JP5844969B2 - Image decoding module, image encoding device, image decoding device, image decoding method, and program - Google Patents

Description

  The present invention relates to an image decoding module, an image encoding device, an image decoding device, an image decoding method, and a program for reconstructing an image signal having a predetermined bit depth.

  The moving image coding algorithms are roughly classified into interframe coding (inter coding) and intraframe coding (intra coding). Interframe coding is an approach for compressing information by utilizing the correlation in the time direction in a moving image. A typical example is inter-frame prediction using motion compensation. On the other hand, intraframe coding is an approach for compressing information using correlation within a single frame. For example, in JPEG (Joint Photographic Experts Group) and MPEG-2 (Moving Picture Experts Group), an approach using a discrete cosine transform (DCT) is used, and in JPEG2000, a discrete wavelet is used. It was taken. Furthermore, H.264 recommended by ITU-T. In H.264 / AVC (Advanced Video Coding), an approach in which spatial direction prediction (see, for example, Non-Patent Document 1) and DCT are combined is taken. In general, the coding efficiency of intra coding is lower than that of inter coding, and further improvement in coding efficiency is required.

In recent years, hierarchical coding has been studied as an intra coding method as a scalable coding method for high bit depth video (see, for example, Non-Patent Document 2). Hereinafter, a signal encoding operation by the hierarchical encoding method will be described.
First, an image signal having a bit depth of N bits is input, and a bit depth conversion process (for example, right shift of bits) is performed to convert the image signal into a low bit depth image of N−Δ bits. This N-Δ bit signal is called a lower layer signal. Next, a lower layer signal is encoded, and a decoded image is generated by performing a decoding process on the encoded signal. Next, an inverse bit depth conversion process (for example, bit shift to the left) is performed on the decoded image to generate an N-bit image signal. This N-bit signal is called an upper layer prediction signal. Next, the difference signal between the upper layer prediction signal and the input image is encoded. This difference signal is called an upper layer residual signal. Then, an encoded stream of the lower layer signal and the upper layer residual signal is output.

Reconstruct the original image signal by decoding the lower layer signal included in this encoded stream, performing inverse bit depth conversion processing, and then adding the upper layer residual signal included in the encoded stream Can do.
As described above, the hierarchical coding scheme is a technique corresponding to the bit-depth scalable coding.

Kakuno Shinya et al., “Revised Third Edition H.264 / AVC Textbook”, Impress R & D, Inc., January 1, 2009, p. 131 M.M. Winken, D.W. Marpe, H.M. Schwarz, and T.W. Wiegand, “Bit Depth Scalable Video Coding”, IEEE Intl. Conf. on Images Processing, 2007

  However, in the conventional hierarchical coding scheme, when the inverse bit depth conversion process is performed on the decoded lower layer signal, the bit is left-shifted and the lower bit is padded with zeros. Therefore, there is a problem that the difference between the upper layer prediction signal and the input image becomes large, and the data amount of the upper layer residual signal increases.

The present invention has been made to solve the above problem, and is an image decoding module for reconstructing an image signal having a first bit depth which is a predetermined bit depth, and is lower than the first bit depth. A code input unit that inputs an encoded image signal having a second bit depth that is a bit depth, an image signal decoding unit that decodes the encoded image signal input by the code input unit, and the image signal decoding A bit depth extension unit that extends the bit depth of the image signal to the first bit depth using each bit of the image signal decoded by the unit, and an image signal in which the bit depth extension unit extends the bit depth, And a reconstructing unit that reconstructs an image signal by adding a predetermined bit value having the same number of bits as the difference between the first bit depth and the second bit depth.

  In the present invention, the image signal decoding unit performs a real number operation on the image signal input by the code input unit, outputs an integer part of the calculation result as a decoded image signal, and the reconstruction unit It is preferable that the predetermined bit value to be added is a fractional part of the calculation result of the real number calculation by the image signal decoding unit.

  In the present invention, it is preferable that the predetermined bit value added by the reconstruction unit is a value that minimizes an approximation error with respect to the lower bits of the original image signal.

  In the present invention, the second bit depth value that minimizes an error caused by approximation to the first bit depth value is stored in association with the first bit depth value. A bit depth expansion unit, wherein each bit of the image signal decoded by the image signal decoding unit has a first bit depth associated with a second bit depth value stored in the table. It is preferable to extend the bit depth of the image signal to the first bit depth by replacing with a value.

  In the present invention, the second bit depth is set for each color channel, and the image signal decoding unit, the bit depth extending unit, and the reconstruction unit execute a decoding process for each color channel. It is preferable to do.

The present invention is also an image coding apparatus including the image decoding module, wherein the image signal input unit inputs the image signal having the first bit depth, and the bits of the image signal input by the image signal input unit. a bit-depth reduction unit for reducing the depth to the second bit depth is lower bit depth than the first bit depth, image signals the bit depth reduction unit encodes the image signal obtained by reducing the bit depth, and the coded Is output to the code input unit of the image decoding module, the image signal reconstructed by the reconstruction unit of the image decoding module, and the difference between the image signal input by the image signal input unit A difference generation unit that generates a difference signal.

  Further, in the present invention, the magnitude of the error of the image signal reconstructed by the reconstruction unit of the image decoding module with respect to the image signal input by the image signal input unit and the image signal encoded by the image signal encoding unit A first cost calculation unit that calculates a first cost by weighted addition with a data amount, an image signal encoded by the image signal encoding unit, and a data amount of a difference signal generated by the difference generation unit. A second cost calculating unit that calculates a cost of 2, and when the first cost is lower than the second cost, the image signal decoded by the image signal decoding unit is output as an encoded stream; When the cost of the image is higher than the second cost, the image signal decoded by the image signal decoding unit and the difference signal generated by the difference generation unit are output as an encoded stream. It is preferable and a stream output portion.

  The present invention is also an image decoding apparatus comprising the image decoding module, wherein the second bit depth encoded image signal, and an image signal and an original image signal that can be reconstructed using the image decoding module An encoded stream input unit that inputs an encoded stream including a difference signal indicating a difference between the encoded stream and an encoded signal output unit that outputs an image signal included in the encoded stream to a code input unit of the image decoding module; It is characterized by providing.

  Moreover, in this invention, it is preferable to provide the difference reflection part which adds the difference data contained in the said encoding stream to the image signal reconstructed by the reconstruction part of the said image decoding module.

The present invention is also an image decoding method using an image decoding module for reconstructing an image signal having a first bit depth that is a predetermined bit depth, wherein the code input unit is lower than the first bit depth. The encoded image signal of the second bit depth which is a bit depth is input, the image signal decoding unit decodes the image signal input by the code input unit, and the bit depth extension unit is the image signal decoding unit Is used to extend the bit depth of the image signal to the first bit depth, and the reconstruction unit adds the bit depth of the image signal to the image signal. The image signal is reconstructed by adding a predetermined bit value having the same number of bits as the difference between the first bit depth and the second bit depth.

  In the present invention, the image signal decoding unit performs a real number operation on the image signal input by the code input unit, outputs an integer part of the calculation result as a decoded image signal, and the reconstruction unit It is preferable that the predetermined bit value to be added is a fractional part of the calculation result of the real number calculation by the image signal decoding unit.

  In the present invention, it is preferable that the predetermined bit value added by the reconstruction unit is a value that minimizes an approximation error with respect to the lower bits of the original image signal.

  In the present invention, the second bit depth value that minimizes an error caused by approximation to the first bit depth value is stored in association with the first bit depth value. A bit depth expansion unit, wherein each bit of the image signal decoded by the image signal decoding unit has a first bit depth associated with a second bit depth value stored in the table. It is preferable to extend the bit depth of the image signal to the first bit depth by replacing with a value.

  In the present invention, the second bit depth is set for each color channel, and the image signal decoding unit, the bit depth extending unit, and the reconstruction unit execute a decoding process for each color channel. It is preferable to do.

The present invention also provides an image decoding module for reconstructing an image signal having a first bit depth that is a predetermined bit depth, encoded with a second bit depth that is lower than the first bit depth. A bit depth of the image signal using each bit of the image signal decoded by the image signal decoding unit and the image signal decoding unit decoding the image signal input by the code input unit A bit depth extension unit that extends the bit depth to the first bit depth, and the number of bits equal to the difference between the first bit depth and the second bit depth in the image signal in which the bit depth extension unit has extended the bit depth is a program for functioning as a reconstruction unit for reconstructing an image signal by adding a predetermined bit value having.

  According to the present invention, after extending the bit depth of the image signal decoded by the image signal decoding unit to the first bit depth, a predetermined bit value is added to the lower bit to generate an upper layer prediction signal. . As a result, the difference between the original image signal and the upper layer prediction signal is reduced, so that the data amount of the upper layer residual signal can be reduced.

It is a schematic block diagram which shows the structure of an encoding apparatus provided with the decoding module by one Embodiment of this invention. It is a schematic block diagram which shows the structure of the decoding module by one Embodiment of this invention. It is a flowchart which shows the encoding operation | movement of the moving image by an encoding apparatus. It is a flowchart which shows the reconstruction operation | movement of the upper layer prediction signal by a decoding module. It is a schematic block diagram which shows the structure of a decoding apparatus provided with the decoding module by one Embodiment of this invention. It is a flowchart which shows the decoding operation | movement of the moving image by a decoding apparatus. It is a schematic block diagram which shows the structure of an image signal decoding part. It is a flowchart which shows the operation | movement of the decoding process of the lower hierarchy signal by this embodiment. It is a flowchart which shows operation | movement of the encoding apparatus in a 1st modification. It is a flowchart which shows operation | movement of the decoding apparatus in a 1st modification. It is a schematic block diagram which shows the structure of the image signal decoding part in a 1st modification. It is a flowchart which shows operation | movement of the image signal decoding part in a 1st modification. It is a schematic block diagram which shows the structure of the image signal decoding part in a 2nd modification. It is a flowchart which shows operation | movement of the image signal decoding part in a 2nd modification. It is a schematic block diagram which shows the structure of the image signal decoding part in a 3rd modification. It is a flowchart which shows operation | movement of the image signal decoding part in a 3rd modification.

Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings.
First, an encoding device (image encoding device) including a decoding module (image decoding module) according to an embodiment of the present invention will be described.
FIG. 1 is a schematic block diagram illustrating a configuration of an encoding device 10 including a decoding module 50 according to an embodiment of the present invention.
The encoding apparatus 10 is an apparatus that encodes an image signal by the hierarchical encoding method according to the present invention, and includes an image signal input unit 11, a bit depth conversion amount setting unit 12, a bit depth reduction unit 13, and an image signal encoding unit 14. , A decoding module 50, a difference generation unit 15, a first cost calculation unit 16, a second cost calculation unit 17, a bit depth conversion prediction determination unit 18, and an encoded stream output unit 19. Here, a “module” is a subset of a device and a group of parts that realize a certain function. The encoding apparatus 10 executes various intra-encoding schemes and adds an encoding system that performs encoding using an optimal encoding scheme for each intra-encoded block to the hierarchical encoding scheme. It is assumed that it will be provided as a module. As an example of the intra coding method, H.264 is used. Intra prediction specified in H.264 can be mentioned.

The image signal input unit 11 inputs an N-bit (first bit depth) image signal to be encoded from the outside.
The bit depth conversion amount setting unit 12 sets a bit depth conversion amount Δ indicating the number of bits for reducing the bit depth of the image signal input by the image signal input unit 11.
The bit depth reduction unit 13 reduces the bit depth of the image signal input by the image signal input unit 11 by the bit depth conversion amount Δ, and generates a lower layer signal of N−Δ bits (second bit depth). Specifically, the bit depth of the image signal input by the image signal input unit 11 is shifted to the right by the bit depth conversion amount Δ.
The image signal encoding unit 14 encodes the lower layer signal generated by the bit depth reduction unit 13.

The decoding module 50 is a module that decodes an N−Δ-bit lower layer signal and generates an upper layer prediction signal that is an N-bit image signal. Note that the upper layer prediction signal is a signal decoded from the lower layer signal, and does not necessarily match the N-bit image signal input by the image signal input unit 11.
The decoding module 50 provided in the encoding device 10 reconstructs an upper layer prediction signal from the lower layer signal encoded by the image signal encoding unit 14.
The difference generation unit 15 generates an upper layer residual signal (difference signal) indicating a difference between the upper layer prediction signal reconstructed by the decoding module 50 and the image signal input by the image signal input unit 11.

The first cost calculation unit 16 includes the noise of the upper layer prediction signal generated by the decoding module 50 for the image signal input by the image signal input unit 11 and the data amount of the lower layer signal encoded by the image signal encoding unit 14. RD (Rate-Distortion) cost is calculated.
The second cost calculation unit 17 calculates the RD cost using the lower layer signal encoded by the image signal encoding unit 14 and the data amount of the upper layer residual signal generated by the difference generation unit 15.

  The bit depth conversion prediction determination unit 18 compares the RD costs calculated by the first cost calculation unit 16 and the second cost calculation unit 17 and indicates whether or not to perform decoding using an upper layer residual signal It is determined whether the depth prediction error skip flag is set to ON or OFF. When the bit depth prediction error skip flag included in the encoded stream is ON, the upper layer residual signal is not used when decoding the encoded stream. On the other hand, when the bit depth prediction error skip flag is OFF, the upper layer residual signal is used when decoding the encoded stream.

  The encoded stream output unit 19 generates the lower layer signal encoded by the image signal encoding unit 14, the bit depth conversion amount Δ set by the bit depth conversion amount setting unit 12, and the bit depth conversion prediction determination unit 18. An encoded stream including at least a bit depth prediction error skip flag is generated. When the bit depth prediction error skip flag is ON, the encoded stream output unit 19 includes the higher layer residual signal generated by the difference generation unit 15 in the encoded stream, and when the bit depth prediction error skip flag is OFF. The upper layer residual signal is not included in the encoded stream.

Next, the configuration of the decoding module 50 included in the encoding device 10 will be described.
FIG. 2 is a schematic block diagram showing the configuration of the decoding module 50 according to an embodiment of the present invention.
The decoding module 50 includes a code input unit 51, an image signal decoding unit 52, a bit depth extension unit 53, and a reconstruction unit 54.
The code input unit 51 inputs an encoded image signal to be decoded.
The image signal decoding unit 52 decodes the encoded image signal input by the code input unit 51. Note that in this embodiment, the image signal decoding unit 52 performs decoding processing of the image signal, and includes processing that requires real number operations such as inverse quantization processing and in-loop filter processing, and is obtained by such processing. An integer part (N-Δ bit signal) of the image signal and a decimal part (higher Δ bit signal) of the image signal are output. Examples of the in-loop filter include a deblocking filter for reducing block distortion due to block division of an image signal and a Wiener filter for minimizing the sum of square errors of a difference signal between a decoded signal and an original signal. The output signal of the in-loop filter is used to generate a reference signal for inter-frame prediction, and is located inside the encoding loop, so it is called an in-loop filter.
The bit depth extension unit 53 receives the bit depth conversion amount Δ from the outside, and extends the bit depth of the image signal decoded by the image signal decoding unit 52 by the bit depth conversion amount Δ. Specifically, the integer part (N−Δ bits) of the image signal and the upper Δ bits of the decimal part are shifted to the left by the bit depth conversion amount Δ.
The reconstructing unit 54 shifts the integer part (N−Δ bits) of the image signal output by the bit depth extending unit 53 to the left by the bit depth conversion amount Δ and the image signal. Are added to an integer value signal (Δ bits) obtained by shifting the upper Δ bits of the decimal part to the left by the bit depth conversion amount Δ, thereby generating an upper layer prediction signal.

Next, a moving image encoding operation by the encoding device 10 will be described.
When an input system reads an input signal, the encoding system including the encoding apparatus 10 performs encoding using various intra encoding schemes for each intra encoding block, and outputs codes based on the optimal intra encoding scheme as encoded streams. To do. The present encoding device 10 performs a hierarchical encoding method among the intra encoding methods.

FIG. 3 is a flowchart showing a moving image encoding operation performed by the encoding apparatus 10.
When the encoding system outputs an intra-encoded block (N-bit image signal) to a module that executes each intra-encoding method, the image signal input unit 11 of the encoding device 10 according to the present embodiment receives the N from the encoding system. A bit image signal is input (step S101). Next, the bit depth reduction unit 13 reads the bit depth N from the image signal input by the image signal input unit 11 (step S102).

Next, the bit depth conversion amount setting unit 12 sets the bit depth conversion amount Δ used for reducing the bit depth of the image signal input by the image signal input unit 11 (step S103). The bit depth conversion amount setting unit 12 may set the bit depth conversion amount Δ by predicting the optimum bit depth conversion amount Δ from the image signal, or may set a predetermined value as the bit depth conversion amount Δ. May be set as
Next, the bit depth reduction unit 13 right-shifts each bit of the image signal input by the image signal input unit 11 by the bit depth conversion amount Δ, thereby generating an N−Δ bit lower layer signal ( Step S104).

  Next, the image signal encoding unit 14 generates encoded data of the lower layer signal by performing an encoding process on the lower layer signal generated by the bit depth reduction unit 13 (step S105). Next, the decoding module 50 decodes the lower layer signal and reconstructs an N-bit upper layer prediction signal (step S106).

Below, the reconstruction operation | movement of the upper layer prediction signal by the decoding module 50 is demonstrated.
FIG. 4 is a flowchart showing the reconstruction operation of the upper layer prediction signal by the decoding module 50.
The code input unit 51 of the decoding module 50 inputs the lower layer signal encoded from the image signal encoding unit 14 (step S501). Next, the image signal decoding unit 52 decodes the lower layer signal input by the code input unit 51 by real number calculation in the in-loop filter process (step S502). At this time, the image signal decoding unit 52 outputs the integer part (N−Δ bits) of the calculation result obtained by the real number calculation and the upper Δ bits of the decimal part to the bit depth extension unit 53.

Next, the bit depth extension unit 53 reads the bit depth conversion amount Δ from the bit depth conversion amount setting unit 12, and the integer part (N−Δ bits) and the upper part of the decimal part of the image signal input from the image signal decoding unit 52. The Δ bits are shifted to the left by the bit depth conversion amount Δ (step S503). Next, the reconstruction unit 54 uses an integer value signal (N bits) obtained by shifting the integer part (N−Δ bits) of the image signal output by the bit depth extension unit 53 to the left by the bit depth conversion amount Δ. An upper layer prediction signal (N bits) is reconstructed by adding an integer value signal (Δ bits) obtained by shifting the upper Δ bits of the decimal part of the image signal to the left by the bit depth conversion amount Δ (step) S504).
In this way, the image signal decoding unit 52 of the decoding module 50 decodes the real number part and the decimal part that have been decoded by rounding the decimal part of the real number calculation result in the conventional in-loop filter processing. Output and add it. That is, according to the present embodiment, decoding is performed in accordance with the accuracy of the real number calculation in the in-loop filter process in accordance with the bit depth of the upper layer. Thereby, it is possible to generate an upper layer prediction signal that is closer to the input signal than the conventional decoding process.

  When the decoding module 50 reconstructs the upper layer prediction signal (step S106), the difference generation unit 15 calculates the difference between the upper layer prediction signal reconstructed by the decoding module 50 and the image signal input by the input unit, An upper layer residual signal that is a difference signal is generated (step S107). Next, the difference generation unit 15 encodes the higher layer residual signal (step S108).

Next, the first cost calculation unit 16 calculates a first cost that is an RD cost when the upper layer residual signal is not used (step S109). Specifically, the first cost calculation unit 16 first sums the squares of the difference values between the image signal input by the image signal input unit 11 and the upper layer prediction signal reconstructed by the reconstruction unit 54 of the decoding module 50. by calculating the calculated size D ₁ of the error. Next, the first cost calculation unit 16 reads the data amount R ₁ of the lower layer signal encoded by the image signal encoding unit 14. Then, using equation (1) below, to calculate the first cost C _1.
C ₁ = D ₁ + λ · R ₁ (1)
Where λ is a Lagrange multiplier.

Next, the second cost calculation unit 17 calculates a second cost that is an RD cost when the higher-layer residual signal is used (step S110). Specifically, the second cost calculation unit 17 includes the data amount R _2a of the lower layer signal encoded by the image signal encoding unit 14 and the data amount R _{2b of} the upper layer residual signal generated by the difference generation unit 15. Is read. Then, using equation (2) below, to calculate the second cost C _2.
C ₂ = D ₂ + λ · (R _2a + R _2b ) (2)
However, D ₂ is to the image signal by the image signal input unit 11 is input, is a value indicating the magnitude of the sum of the squares of the difference signal reconstructed decoded signal using the upper layer prediction signal and the upper layer residual signal .

  Next, the bit depth conversion prediction determination unit 18 determines whether or not the first cost is less than the second cost (step S111). When the first cost is less than the second cost (step S111: YES), the bit depth conversion prediction determination unit 18 indicates that the RD cost is lower when the higher layer residual signal is not used. The prediction error skip flag is turned on (step S112). The encoded stream output unit 19 generates the lower layer signal generated by the image signal encoding unit 14, the bit depth conversion amount Δ set by the bit depth conversion amount setting unit 12, and the bit depth conversion prediction determination unit 18. The bit depth prediction error skip flag is output as an encoded stream (step S113).

  On the other hand, when the first cost is equal to or higher than the second cost (step S111: NO), the RD cost is lower when the higher layer residual signal is used. The depth prediction error skip flag is turned off (step S114). Then, the encoded stream output unit 19 includes a lower layer signal generated by the image signal encoding unit 14, an upper layer residual signal generated by the difference generation unit 15, and a bit depth conversion amount set by the bit depth conversion amount setting unit 12. Δ and the bit depth prediction error skip flag generated by the bit depth conversion prediction determination unit 18 are output as an encoded stream (step S115).

  As a result, the encoding apparatus 10 can output an encoded stream that uses an upper layer residual signal and an encoded stream that does not use the lower RD cost. Further, according to the present embodiment, since the decoding module 50 generates an upper layer prediction signal that is closer to the input signal than in the conventional decoding process, the upper layer residual signal generated by the difference generation unit 15 The data amount is smaller than the difference signal.

Next, a decoding device (image decoding device) including the decoding module 50 according to an embodiment of the present invention will be described.
FIG. 5 is a schematic block diagram illustrating a configuration of the decoding device 20 including the decoding module 50 according to an embodiment of the present invention.
The decoding device 20 is a device that decodes the image signal encoded by the encoding device 10, and includes an encoded stream input unit 21, a bit depth conversion amount decoding unit 22, a code signal output unit 23, a decoding module 50, a bit depth. A prediction error skip flag decoding unit 24, a difference decoding unit 25, and a difference reflection unit 26 are provided. Note that it is assumed that the present decoding device 20 is provided as a module for a hierarchical coding system in a decoding system that decodes an encoded stream including intra-coded blocks encoded by various intra-coding systems. .

The encoded stream input unit 21 inputs an encoded stream to be decoded from the outside.
The bit depth conversion amount decoding unit 22 decodes the bit depth conversion amount Δ included in the encoded stream.
The code signal output unit 23 extracts the lower layer signal included in the encoded stream input unit 21 and outputs it to the decoding module 50.
The decoding module 50 receives the lower layer signal from the code signal output unit 23 and generates an upper layer prediction signal. Note that the decoding module 50 included in the decoding device 20 has the same function as the decoding module 50 included in the encoding device 10.
The bit depth prediction error skip flag decoding unit 24 decodes the bit depth prediction error skip flag included in the encoded stream input unit 21.
The difference decoding unit 25 decodes the upper layer residual signal included in the encoded stream input unit 21 when the bit depth prediction error skip flag decoded by the bit depth prediction error skip flag decoding unit 24 indicates OFF.
When the bit depth prediction error skip flag decoded by the bit depth prediction error skip flag decoding unit 24 indicates OFF, the difference reflection unit 26 and the higher layer prediction signal decoded by the decoding module 50 and the upper layer decoded by the difference decoding unit 25 A decoded signal is generated by adding the hierarchical residual signal. When the bit depth prediction error skip flag decoded by the bit depth prediction error skip flag decoding unit 24 indicates ON, the difference reflection unit 26 outputs the upper layer prediction signal decoded by the decoding module 50 as it is as a decoded signal.

Next, a moving image decoding operation by the decoding device 20 will be described.
When the decoding system including the present decoding device 20 reads the encoded stream, the decoding system 20 identifies the intra encoding scheme and outputs the encoded stream to a module that performs decoding of the intra encoding scheme. The present decoding device 20 performs decoding using a hierarchical coding method among intra coding methods.

FIG. 6 is a flowchart showing a moving image decoding operation performed by the decoding device 20.
When the decoding system outputs the encoded stream to the decoding device 20, the encoded stream input unit 21 of the decoding device 20 according to the present embodiment inputs the encoded stream from the decoding system (step S201). Next, the bit depth conversion amount decoding unit 22 decodes the bit depth conversion amount Δ included in the encoded stream input by the encoded stream input unit 21 (step S202).

  Next, the code signal output unit 23 extracts a lower layer signal from the encoded stream input by the encoded stream input unit 21, and outputs the lower layer signal to the decoding module 50 (step S203). Next, the decoding module 50 reproduces the upper layer prediction signal using the lower layer signal input from the code signal output unit 23 and the bit depth conversion amount Δ decoded by the bit depth conversion amount decoding unit 22 (step S204). . In addition, the reproduction | regeneration operation | movement of the upper hierarchy prediction signal by the decoding module 50 is the same as the operation | movement of step S501-S504 (FIG. 4) mentioned above.

  Next, the bit depth prediction error skip flag decoding unit 24 decodes the bit depth prediction error skip flag included in the encoded stream (step S205). Next, the bit depth prediction error skip flag decoding unit 24 determines whether or not the bit depth prediction error skip flag indicates ON (step S206).

When the bit depth prediction error skip flag indicates ON (step S206: YES), the encoded stream does not include an upper layer residual signal, and instructs decoding without using the upper layer residual signal. Therefore, the difference reflection unit 26 outputs the upper layer predicted signal decoded by the decoding module 50 as it is as a decoded signal (step S207).
On the other hand, when the bit depth prediction error skip flag indicates OFF (step S206: NO), the differential decoding unit 25 decodes the upper layer residual signal included in the encoded stream (step S208). Next, the difference reflection unit 26 adds the upper layer residual signal decoded by the difference decoding unit 25 to the upper layer prediction signal decoded by the decoding module 50, and outputs the signal as a decoded signal (step S209).
Thereby, the decoding device 20 can decode the image signal encoded by the encoding device 10.

Below, the process of the image signal decoding part 52 with which the decoding module 50 by this embodiment is provided is demonstrated.
Next, the configuration of the image signal decoding unit 52 included in the decoding module 50 will be described.
FIG. 7 is a schematic block diagram showing the configuration of the image signal decoding unit 52.
The image signal decoding unit 52 includes an entropy decoding processing unit 521, an inverse transform / inverse quantization processing unit 522, an intra prediction processing unit 523, a frame storage unit 524, a motion vector decoding unit 525, an inter prediction processing unit 526, an addition unit 527, An in-loop filter processing unit 528 is provided.

The entropy decoding processing unit 521 performs entropy decoding processing on the encoded image signal to generate a quantization index. Also, information indicating the prediction method is decoded from the encoded image signal.
The inverse transform / inverse quantization processing unit 522 performs inverse quantization / inverse transform processing on the quantization index generated by the entropy decoding processing unit 521 to generate a prediction error. Examples of the inverse transform process include discrete cosine transform.
When the information indicating the prediction method decoded by the entropy decoding processing unit 521 indicates an intra prediction process, the intra prediction processing unit 523 decodes prediction mode information necessary for intra prediction from the encoded image information. Prediction is performed and a prediction signal is generated.

The frame storage unit 524 stores image signals of past frames.
The motion vector decoding unit 525 decodes motion vector information from the encoded image signal. Here, when using a plurality of reference frames, the motion vector information includes an index for designating the reference frame. In addition, when referring to the decimal pixel position and the interpolation filter coefficient necessary for generating the interpolation image is transmitted, the motion vector information includes information on such a filter.
When the information indicating the prediction method decoded by the entropy decoding processing unit 521 indicates inter prediction processing, the inter prediction processing unit 526 performs inter prediction using the decoded motion vector information, and generates a prediction signal.
The adding unit 527 generates a decoded signal by adding the prediction signal generated by the intra prediction processing unit 523 or the inter prediction processing unit 526 and the prediction error generated by the inverse transform / inverse quantization processing unit 522.
The in-loop filter processing unit 528 performs a deblocking process on the decoded signal or a coding distortion (square error sum of difference signal between the decoded signal and the original signal) reduction process based on a Wiener filter on the decoded signal. An integer part signal (N-Δ bit signal) and a decimal part signal (Δ bit signal) are output from the decoded signal. Further, the signal of the integer part is recorded in the frame storage unit 524.

Next, the detailed operation of the above-described decoding process of the lower layer signal (step S502) will be described.
FIG. 8 is a flowchart showing the operation of the lower layer signal decoding process according to this embodiment.
First, the entropy decoding processing unit 521 reads encoded data of a lower layer signal (step S601). When the encoded data of the lower layer signal is read, the entropy decoding processing unit 521 performs the following processes in steps S603 to S612 for each rectangular area (referred to as an encoding unit unit) serving as a unit of intra prediction processing and inter prediction processing. Execute (step S602).

  The entropy decoding processing unit 521 performs entropy decoding processing on the encoded data of the lower layer signal to generate a quantization index (step S603). Next, the inverse transform / inverse quantization processing unit 522 performs inverse quantization / inverse transform processing on the quantization index generated by the entropy decoding processing unit 521 to generate a prediction residual signal (step S604). Next, the entropy decoding processing unit 521 decodes information indicating the prediction method, and determines whether intra prediction processing or inter prediction processing is specified as the prediction method (step S605).

  When the intra prediction process is designated as the prediction method (step S605: YES), the intra prediction processing unit 523 decodes the prediction mode information necessary for the intra prediction, performs the intra prediction, and generates a prediction signal ( Step S606). On the other hand, when the inter prediction process is designated as the prediction method (step S605: NO), the motion vector decoding unit 525 inputs the encoded data of the motion vector information included in the encoded data of the lower layer signal and performs entropy. Decoding processing is performed to generate motion vector information (step S607). Next, the inter prediction processing unit 526 performs inter prediction using the motion vector information decoded by the motion vector decoding unit 525 and the past frame information stored in the frame storage unit 524, and generates a prediction signal (step S608). ).

  When the prediction signal is generated in step S606 or step S608, the adding unit 527 generates the prediction signal generated by the intra prediction processing unit 523 or the inter prediction processing unit 526 and the prediction residual signal generated by the inverse transform / inverse quantization processing unit 522. And a decoded signal is generated (step S609). Next, the in-loop filter processing unit 528 performs deblocking processing on the decoded signal and outputs the processed decoded signal as a real value (step S610). Next, the in-loop filter processing unit 528 outputs an integer part signal and a decimal part signal from the decoded signal after processing (step S611). Further, the in-loop filter processing unit 528 records the integer part signal in the frame storage unit 524 (step S612).

  When the processes in steps S603 to S612 are performed on all the coding unit units (step S613), the image signal decoding unit 52 ends the decoding process.

  According to the decoding module 50 according to the present embodiment, an integer value signal (N bits) obtained by shifting the integer part (N−Δ bits) of the image signal decoded by the image signal decoding unit 52 to the left by the bit depth conversion amount, An upper layer prediction signal is generated by adding an integer value signal (Δ bits) obtained by shifting the upper Δ bits of the decimal part of the image signal to the left by the bit depth conversion amount Δ. As a result, the difference between the original image signal and the upper layer prediction signal is reduced, so that the data amount of the upper layer residual signal by the processing of the encoding device 10 can be reduced.

As described above, the embodiment of the present invention has been described in detail with reference to the drawings. However, the specific configuration is not limited to the above, and various design changes and the like can be made without departing from the scope of the present invention. It is possible to
For example, in this embodiment, the case where the bit shift method is used as the bit depth conversion method has been described. However, the present invention is not limited to this, and a rounding method, a table reference method, or the like may be used.

The rounding off method is a bit depth conversion method shown below.
In the bit depth conversion, 1 is added to the lower Δth bit of the input signal, and each bit in the same register is right shifted by Δ bits. On the other hand, at the time of inverse bit depth conversion, the left bit is shifted by Δ bits and lower bits are filled with 0.

The table reference method is a bit depth conversion method described below.
The conversion process from an N-bit image signal to an N-Δ bit low bit depth image signal is performed by converting 2 ^N signal values x (x = 0,..., 2 ^N −1) into 2 ^N-Δ signals. This is a process of associating with the value y (y = 0,..., 2 ^N−Δ −1). Therefore, describing this correspondence relationship defines the conversion process to a low bit depth image signal. Therefore, a bit depth conversion process is performed based on a separately determined correspondence, and further, the signal value y (y = 0,..., 2 ^N−Δ −1) is handled based on the correspondence. A look-up table storing signal values x (x = 0,..., 2 ^N−1 ) is prepared. Then, an inverse bit depth conversion process is performed based on the lookup table. The look-up table is a 2 ^N-Δ dimensional array. As an example of the correspondence relationship in the lookup table, it is preferable to use a correspondence relationship that minimizes the sum of squares of the bit depth prediction error signal for each image.

An example of the correspondence relationship that minimizes the sum of squares of the bit depth prediction error signal is shown below. x is a signal that can take four values (0, 1, 2, 3) (corresponding to N = 2), and the frequency of each value (0, 1, 2, 3) is 10, 1, 1, Suppose that it was 10. At this time, it is assumed that Δ = 1 is set and the above x is associated with y having two values. Here, in order to minimize the sum of squares of the bit depth prediction error signal, x = 0: 1 is converted to y = 0, and x = 2; 3 is converted to y = 1. At this time, the lookup table T [] used for the inverse bit depth conversion processing is configured as T [0] = 0, T [1] = 3.
As described above, reads the image signal formed by the pixel values of 2 ^N Street, when approximating the same pixel value as 2 ^N-delta, 2 so as to minimize the error caused by approximation ^N-delta By setting the street pixel value, it is possible to reduce the upper layer residual signal and improve the coding efficiency.

  In the present embodiment, the case has been described in which the encoding device 10 designates the bit depth conversion amount Δ in units of intra-encoded blocks, but is not limited thereto. For example, the bit depth conversion amount Δ may be specified in units of sequences, may be specified in units of frames, or a predetermined fixed value may be used as the bit depth conversion amount Δ. When a fixed value is used as the bit depth conversion amount Δ, if the encoding device 10 and the decoding device 20 share the value of the bit depth conversion amount Δ in advance, there is no need to include and transmit in the encoded stream. Also, the bit depth conversion amount Δ may be shared in a set of a plurality of intra-coded blocks. In this case, based on the information indicating that the bit depth conversion amount is shared in the same set, when the same information indicates sharing of the bit depth conversion amount, subsequently, by transmitting the bit depth conversion amount Δ only once, It is not necessary to encode the bit depth conversion amount for each intra-coded block.

When the encoding target is a color image, a common bit depth conversion amount Δ may be used for all color channels, or an individual bit depth conversion amount Δ may be used for each color channel.
Specifically, the encoding apparatus 10 reads the signal of each color channel, and sets the bit depth conversion amount for each color channel that can suppress the generated code amount required to realize the same encoding distortion. To adaptively control. As a result, the amount of generated codes for all color channel signals can be reduced, and the encoding efficiency is improved. In this case, the code input unit 51, the image signal decoding unit 52, the bit depth extension unit 53, and the reconstruction unit 54 of the decoding module 50 included in the encoding device 10 and the decoding device 20 perform a decoding process for each color channel.

  In addition, although this embodiment demonstrated the case where the encoding apparatus 10 designated the bit depth prediction error skip flag per intra coding block, it is not restricted to this. For example, the bit depth prediction error skip flag may be specified in sequence units, may be specified in frame units, or a predetermined fixed value may be used as the bit depth prediction error skip flag. Note that when a fixed value is used as the bit depth prediction error skip flag, the encoding device 10 may not include the first cost calculation unit 16 and the second cost calculation unit 17. In particular, when the bit depth prediction error skip flag indicates ON with a fixed value, the encoding device 10 does not need to include the difference generation unit 15.

FIG. 9 is a flowchart showing the operation of the encoding apparatus 10 in the first modification.
When the bit depth prediction error skip flag indicates a fixed value of ON, the encoding apparatus 10 differs from the operation of the encoding apparatus 10 in the above embodiment (see FIG. 3), and generates the upper layer residual signal and the first There is no need to calculate and compare the cost and the second cost. Therefore, when the bit depth prediction error skip flag indicates a fixed value of ON, the encoding apparatus 10 may perform at least steps S101 to S106 as shown in FIG. 9, and steps S107 to S107 shown in FIG. There is no need to perform step S115.

FIG. 10 is a flowchart showing the operation of the decoding device 20 in the first modification.
When the bit depth prediction error skip flag indicates a fixed value of ON, the decoding device 20 does not need to determine the bit depth prediction error skip flag, unlike the operation of the decoding device 20 in the above embodiment (see FIG. 6). . Therefore, when the bit depth prediction error skip flag indicates a fixed value of ON, the decoding device 20 does not need to perform the determination process in step S206 illustrated in FIG. 6 as illustrated in FIG. 10, and the step illustrated in FIG. Since it is not necessary to perform the process accompanied by the upper layer residual signal in S208 and Step S209, the calculation amount of the encoding / decoding process can be reduced.

In the present embodiment, the image signal decoding unit 52 of the decoding module 50 outputs an integer part image signal and a fractional part image signal generated by the real number calculation, respectively, and the reconstruction unit 54 outputs the integer part image signal and the integer part image signal. Although the case has been described in which the upper layer prediction signal is generated by bit-shifting the image signal of the decimal part to the left and then adding both signals, the present invention is not limited thereto. For example, the image signal decoding unit 52 outputs only the integer portion of the image signal, and the reconstruction unit 54 adds the integer portion of the image signal and a predetermined fixed value of Δ bits to the upper layer. A prediction signal may be generated.
In addition, the reconstruction unit 54 calculates a Δ bit value that minimizes the sum of the upper layer residual signals in the frame, and adds the value and the integer portion of the image signal to add the upper layer prediction signal. May be generated. Specifically, the lower layer residual signal is reduced by reading the lower Δ bits of the original image signal and setting in advance a Δ bit value that minimizes the approximation error for the lower bits. Thus, the encoding efficiency can be improved. Note that “a value of Δ bits that minimizes the approximation error” includes a value that approximates the value in addition to a value that minimizes the approximation error.

When the fractional part is not output, the configuration / action of the image signal decoding unit 52 is as follows.
FIG. 11 is a schematic block diagram showing the configuration of the image signal decoding unit 52 in the first modification.
Unlike the image signal decoding unit 52 (see FIG. 7) in the above embodiment, the image signal decoding unit 52 illustrated in FIG. 11 is configured to be able to acquire the integer part (N−Δ bits) of the image signal from the frame storage unit 524. Thus, the decimal part (Δ bit) of the image signal is not output. This eliminates the need to provide an output buffer after the in-loop filter processing unit 528.

FIG. 12 is a flowchart showing the operation of the image signal decoding unit 52 in the first modification.
In step S611, the in-loop filter processing unit 528 outputs the integer part signal and the fractional part signal from the decoded signal after processing in the image signal decoding part 52 in the above embodiment (see FIG. 8). On the other hand, when the image signal of the decimal part is not used, the frame storage unit 524 outputs the signal of the integer part of the image signal as shown in FIG. 12 instead of the process of step S611 shown in FIG. S614).

  In this embodiment, the case where the image signal decoding unit 52 of the decoding module 50 includes the in-loop filter processing unit 528 has been described. However, the present invention is not limited to this, and the in-loop filter processing unit 528 may not be provided. In this case, the integer part image signal and the decimal part image signal are output from the inverse transform / inverse quantization processing unit 522.

When the image signal decoding unit 52 does not include the in-loop filter processing unit 528, the configuration and operation of the image signal decoding unit 52 are as follows.
FIG. 13 is a schematic block diagram showing the configuration of the image signal decoding unit 52 in the second modification.
Unlike the image signal decoding unit 52 (see FIG. 7) in the above embodiment, the image signal decoding unit 52 illustrated in FIG. 13 does not include the in-loop filter processing unit 528 and is output from the inverse transform / inverse quantization processing unit 522. An integer part storage unit 529 which is a buffer for storing the integer part of the image signal, and a decimal part storage unit 530 which is a buffer for storing the decimal part of the image signal. The adding unit 527 generates a decoded signal by adding the prediction signal generated by the intra prediction processing unit 523 or the inter prediction processing unit 526 and the signal stored in the integer part storage unit 529. The integer part of the image signal can be acquired from the frame storage unit 524, and the decimal part of the image signal can be acquired from the decimal part storage unit 530.

FIG. 14 is a flowchart showing the operation of the image signal decoding unit 52 in the second modification.
When the in-loop filter processing unit 528 is not provided, after the inverse transformation / inverse quantization processing unit 522 generates a prediction residual signal in step S604, the inverse transformation / inverse quantization unit 522 performs an integer part of the prediction residual signal. Are stored in the integer part storage unit 528, and the decimal part of the prediction residual signal is stored in the decimal part storage unit 529. As a result, the decimal part of the prediction residual signal is output from the integer part storage unit 528 (step S615). Then, the in-loop filter process in step S610 shown in FIG. 8 is skipped, and instead of the process in step S611 shown in FIG. 8, the frame storage unit 524 outputs the signal of the integer part of the image signal (step S614).

Further, when the image signal decoding unit 52 does not include the in-loop filter processing unit 528, when the fractional part is not output, the configuration / action of the image signal decoding unit 52 is as follows.
FIG. 15 is a schematic block diagram showing the configuration of the image signal decoding unit 52 in the third modification.
The image signal decoding unit 52 shown in FIG. 15 does not include the in-loop filter processing unit 528, unlike the image signal decoding unit 52 (see FIG. 7) in the above embodiment. The integer part of the image signal can be acquired from the frame storage unit 524, and the decimal part of the image signal is not output.

FIG. 16 is a flowchart showing the operation of the image signal decoding unit 52 in the third modification.
According to the image signal decoding unit 52 in the third modified example, the in-loop filter process in step S610 illustrated in FIG. 8 is skipped, and the frame storage unit 524 replaces the process in step S611 illustrated in FIG. The signal of the integer part of is output (step S614).

  The decoding module 50, the encoding device 10, and the decoding device 20 described above have a computer system therein. The operation of each processing unit described above is stored in a computer-readable recording medium in the form of a program, and the above processing is performed by the computer reading and executing this program. Here, the computer-readable recording medium means a magnetic disk, a magneto-optical disk, a CD-ROM, a DVD-ROM, a semiconductor memory, or the like. Alternatively, the computer program may be distributed to the computer via a communication line, and the computer that has received the distribution may execute the program.

  The program may be for realizing a part of the functions described above. Furthermore, what can implement | achieve the function mentioned above in combination with the program already recorded on the computer system, and what is called a difference file (difference program) may be sufficient.

  DESCRIPTION OF SYMBOLS 10 ... Coding apparatus 11 ... Image signal input part 12 ... Bit depth conversion amount setting part 13 ... Bit depth reduction part 14 ... Image signal encoding part 15 ... Difference generation part 16 ... 1st cost calculation part 17 ... 2nd Cost calculation unit 18 ... Bit depth conversion prediction determination unit 19 ... Encoded stream output unit 20 ... Decoding device 21 ... Encoded stream input unit 22 ... Bit depth conversion amount decoding unit 23 ... Code signal output unit 24 ... Bit depth prediction error skip Flag decoding unit 25 ... difference decoding unit 26 ... difference reflection unit 50 ... decoding module 51 ... code input unit 52 ... image signal decoding unit 53 ... bit depth extension unit 54 ... reconstruction unit 521 ... entropy decoding processing unit 522 ... inverse transformation / Inverse quantization processing unit 523 ... Intra prediction processing unit 524 ... Frame storage unit 525 ... Motion vector decoding unit 526 ... Centers prediction processing unit 527 ... adding section 528 ... in-loop filter processing unit

Claims (15)

An image decoding module for reconstructing an image signal having a first bit depth that is a predetermined bit depth,
A code input unit for inputting a coded image signal having a second bit depth that is lower than the first bit depth;
An image signal decoding unit for decoding the encoded image signal input by the code input unit;
A bit depth extending unit that extends the bit depth of the image signal to the first bit depth using each bit of the image signal decoded by the image signal decoding unit;
The image signal is regenerated by adding a predetermined bit value having the same number of bits as the difference between the first bit depth and the second bit depth to the image signal whose bit depth has been expanded by the bit depth extension unit. An image decoding module comprising: a reconstructing unit for constructing. The image signal decoding unit performs a real number operation on the image signal input by the code input unit, and outputs an integer part of the calculation result as a decoded image signal,
The image decoding module according to claim 1, wherein the predetermined bit value added by the reconstruction unit is a decimal part of a result of a real number operation performed by the image signal decoding unit. The image decoding module according to claim 1, wherein the predetermined bit value added by the reconstruction unit is a value that minimizes an approximation error with respect to lower bits of the original image signal. A table for storing the second bit depth value in association with the first bit depth value so as to minimize an error caused by approximation to the first bit depth value;
The bit depth extension unit replaces each bit of the image signal decoded by the image signal decoding unit with the first bit depth value associated with the second bit depth value stored in the table. The image decoding module according to any one of claims 1 to 3, wherein the bit depth of the image signal is expanded to the first bit depth. The second bit depth is set for each color channel;
The image decoding module according to any one of claims 1 to 4, wherein the image signal decoding unit, the bit depth extension unit, and the reconstruction unit execute a decoding process for each color channel. . An image encoding device comprising the image decoding module according to any one of claims 1 to 5,
An image signal input unit for inputting an image signal of the first bit depth;
A bit depth reduction unit that reduces the bit depth of the image signal input by the image signal input unit to a second bit depth that is lower than the first bit depth;
An image signal encoding unit that encodes an image signal in which the bit depth reduction unit has reduced the bit depth and outputs the encoded image signal to a code input unit of the image decoding module;
A difference generation unit that generates a difference signal indicating a difference between the image signal reconstructed by the reconstruction unit of the image decoding module and the image signal input by the image signal input unit;
An image encoding device comprising: Weighted addition of the magnitude of the error of the image signal reconstructed by the reconstruction unit of the image decoding module and the amount of data of the image signal encoded by the image signal encoding unit with respect to the image signal input by the image signal input unit A first cost calculation unit for calculating a first cost by:
A second cost calculation unit that calculates a second cost according to the data amount of the image signal encoded by the image signal encoding unit and the difference signal generated by the difference generation unit;
When the first cost is lower than the second cost, the image signal decoded by the image signal decoding unit is output as an encoded stream, and when the first cost is higher than the second cost, the image The image encoding apparatus according to claim 6, further comprising: an encoded stream output unit that outputs an image signal decoded by the signal decoding unit and a difference signal generated by the difference generation unit as an encoded stream. An image decoding apparatus comprising the image decoding module according to any one of claims 1 to 5,
An encoded stream input for inputting an encoded stream including the encoded image signal having the second bit depth and a difference signal indicating a difference between the image signal reconstructed using the image decoding module and the original image signal And
A code signal output unit that outputs an image signal included in the encoded stream to a code input unit of the image decoding module;
An image decoding apparatus comprising: The image decoding apparatus according to claim 8, further comprising: a difference reflection unit that adds the difference data included in the encoded stream to the image signal reconstructed by the reconstruction unit of the image decoding module. An image decoding method using an image decoding module for reconstructing an image signal having a first bit depth which is a predetermined bit depth,
The code input unit inputs an encoded image signal having a second bit depth that is lower than the first bit depth,
The image signal decoding unit decodes the image signal input by the code input unit,
The bit depth extension unit uses each bit of the image signal decoded by the image signal decoding unit to extend the bit depth of the image signal to the first bit depth,
The reconstruction unit adds a predetermined bit value having the same number of bits as the difference between the first bit depth and the second bit depth to the image signal whose bit depth has been expanded by the bit depth expansion unit. An image decoding method characterized by reconstructing an image signal by using the method. The image signal decoding unit performs a real number operation on the image signal input by the code input unit, and outputs an integer part of the calculation result as a decoded image signal,
The image decoding method according to claim 10, wherein the predetermined bit value added by the reconstructing unit is a decimal part of a calculation result of a real number calculation by the image signal decoding unit. The image decoding method according to claim 10, wherein the predetermined bit value added by the reconstruction unit is a value that minimizes an approximation error with respect to lower bits of the original image signal. A table for storing the second bit depth value in association with the first bit depth value so as to minimize an error caused by approximation to the first bit depth value;
The bit depth extension unit replaces each bit of the image signal decoded by the image signal decoding unit with the first bit depth value associated with the second bit depth value stored in the table. The image decoding method according to any one of claims 10 to 12, wherein the bit depth of the image signal is expanded to the first bit depth. The second bit depth is set for each color channel;
The image decoding method according to any one of claims 10 to 13, wherein the image signal decoding unit, the bit depth extending unit, and the reconstruction unit execute a decoding process for each color channel. . An image decoding module for reconstructing an image signal of a first bit depth that is a predetermined bit depth;
A code input unit for inputting a coded image signal having a second bit depth that is lower than the first bit depth;
An image signal decoding unit for decoding the image signal input by the code input unit;
A bit depth extending unit that extends the bit depth of the image signal to the first bit depth using each bit of the image signal decoded by the image signal decoding unit;
The image signal is regenerated by adding a predetermined bit value having the same number of bits as the difference between the first bit depth and the second bit depth to the image signal whose bit depth has been expanded by the bit depth extension unit. A program for functioning as a reconstruction unit to be constructed.

Images