calculates the difference between the residual data bit depth between components (bit depth difference), as in Equation (5) through (8), the minute of the bit depth difference it may be performed only bit shifting. At that time, the bit shift is as shown in equation (6) or expression (8), may be performed by the amount of the absolute value of the bit depth difference. In that case, for example, depending on whether or not the condition equation (5) or expression (7), may be shifted direction is determined by the equation (6) or expression (8).
[0073]
[Number 4]

[0074]
　In the formula (4) through (. 8), BitDepth Y- indicates the bit depth of residual data of the luminance component (Y). BitDepth C indicates the bit depth of the residual data of the color difference component (Cb or Cr). Δbitdepth the bit depth differences between components (residual data bit depth of the luminance component (Y-) (BitDepth Y- ) and color difference components (Cb or Cr) residual data bit depth (BitDepth C the difference between)) show.
[0075]
　Further, r c (x, y) indicates the residual data of the color difference component (Cb or Cr). Further, r L (x, y) indicates the residual data of the luminance component (Y). Furthermore, .DELTA.r c (x, y) represents a prediction residual data between components (which residual data of the color difference component (Cb or Cr) is predicted by the residual data of the luminance component (Y)). Incidentally, shown in these, x, y, the position in the image (coordinate).
[0076]
　In addition, α represents a weighting factor. The value of this α is arbitrary, for example, set to any value of ± (0,1,2,4,8). This value is set by the TU units (i.e., are transmitted to the decoding side in TU units). Further, >> represents a bit shift to the right (right shift), << represents a bit shift to the left (left shift). For example, >> n denotes a right shift of n bits, << m represents a left shift of m bits.
[0077]
　In this example, equation (4) to as shown in equation (8), the equation (4), bit depth differences between components (Derutabitdepth) is calculated. When this value satisfies the condition of formula (5), i.e., if Δbitdepth is 0 or more, the prediction residual data (.DELTA.r c (x, y)) is calculated by the equation (6). As shown in equation (6), the calculation method in this case is basically the same as Equation (3), the residual data of the luminance component (Y-) (r L (x, y)) bits depth difference (Δbitdepth) to the right-shifted prediction residual data (Derutaaru C (X, Y)) is calculated.
[0078]
　In contrast, the bit depth differences between components (Derutabitdepth) is a negative value (less than 0), (satisfies the conditions of formula (7)) When the condition is not satisfied in the formula (5), the predicted residual data (.DELTA.r c (x, y)) is calculated by the equation (8). As shown in equation (8), method of calculating the case, but basically the same as Equation (3), the residual data of the luminance component (Y-) (r L (x, y)) bits depth difference (-Δbitdepth) to the left shifted by prediction residual data (Derutaaru C (X, Y)) is calculated.
[0079]
　That is, in these cases, the residual data (r of the luminance component (Y-) L (x, y)) bit depth (BitDepth Y- is), the residual data of the color difference component (Cb or Cr) (r c (x, bit depth (bitDepth of y)) C prediction residual data aligned to) (.DELTA.r c (x, y)) is calculated. More specifically, the difference in bit depth between the two components to be used in residue prediction is determined, two one component residual bit shifts corresponding to a difference between the bit depth with respect to data of the components ( scaling) is performed, is multiplied by the predetermined weighting factor to the bit shifted residual data, the multiplication result is bit shifted a predetermined number of bits are performed on the residual data and the bit shift of the other component the difference between the multiplication result of the is obtained. It may be performed prediction as described above.
[0080]
　In the case Δbitdepth is 0, the residual data (r of the luminance component (Y-) L (x, y)) the bit shift amount is 0, the prediction residual data (.DELTA.r c (x, y)) is , it can be said is calculated as equation (3).
[0081]
　By calculating as described above, also bit depth of the residual data is a vary between components, the prediction residual data (.DELTA.r c can correctly calculate the (x, y)). Accordingly, it is possible to suppress the decrease in coding efficiency.
[0082]
　　
　for scaling the above-mentioned bit depth, may be the semantics description, such as in the example of FIG. Portion underlined semantics shown in FIG. 2, the above-mentioned scaling (e.g., Equation (5) through (8)) is a description corresponding to.
[0083]
　　
　FIG. 3 is a block diagram showing an example of a configuration of an image encoding apparatus which is one embodiment of an image processing apparatus according to an embodiment of the present technique. The image coding apparatus 100 shown in FIG. 3, for example, encodes image data of a moving picture using a prediction process of predicting process or method analogous thereto, the HEVC. In the following, a case the color space of the input image is a YUV444 for example.
[0084]
　The picture coding apparatus 100 as shown in FIG. 3, a screen rearrangement buffer 102, arithmetic unit 103, orthogonal transform unit 104, a quantization unit 105, a reversible encoding unit 106, storage buffer 107, an inverse quantization unit 108, and an inverse orthogonal transform unit 109. Further, the image encoding apparatus 100 includes operation unit 110, a loop filter 111, a frame memory 112, an intra prediction unit 113, an inter prediction unit 114, a predicted image selecting unit 115 and rate control unit 116. Furthermore, the picture coding apparatus 100 includes a header processing unit 121, residual prediction unit 122, and the residual restoring unit 123.
[0085]
　Screen rearrangement buffer 102, stores the image of each frame of the input image data to the display order, the picture of the stored display order of the frame, in accordance with GOP (Group Of Picture), for encoding rearrange the order of the frame, an image rearranges the order of the frames, and supplies the arithmetic unit 103. Further, the screen rearrangement buffer 102 also supplies an image rearranges the order of the frames, the intra prediction unit 113 and the inter prediction section 114.
[0086]
　Calculation section 103, from the image read from the screen rearrangement buffer 102, subtracts the predicted image supplied from the intra prediction unit 113 or the inter prediction section 114 via the predicted image selection unit 115, the difference information (remaining and it supplies the difference data) to the residual prediction unit 122. For example, if an image intra encoding is performed, the arithmetic unit 103, the image read from the screen rearrangement buffer 102, subtracts the predicted image supplied from the intra prediction unit 113. For example, when an image inter-encoding is performed, the arithmetic unit 103, the image read from the screen rearrangement buffer 102, subtracts the predicted image supplied from the inter prediction unit 114.
[0087]
　Orthogonal transform unit 104, with respect to prediction residual data of the residual data and chrominance component of the luminance component supplied from the residual prediction unit 122 performs orthogonal transform such as discrete cosine transform or Karhunen-Loeve transform. Orthogonal transform unit 104 supplies the transform coefficients obtained by the orthogonal transformation to the quantization unit 105.
[0088]
　Quantization unit 105 quantizes the transform coefficient supplied from the orthogonal transformation unit 104. Quantization unit 105 sets the quantization parameter based on information on the target value of the code amount supplied from the rate control unit 116 performs the quantization. Quantization unit 105 supplies the transform coefficients quantized in the lossless coding unit 106.
[0089]
　Lossless encoding unit 106 encodes the transform coefficient quantized in the quantization unit 105 in any encoding method. Coefficient data because it is quantized under the control of the rate control unit 116, the code amount (approximate to or target value) becomes a target value rate control unit 116 has set.
[0090]
　Further, the reversible encoding unit 106 acquires such information indicating the mode of intra prediction from the intra prediction unit 113, obtains such information and differential motion vector information indicating the mode of inter prediction from the inter prediction unit 114.
[0091]
　Lossless encoding unit 106, these various information coded by any encoding scheme, as part of the header information of the encoded data (also referred to as code stream) (multiplexing). Lossless encoding unit 106, to accumulate and supplies the encoded data obtained by encoding in the storage buffer 107.
[0092]
　The encoding method for lossless encoding unit 106, for example, variable length coding or arithmetic coding or the like. The variable-length coding, for example, H. Such as 264 / AVC system in are defined CAVLC (Context-Adaptive Variable Length Coding), and the like. The arithmetic coding, for example, and the like CABAC (Context-Adaptive Binary Arithmetic Coding).
[0093]
　Storage buffer 107, the coded data supplied from the lossless coding unit 106, and temporarily held. Storage buffer 107 at a predetermined timing, the encoded data stored, and outputs to the outside of the image coding apparatus 100. That is, the storage buffer 107 is also a transmission unit for transmitting the encoded data.
[0094]
　The conversion coefficient quantized in the quantization unit 105 are also supplied to the inverse quantization unit 108. Inverse quantization unit 108, the quantized transform coefficients to inverse quantization in a manner corresponding to the quantization by the quantization unit 105. Inverse quantization unit 108, a transformation coefficient obtained by the inverse quantization, and supplies the inverse orthogonal transform unit 109.
[0095]
　Inverse orthogonal transform unit 109, a transform coefficient supplied from the inverse quantization unit 108 performs inverse orthogonal transformation in a manner corresponding to the orthogonal transform processing performed by the orthogonal transform unit 104. Inverse orthogonal transform unit 109, supplies the inverse orthogonal transform output of the (prediction residual data of the residual data and color difference component of the restored luminance component) to the residual restoring unit 123. The inverse orthogonal transform unit 109 also supplies the residual prediction unit 122 and the residual data of the restored luminance component.
[0096]
　Calculation unit 110, supplied from the residual restoring unit 123, to the restored respective components of the residual data, adds the predicted image from the intra prediction unit 113 or the inter prediction section 114 via the predicted image selecting unit 115 , locally reconstructed image (hereinafter, referred to as reconstructed image) is obtained. Its reconstructed image is supplied to the loop filter 111 or the intra prediction unit 113.
[0097]
　Loop filter 111 includes a deblocking filter or an adaptive loop filter or the like, and appropriate filtering processing on the reconstructed image supplied from the calculating unit 110. For example, the loop filter 111 removes block distortion of the reconstructed image by performing deblocking filtering on the reconstructed image. Further, for example, the loop filter 111, for the deblocking filtering result (reconstructed image removal of block distortion is performed), image quality improvement by performing the loop filter processing by using the Wiener filter (Wiener the Filter) I do.
[0098]
　Incidentally, the loop filter 111 is further on the reconstructed image may be performed any other filtering. Further, the loop filter 111, if necessary, information such as filter coefficients used for the filtering is supplied to the lossless encoding unit 106, it can also be adapted to coding.
[0099]
　The loop filter 111 supplies filter processing result (hereinafter, referred to as decoded image) to the frame memory 112.
[0100]
　Frame memory 112 stores the decoded image which is supplied, at a predetermined timing, and supplies the inter prediction section 114 a decoded image stored as a reference picture.

The scope of the claims
[Claim 1]
　When performing prediction between the components with respect to residual data between the input image and the predicted image comprising a plurality components, the residual prediction unit the bit depth of the residual data performs the prediction aligned between the components ,
　an encoding unit for encoding the prediction residual data generated by the prediction by the residual prediction unit
　image processing apparatus comprising a.
[Claim 2]
　The residual prediction unit aligns the bit depth of the residual data by bit shifting
　the image processing apparatus according to claim 1.
[Claim 3]
　The residual prediction unit, said if the difference of the bit depth of between the two components that make predictions is not 0, perform the prediction aligned between components the bit depth of the residual data by the bit shift
　claim the image processing apparatus according to claim 2.
[Claim 4]
　The residual prediction unit, when the difference of the bit depth is positive, performs the prediction aligned between components the bit depth of the residual data by the bit shift, if the difference of the bit depth is negative, It omitted the predicted
　image processing apparatus according to claim 3.
[Claim 5]
　The residual prediction unit, when the color space of the input image is not RGB space, performs the prediction aligned between components the bit depth of the residual data by the bit shift, the color space of the input image is RGB for spatial omitted the predicted
　image processing apparatus according to claim 2.
[6.]
　A color space YUV space of the input image,
　the residual prediction unit, between the luminance component and color difference components and performs the prediction by aligning the bit depth of the residual data by the bit shift
　claims the apparatus according to 2.
[7.]
　A color space RGB space of the input image,
　the residual prediction unit, between the G component and the R component or B component, the prediction by aligning the bit depth of the residual data by the bit shift performing
　image processing apparatus according to claim 2.
[8.]
　The residual prediction unit obtains a difference between the bit depth of between two components performing the prediction, according to a difference of the bit depth with respect to the residual data of one component of said two components It performs the bit shift, the bit-shifted the residual data is multiplied by a predetermined weighting factor, performs a bit shift of the predetermined number of bits for the multiplication result, and the residual data of the other component by obtaining a difference between the multiplication result of the bit shift, and performs the prediction
　image processing apparatus according to claim 2.
[9.]
　The residual prediction unit sets a common said weighting coefficients by a plurality of component
　image processing apparatus according to claim 8.
[10.]
　When performing prediction on the residual data between the input image and the prediction image made up of multiple components among the components performs the prediction bit depth of the residual data by aligning between the components,
　generated by the prediction encoding the prediction residual data that is
　an image processing method.
[11.]
　An image composed of a plurality components and decoding unit prediction residual data is the prediction results between the components of the residual data and the predicted image to decode the coded data,
　the coded data by said decoding section when someone to restore the residual data using the prediction residual data obtained by decoding, and the residual restoration unit performing the restore bit depth of the residual data by aligning between the component
　and image processing apparatus comprising.
[12.]
　The residual restoring unit aligns the bit depth of the residual data by bit shifting
　the image processing apparatus according to claim 11.
[13.]
　Further comprising a receiving unit that receives information on the bit depth,
　the residual restoring unit, based on information on the bit depth received by said receiving portion, of the bit depth between the two components performing the prediction It obtains a difference by performing the bit-shift based on a difference of the bit depth obtained, align the bit depth of the residual data
　image processing apparatus according to claim 12.
[14.]
　The residual restoring unit, when the difference of the bit depth obtained is not 0, align the bit depth of the residual data by the bit shift
　image processing apparatus according to claim 13.
[15.]
　The residual restoring unit, when the difference of the bit depth obtained is positive, performs the restoration by aligning the bit depth of the residual data by the bit shift, if the difference of the bit depth obtained is negative It omitted the restored
　image processing apparatus according to claim 13.
[16.]
　Said receiving unit further receives information relating to the color space of the image,
　the residual restoring unit, based on the information regarding the color space of the image received by said receiving unit, the color space of the image is not RGB space case, performs the restoration aligned between components the bit depth of the residual data by the bit shift, if the color space of the image is RGB space, omitted the restored
　image processing apparatus according to claim 13 .
[17.]
　The color space of the image is the YUV space,
　the residual restoring unit, between the luminance component and color difference components and performs the restoration by aligning the bit depth of the residual data by the bit shift
　Claim 12 the image processing apparatus according to.
[18.]
　Color space of the image is an RGB space,
　the residual restoration unit, between the G and R components or B components, performs the restoration by aligning the bit depth of the residual data by the bit shift
　the image processing apparatus according to claim 12.
[19.]
　The residual restoring unit obtains a difference between the bit depth of between two components performing the restoration, the difference of the bit depth with respect to the residual data Restored of one component of said two components performs the bit shifting in response to the bit-shifted the residual data is multiplied by a predetermined weighting factor, it performs a bit shift of the predetermined number of bits for the multiplication result, which is the bit-shifted the multiplication result by the adding the prediction residual data, performs the restoring of the residual data of the other component
　image processing apparatus according to claim 12.
[20.]
　Prediction residual data is the prediction results between the components of the residual data of the image comprising a plurality components and their prediction image by decoding the coded data,
　the coded data obtained by decoding the on using the prediction residual data when performing restoration of the residual data, performs the restore bit depth of the residual data by aligning between the component
　image processing method.
It corrected the scope of the claims (Convention Article 19)
[April 13, 2015 (13.04.2015) The International Bureau acceptance]
[1]
If the [corrected] bit depth of the residual data of the color difference component of the image and the residual data bit depth of the luminance component of the image is different, the remaining bit depth and the luminance component of the residual data of the color difference component a setting section to align the bit depth of the differential data,
　according to the bit depth set by the setting unit, and predicts the residual data of the color difference component by using the residual data of the luminance component, the prediction residual data and generating residual prediction unit that,
　a coding unit for encoding the prediction residual data generated by the residual prediction unit
　image processing apparatus comprising a.
[2]
[Corrected] The setting unit, the bit depth of residual data of the luminance component as the bit depth of the residual data of the color difference component and the bit depth of residual data of the luminance component is the same performs bit shifting
　image processing apparatus according to claim 1.
[3]
[Corrected] The setting unit, the bit depth of residual data of the luminance component to match the bit depth of the residual data of the color difference component, the bit shift on the bit depth of the residual data of the luminance component performing
　image processing apparatus according to claim 2.
[4]
[Corrected] The setting unit, when the difference between the residual data bit depth bit depth and the color difference component of the residual data of the luminance component is a positive value, the residual data of the luminance component to match the bit depth bit depth of residual data of the color difference component, it performs bit shift on the bit depth of the residual data of the luminance component
　image processing apparatus according to claim 3.
[5]
[Corrected] The image is 4 the number of pixels between pixel number and the color difference component of the luminance component is equal: 4: 4 format image of the
　image processing apparatus according to claim 4.
[6]
If the [corrected] bit depth of the residual data of the color difference component of the image and the residual data bit depth of the luminance component of the image is different, the remaining bit depth and the luminance component of the residual data of the color difference component align the bit depth of the differential data,
　according to the set bit depth, and predicts the residual data of the color difference component by using the residual data of the luminance component, to generate a prediction residual data,
　the generated predicted coding the residual data
　image processing method.
[7]
And [corrected] decoding unit for decoding the encoded data of prediction residual data predicted residual data of the color difference component of the image using the residual data of the luminance component of the image,
　the residual data of the color difference component in the case where the bit depth of the residual data of the luminance component and the bit depth is different, a setting unit to align the bit depth of residual data bit depth and the luminance component of the residual data of the color difference component,
　the setting unit according to the bit depth set by, restoring the residual data of the luminance component, by using the prediction residual data to which the encoded data obtained by decoding by the decoding unit, the residual data of the color difference component and residual restoration unit for
　an image processing apparatus including a.
[8]
[Corrected] The setting unit, the bit depth of residual data of the luminance component as the bit depth of the residual data of the color difference component and the bit depth of residual data of the luminance component is the same performs bit shifting
　image processing apparatus according to claim 7.
[9]
[Corrected] The setting unit, the bit depth of residual data of the luminance component to match the bit depth of the residual data of the color difference component, the bit shift on the bit depth of the residual data of the luminance component performing
　image processing apparatus according to claim 8.
[10]
[Corrected] The setting unit, when the difference between the residual data bit depth bit depth and the color difference component of the residual data of the luminance component is a positive value, the residual data of the luminance component to match the bit depth bit depth of residual data of the color difference component, it performs bit shift on the bit depth of the residual data of the luminance component
　image processing apparatus according to claim 9.
[11]
[Corrected] The image is 4 the number of pixels between pixel number and the color difference component of the luminance component is equal: 4: 4 format image of the
　image processing apparatus according to claim 10.
[12]
Using the residual data of the luminance component of [corrected] image by decoding the encoded data of prediction residual data predicted residual data of the color difference component of the image,
　a bit depth of residual data of the color difference component It said if the bit depth of residual data of the luminance component different, align the bit depth of residual data bit depth and the luminance component of the residual data of the color difference component,
　according to the set bit depth, the luminance by using the residual data of the components, and the prediction residual data that the encoded data obtained by decoding, to restore the residual data of the color difference component
　image processing method.