JP2012070249A - Image encoding method, image decoding method, image encoding device, image decoding device, and program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To improve encoding efficiency of encoding even if use of a code amount is restricted at a low rate.SOLUTION: When bit depth conversion predictive encoding is selected in intra-encoding, a bit depth conversion predictive encoding unit 60 performs bit depth conversion processing on an N-bit image signal for a designated encoding unit to convert the image signal to a low bit depth image of N-Δ bits, performs encoding/decoding processing on the lower layer signal, and performs inverse bit depth conversion processing on the decoded image. Finally, the unit encodes a difference signal between the N-bit image and an input image, and multiplexes the difference signal and the low bit depth image signal to output them as an encoded stream.

Description

  The present invention relates to an image encoding method, an image decoding method, an image encoding device, an image decoding device, and a program.

  The moving picture coding algorithms are roughly classified into inter-frame coding (inter coding) and intra-frame coding (intra coding). Interframe coding is an approach for compressing information by using 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. In JPEG and MPEG-2, an approach using a discrete cosine transform (DCT) is used, and in JPEG 2000, an approach using a discrete wavelet transform is used. Further, H.C. H.264 / AVC employs an approach that combines spatial direction prediction and DCT. Generally, compared with inter coding, the coding efficiency of intra coding is low, and further improvement in coding efficiency is required.

  FIG. 15 is a block diagram showing a configuration of an intra coder according to the prior art. In FIG. 15, an intra coding method designation information coding unit 1 inputs an input video signal. The intra coding scheme setting unit 2 applies any intra code among the intra coding units 3-1 to 3-M described later to the input video signal input to the intra coding scheme designation information coding unit 1. The switch unit 4 is switched and the input video signal is supplied to the set intra encoding units 3-1 to 3 -M. The intra coding units 3-1 to 3-M each have different intra coding schemes 1 to M, and perform intra coding according to each of the intra coding schemes 1 to M on the input video signal. , And output via the switch unit 5 (in conjunction with the switch unit 4). Information indicating the set intra coding method and the encoded video signal are output as an encoded stream.

  FIG. 16 is a flowchart for explaining the operation of the conventional intra encoder. First, the intra coding method designation information encoding unit 1 reads an input video signal (step S1). Next, in steps S2 to S5, intra-frame encoding is repeatedly performed for each predetermined partial region (encoding unit). First, the intra coding method setting unit 2 sets which intra coding method of the intra coding units 3-1 to 3-M, which will be described later, to be used for the inputted input video signal (step S3). ), Intra encoding is performed using any of the intra encoding units 3-1 to 3-M corresponding to the set intra encoding method (step S4). Then, when intra coding is completed for all the coding units, the processing is finished.

  FIG. 17 is a block diagram showing a configuration of an intra decoder according to the prior art. In FIG. 17, an intra encoding scheme designation information decoding unit 10 receives an encoded stream and decodes information indicating a set intra encoding scheme included in the encoded stream. The intra decoding scheme setting unit 11 sets (selects) whether to use any of the intra decoding units 12-1 to 12-M, which will be described later, according to the information indicating the decoded intra coding scheme. The switch unit 13 is switched to supply the encoded video signal to the set intra decoding units 12-1 to 12-M. The intra decoding units 12-1 to 12-M respectively have different intra decoding schemes 1 to M (corresponding to the intra coding schemes 1 to M), and each of the intra decoding schemes 1 to M for the encoded stream. Is output as a decoded image signal via the switch unit 14 (in conjunction with the switch unit 13).

  FIG. 18 is a flowchart for explaining the operation of the conventional intra decoder. First, the intra-encoding method designation information decoding unit 10 reads an encoded stream, and decodes information indicating the set intra-encoding method (step S10). Next, in steps S11 to S15, intra-frame decoding is performed for each coding unit. First, the intra decoding scheme setting unit 11 sets which intra decoding scheme of the intra decoding units 12-1 to 12-M, which will be described later, to be used (step S12), and the intra decoding unit corresponding to the set intra decoding scheme. Intra decoding is performed in any of 12-1 to 12-M (step S14). Then, when intra decoding is completed for all the coding units, the processing is terminated.

  In recent years, hierarchical coding having a hierarchical structure has been studied as a scalable coding method for high bit depth video (see, for example, Non-Patent Document 1). In this method, first, an N-bit image signal is input, and a bit depth conversion process is performed to convert it into an N-Δ bit low bit depth image. Encoding / decoding processing is performed on the low bit depth image. Further, an inverse bit depth conversion process is performed on the decoded image to generate an N-bit image signal. Finally, the difference signal between the generated N-bit image and the input image is encoded. The output is an encoded stream of the difference signal and the low bit depth image signal. As described above, this method is a method corresponding to the scalable coding of the bit depth.

  Hereinafter, a layer that performs encoding / decoding processing on an N-Δ bit signal is referred to as a lower layer, and a layer that performs encoding on a difference signal between an N-bit image and an input image is referred to as an upper layer.

“Bit-Depth Scalable Video Coding”, M. Winken, D. Marpe, H. Schwarz, and T. Wiegand, Proc. IEEE Intl. Conf. On Images Processing (ICIP2007), 2007.

  In the above-described hierarchical encoding, encoding processing over two layers with bit depth conversion is uniformly applied to a frame. Hereinafter, hierarchical encoding with bit depth conversion is referred to as bit depth conversion hierarchical encoding. Since this bit depth conversion hierarchical encoding is based on the assumption of scalable encoding, the use of a uniform bit depth conversion amount for a frame is imposed as a restriction.

By the way, in the case of non-scalable coding, such a restriction is not always necessary. In non-scalable coding, it is also possible to apply a coding method based partially on bit depth conversion to a specified region in a frame. That is, bit depth conversion hierarchical coding can be used as a new coding mode for intra coding. The restriction on the premise of scalable coding causes a decrease in coding efficiency, and is not necessary when used as a non-scalable coding method.
However, when the scheme based on scalable coding is simply diverted to the non-scalable coding scheme, there is a problem that there remains room for improvement in improving the coding efficiency.

  In addition, regarding “the intra prediction method in H.264”, a prediction signal is generated with reference to the encoded pixel values located around the encoding target block, and the prediction error is encoded. However, since the method is based on limited spatial prediction, there is a problem that there remains room for improvement in improving the coding efficiency.

  “Bit depth scalable coding” is scalable coding that supports generation of decoded images at two bit depths of N bits and N−Δbit (N is 8 or 10). However, the value of the bit depth conversion amount (Δ) is constrained to a fixed value in the frame in order to realize the scalability function, and the spatial locality is considered, that is, the region in the frame (for example, There is a problem that the value of the bit depth conversion amount Δ that differs for each block) cannot be adaptively set.

  The present invention has been made in view of such circumstances, and its purpose is to improve the coding efficiency of coding even when the use of the code amount is restricted at a low transmission rate. An image encoding method, an image decoding method, an image encoding device, an image decoding device, and a program are provided.

  In order to solve the above-described problems, the present invention reads an input video signal to be image-encoded and performs first intra coding using bit depth conversion and second intra coding not using bit depth conversion. To calculate the cost when the input video signal is encoded using a cost function based on the weighted sum of the code amount and the encoding distortion, and perform intra coding corresponding to the cost with the lowest calculated cost. A first step of selecting, a second step of encoding the input video signal by the first intra encoding, a third step of encoding the input video signal by the second intra encoding, and the first In the step, when the first intra encoding is selected, the input video signal encoded in the second step and information indicating the first intra encoding Is output as an encoded stream, and when the second intra encoding is selected, the input video signal encoded in the third step and information indicating the second intra encoding are output as an encoded stream. A fourth step of encoding the image.

  In the invention described above, the second step may include a fifth step of converting the input video signal based on a bit depth conversion amount determined according to the input video signal, and the bit depth. A sixth step of generating a first encoded signal by performing first encoding on the converted input video signal, and generating a decoded signal from the first encoded signal by decoding corresponding to the first encoding A seventh step, an eighth step of converting the bit depth of the decoded signal to the same bit depth as the input video signal by inverse bit depth conversion, the decoded signal subjected to the inverse bit depth conversion, and the input video signal And select whether to generate the second encoded signal by performing the second encoding on the difference signal indicating the calculated difference, or to omit the second encoding for the difference signal. It characterized in that it comprises a ninth step of outputting the first encoded signal and the second coded signal, or the first encoded signal as the coded stream.

In the invention described above, the bit depth conversion amount may be a predetermined number of bits according to the input video signal, or bit depth conversion may be performed on the input video signal. In this case, the cost function is a bit depth reduction bit number that minimizes the cost function.
In the present invention described above, the bit depth conversion amount and the determination information indicating the presence / absence of second encoding of the difference signal are set for each frame included in the input video signal. It is characterized by.
In the present invention described above, the bit depth conversion amount and the determination information indicating the presence / absence of second encoding of the difference signal may be further determined in advance for each frame included in the input video signal. It is set for each set area.
In the invention described above, when the input video signal includes a plurality of color channels, the bit depth conversion amount is set for each color channel.

  The present invention also includes a first step of reading an encoded stream to be decoded and decoding information indicating intra encoding obtained by encoding the encoded stream, and performing inverse bit depth conversion on the encoded stream. A second step of decoding using the first intra decoding method used, a third step of decoding the encoded stream using a second intra decoding method that does not use inverse bit depth conversion, and the decoded intra coding. And a fourth step of outputting the decoded signal decoded in one of the first step and the second step as a decoded image signal based on the indicated information.

  In the present invention described above, in the above-described invention, the second step includes a fifth step of separating data in the encoded stream, and a sixth step of decoding a bit depth conversion amount from the separated data. And a seventh step of performing first decoding on the separated data to generate a first decoded signal, and an inverse bit for increasing the number of bits corresponding to the bit depth conversion amount with respect to the first decoded signal An eighth step of performing depth conversion, a ninth step of decoding determination information indicating the presence or absence of a differential signal as a second encoded signal from the separated data, and the separated data based on the judgment information Whether to generate the second decoded signal by performing the second decoding on the above, or to omit the second decoding, and the difference between the decoded signal after the inverse bit depth conversion and the second decoded signal No., or a decoded signal either after the inverse bit depth conversion, characterized in that it comprises a tenth step of outputting as a decoded video signal.

In the invention described above, the present invention is characterized in that the bit depth conversion amount and the determination information are decoded for each frame included in the decoded video signal.
In the invention described above, the bit depth conversion amount and the determination information are further decoded for each predetermined region of each frame included in the decoded video signal. And
Further, in the present invention described above, when the encoded stream to be decoded includes a plurality of color channels, the bit depth conversion amount is set for each color channel. And

  Further, the present invention reads the input video signal to be image-encoded and uses the first intra encoding using bit depth conversion and the second intra encoding not using bit depth conversion, respectively. Selecting means for calculating the cost at the time of encoding the input video signal using a cost function based on the weighted sum of the code amount and the encoding distortion, and selecting the intra encoding corresponding to the cost with the lowest calculated cost; The first encoding means for encoding the input video signal by first intra encoding, the second encoding means for encoding the input video signal by second intra encoding, and the selection means, When 1 intra encoding is selected, the input video signal encoded by the first encoding means and information indicating the first intra encoding are output as an encoded stream, and the second When the intra coding is selected, an output unit that outputs an input video signal encoded by the second encoding unit and information indicating the second intra encoding as an encoded stream is provided. An image encoding device.

  Further, the present invention provides the bit depth conversion means for converting the input video signal based on a bit depth conversion amount set according to the input video signal, wherein the first encoding means is the invention described above. A first encoding means for generating a first encoded signal by performing a first encoding on the input video signal subjected to the bit depth conversion, and the first encoded signal by decoding corresponding to the first encoding. Decoding means for generating a decoded signal from, a reverse bit depth conversion means for converting the bit depth of the decoded signal to the same bit depth as the input video signal by reverse bit depth conversion, and the decoded signal subjected to the reverse bit depth conversion And the difference between the input video signal and the difference signal indicating the calculated difference are subjected to second encoding to generate a second encoded signal, or the second encoding of the difference signal is omitted. Select Luca, characterized by comprising a first encoded signal and the second encoded signal or switching means for outputting the first encoded signal as the coded stream.

Further, the present invention provides the bit depth conversion amount and the determination information indicating the presence / absence of second encoding of the difference signal in each of the frames included in the input video signal or the input in the invention described above. It is further characterized by further comprising setting means for setting for each predetermined region in the frame included in the video signal.
Further, the present invention is characterized in that, in the above-described invention, the bit depth conversion means sets the bit depth conversion amount for each color channel when the input video signal includes a plurality of color channels. To do.

  In addition, the present invention provides an encoding method restoration unit that reads an encoded stream to be decoded and decodes information indicating intra encoding obtained by encoding the encoded stream, and an inverse bit depth for the encoded stream. First intra decoding means for decoding by a first intra decoding scheme using transform, second intra decoding means for decoding by a second intra decoding scheme not using inverse bit depth conversion for the encoded stream, and the decoding Switching means for outputting a decoded signal decoded by either the first intra decoding means or the second intra decoding means as a decoded image signal based on the information indicating the intra coding performed. This is a featured image decoding apparatus.

  In the present invention described above, the first intra decoding unit decodes the bit depth conversion amount from the data separated by the separation unit and the data separated by the separation unit. Bit depth conversion amount decoding means for performing first decoding on the data separated by the separating means to generate a first decoded signal, and the bit for the first decoded signal Inverse bit depth conversion means for performing reverse bit depth conversion for increasing the number of bits corresponding to the depth conversion amount, and determination information decoding means for decoding determination information indicating presence / absence of a differential signal as second encoded data from the separated data And, based on the determination information, select whether to perform a second decoding on the separated data to generate a second decoded signal, or to omit the second decoding, Output means for outputting either a difference signal between the decoded signal after reverse bit depth conversion and the second decoded signal or the decoded signal after reverse bit depth conversion as a decoded video signal, To do.

Further, the present invention is the above-described invention, wherein the bit depth conversion amount and the determination information indicating the presence / absence of the second decoding are included for each frame included in the decoded video signal or in the decoded video signal. It is set for each predetermined area in each frame.
Further, in the present invention described above, when the encoded stream to be decoded includes a plurality of color channels, the bit depth conversion amount is set for each color channel. And

  The present invention also provides a first intra coding using a bit depth conversion by reading an input video signal to be image-encoded into a computer provided in an image coding apparatus for coding an input video signal. The cost when the input video signal is encoded using each of the second intra encoding not using the bit depth conversion is calculated using a cost function based on the weighted sum of the code amount and the encoding distortion. A first step of selecting an intra coding corresponding to a cost having the lowest cost; a second step of coding the input video signal by the first intra coding; and the input video signal by the second intra coding. When the first intra encoding is selected in the third step and the first step, the input encoded in the second step is encoded. When the video signal and the information indicating the first intra encoding are output as an encoded stream and the second intra encoding is selected, the input video signal encoded in the third step, and the first This is a program for executing the fourth step of outputting information indicating 2-intra encoding as an encoded stream.

  The present invention also reads a coded stream to be decoded into an image decoding apparatus that decodes the coded stream, and decodes information indicating intra coding obtained by coding the coded stream. A first step, a second step of decoding the encoded stream by a first intra decoding method using an inverse bit depth transform, and a second intra decoding not using an inverse bit depth transform of the encoded stream. Based on the third step decoded by the method and the information indicating the decoded intra coding, the decoded signal decoded by either the first step or the second step is output as a decoded image signal. This is a program for executing the fourth step.

  According to the present invention, the encoding efficiency of encoding can be improved even when the use of the code amount is restricted at a low rate.

It is a block diagram which shows the structure of the moving image encoder by embodiment of this invention. It is a block diagram which shows the structure of the moving image decoding apparatus by embodiment of this invention. It is a block diagram which shows the structure of the intra encoder by this 1st Embodiment. It is a block diagram which shows the structure of the bit depth conversion prediction encoding part 60 by this 1st Embodiment. It is a flowchart for demonstrating operation | movement of the intra encoder by this 1st Embodiment. It is a flowchart for demonstrating operation | movement of the intra encoder by this 1st Embodiment. It is a block diagram which shows the structure of the intra decoder by this 1st Embodiment. It is a block diagram which shows the structure of the bit depth conversion prediction decoding part 90 by this 1st Embodiment. It is a flowchart for demonstrating operation | movement of the intra decoder by this 1st Embodiment. It is a flowchart for demonstrating operation | movement of the intra decoder by this 1st Embodiment. It is a flowchart for demonstrating operation | movement of the intra encoder by this 2nd Embodiment. It is a flowchart for demonstrating operation | movement of the intra encoder by this 2nd Embodiment. It is a flowchart for demonstrating operation | movement of the intra decoder by this 2nd Embodiment. It is a flowchart for demonstrating operation | movement of the intra decoder by this 2nd Embodiment. It is a block diagram which shows the structure of the intra encoder by a prior art. 5 is a flowchart for explaining the operation of an intra encoder according to the prior art. It is a block diagram which shows the structure of the intra decoder by a prior art. 6 is a flowchart for explaining the operation of an intra decoder according to the prior art.

  Hereinafter, an embodiment of the present invention will be described with reference to the drawings.

  First, an outline of the present invention will be described. Intraframe coding is performed for each predetermined partial area. This partial area is called an encoding unit. For example, H.M. In the case of H.264, a 16 × 16 pixel area called a macroblock corresponds to the encoding unit. Even in a frame to which interframe coding is applied, intraframe coding may be partially applied. H. In H.264, intra macroblocks in P-picture and B-picture correspond to it. Further, the entire screen may be set as a predetermined partial area.

  In the present invention, the following bit depth conversion encoding and bit depth conversion decoding are introduced as one of the encoding method and decoding method for the encoding unit in intra-frame encoding.

  In the image signal, the upper bit and the lower bit have different properties. In general, the upper bits have high correlation between spatially neighboring pixels, and redundancy can be removed by compression processing such as prediction using the correlation. On the other hand, the lower bits have a low correlation between spatially neighboring pixels, and it is difficult to remove redundancy by compression processing such as prediction. Therefore, bit depth conversion processing (or bit separation processing) is performed to separate the upper bits and lower bits of the image signal, only the upper bits are compressed, and a difference from the input signal is generated. By encoding the difference signal, the entire code amount is reduced.

  First, an N-bit image signal is input to a designated encoding unit, and a bit depth conversion process is performed on the image signal to convert it into an N-Δ bit low bit depth image. This N-Δ bit signal is called a lower layer signal. The bit depth conversion amount Δ is designated as additional information, and indicates a bit depth reduction amount. Encoding / decoding processing is performed on the low bit depth image, that is, the lower layer signal. Further, an inverse bit depth conversion process is performed on the decoded image to generate an N-bit image signal. Finally, the difference signal between the generated N-bit image and the input image is encoded. This difference signal is called an upper layer signal. The image encoding device in the embodiment outputs an encoded stream including the difference signal and the low bit depth image signal.

  Here, the method of extracting N−Δ bit information from the N bit signal in the bit depth conversion can be generalized as a process of classifying into a set of “2 to the N−Δ power” signal values. This method of extracting N-Δ bit information from the N-bit signal is called bit depth conversion processing. The above description of the upper bits and the lower bits is a method of classifying a set of signal values for “2 to the N power” signal values at regular intervals “2 to the power of Δ”. For this reason, the classification method is fixed regardless of the distribution of signal values. On the other hand, according to the distribution of signal values, it is possible to reduce the prediction error by classifying densely distributed locations finely and roughly classifying sparsely distributed locations. .

  It is also possible to share the bit depth conversion amount Δ in a set of encoding units composed of a plurality of encoding units. In this case, based on the information indicating that the bit depth conversion amount is shared in the same set, when the information indicates sharing of the bit depth conversion amount, the bit depth conversion amount Δ is transmitted only once and encoded. The bit depth conversion amount is not encoded for each unit.

  Furthermore, as another form of the bit depth conversion prediction mode, there is a form in which encoding of a difference signal between the generated N-bit image and the input image is omitted. In this case, it is necessary to store information indicating that encoding of the differential signal is omitted as part of the encoded information.

When a color image is to be encoded, the following two forms are possible.
(I) A common bit depth conversion amount Δ is used for all color channels.
(Ii) An individual bit depth conversion amount Δ is used for each color channel.

  Note that the bit depth transform coding described above always performs some kind of coding processing on the upper layer signal. For this reason, the generated code amount includes identifier information for identifying this mode in intra coding, lower layer signal encoding information, and upper layer signal encoding information.

  At low rates, code savings are required. Therefore, a method (bit depth conversion prediction) in which encoding of an upper layer signal is omitted and an N-bit image signal obtained by inverse bit depth conversion processing is used as a decoded signal is introduced. In the bit depth conversion prediction, information (1 bit) indicating whether or not the bit depth conversion prediction can be used is transmitted as encoded information following the identifier for identifying the bit depth conversion encoding. This information is called a bit depth conversion prediction flag.

  The bit depth conversion prediction flag can be set for each encoding unit. Alternatively, it is also possible to share the bit depth conversion prediction flag in a set of encoding units including a plurality of encoding units. Alternatively, there is a format in which the bit depth conversion amount Δ and the bit depth conversion prediction flag are set in pairs. In this case, the unit sharing the bit depth conversion amount Δ (eg, encoding unit, set of the same unit) and the unit sharing the bit depth conversion prediction flag are the same.

  Further, the application of the present invention is not limited to an intra-coded frame, and the same applies to an intra-coded region (eg, an intra macroblock) in a frame (eg, P-picture) to be inter-coded. It is applicable to.

  FIG. 1 is a block diagram showing a configuration of a moving picture coding apparatus according to an embodiment of the present invention. In FIG. 1, an encoding control unit 30 controls the operation of each unit. The switch unit 31 selects predicted image information from either the intra prediction processing unit 40 or the inter prediction processing unit 39. The subtracter 32 subtracts the predicted image information from the input image. The transform / quantization unit 33 performs integer orthogonal transform on the difference data of the image, and performs quantization on a transform coefficient subjected to the integer orthogonal transform with a predetermined quantization scale.

  The inverse transform / inverse quantization unit 34 performs a predetermined inverse quantization process on the transform coefficient quantized by the transform / quantization unit 33, and converts the inverse quantized transform coefficient into the original image data space. Perform inverse integer orthogonal transform to return. The adder 35 adds predicted image information described later to the image difference information returned to the image data space. The deblocking filter 36 corrects the discontinuity in the boundary data of a predetermined pixel block unit with respect to the image data restored by the predicted image information. The frame memory 37 holds the restored image data that has been subjected to block boundary correction processing, which will be described later, for use as a reference image of the predicted image information.

  The motion detection unit 38 divides the current image data input from the intra prediction processing unit 40 into a plurality of pixel blocks, and each pixel block has a position highly correlated with the reference frame candidate stored in the frame memory 37. The motion search process to detect is performed, and the difference data of the position is detected as motion information between frames. The inter (inter-screen) prediction encoding unit 39 performs motion compensation that generates predicted image information of the current image data from the reference frame and the motion information output from the motion detection unit 38. The intra (in-screen) prediction processing unit 40 divides image data into a plurality of pixel blocks, predicts image data of each pixel block from pixels around the block, and generates predicted image information. The entropy encoding unit 41 performs entropy encoding processing on the transform coefficient quantized by the transform / quantization unit, compresses the data, and outputs the result as an encoded stream.

  FIG. 2 is a block diagram showing a configuration of the moving picture decoding apparatus according to the embodiment of the present invention. In FIG. 2, an entropy decoding unit 50 performs variable-length decoding processing on an input stream including encoded moving image data, quantized DCT coefficients (hereinafter referred to as quantized DCT coefficients), and quantization parameters. A QP (Quantization Parameter) is acquired, and decoding information including a reference frame number, a prediction mode, and the like is output.

  The inverse transform / inverse quantization unit 51 performs DCT transform processing on the input DCT coefficient, and outputs the processing result to the adding unit 14 as a residual signal. Based on the quantization parameter QP, the inverse quantization processing is performed. Then, the quantized DCT coefficient is inversely quantized to obtain the DCT coefficient. The motion vector decoding unit 52 decodes a motion vector indicating the motion between frames.

  The adder 55 adds the residual signal and the predicted image data output from the intra prediction processing unit 53 or the inter prediction processing unit 54 to reproduce the frame image data. The deblocking filter 56 performs processing for improving the distortion between encoded blocks on the frame image data output from the adder 55, and outputs the processing result to the frame memory 57. The frame memory 57 stores the decoded image data processed by the deblocking filter 56.

  The intra prediction processing unit 53 performs intra-frame prediction based on the frame image data output from the adder 55 and the quantized DCT coefficient obtained by the entropy decoding unit 50, and generates predicted image data. The inter prediction processing unit 54 performs inter-frame prediction based on the quantized DCT coefficient acquired by the entropy decoding unit 50, the motion vector from the motion vector decoding unit 52, the reference image output from the frame memory 57, and the like. Predictive image data is generated. The switch unit 58 selects predicted image data from either the intra prediction processing unit 53 or the inter prediction processing unit 54 and outputs the selected prediction image data to the adder 55.

  The present invention relates to the intra prediction processing unit 40 shown in FIG. 1 and the intra prediction processing unit 53 shown in FIG.

A. First Embodiment A first embodiment of the present invention will be described. The first embodiment relates to an improvement in the efficiency of “intra prediction (intra coder)” in a moving picture coding apparatus.
FIG. 3 is a block diagram showing the configuration of the intra encoder according to the first embodiment. Note that portions corresponding to those in FIG. 15 are denoted by the same reference numerals and description thereof is omitted. In FIG. 3, a bit depth conversion prediction encoding unit 60 is provided in addition to the intra encoding units 3-1 to 3 -M. The intra coding units 3-1 to 3-M use an existing intra coding method. Intra prediction defined in H.264. That is, the intra encoding units 3-1 to 3-M perform intra encoding that does not use bit depth conversion, and the bit depth conversion prediction encoding unit 60 performs intra encoding that uses bit depth conversion.

  The bit depth conversion predictive encoding unit 60 receives an N-bit image signal as an input for the designated encoding unit, and performs bit depth conversion processing to convert it into an N-Δ bit low bit depth image, The low bit depth image, that is, the lower layer signal is encoded / decoded, and the decoded image is subjected to inverse bit depth conversion to generate an N bit image signal. The differential signal between the N-bit image and the input image is encoded, and the differential signal and the low bit depth image signal are multiplexed and output as an encoded stream.

  FIG. 4 is a block diagram illustrating a configuration of the bit depth conversion prediction encoding unit 60 according to the first embodiment. In FIG. 4, the bit depth conversion amount setting unit 71 reads and sets the bit depth conversion amount Δ and the bit depth conversion prediction flag selected by the intra coding scheme setting unit 2. The bit depth conversion unit 72 performs a bit depth conversion process on the input signal according to the bit depth conversion amount Δ, and outputs a lower layer signal with a bit depth of N−Δ.

  The lower layer encoding unit 73 performs an encoding process on the lower layer signal. The lower layer decoding unit 74 performs a decoding process on the lower layer signal and outputs an N-Δ bit decoded signal (lower layer decoded signal). The inverse bit depth conversion unit 75 performs inverse bit depth conversion on the lower layer decoded signal and outputs an N-bit decoded signal (upper layer prediction signal). The bit depth conversion prediction determination unit 76 determines whether or not the bit depth conversion prediction flag is “1”. The bit depth conversion prediction flag is set by the intra coding scheme setting unit 2 described above so that the cost calculated using the cost function represented by the weighted sum of the code amount and the coding distortion is minimized.

When the bit depth conversion prediction flag is “1”, the switch unit 77 outputs a zero value to the encoded stream multiplexing processing unit 80, and when the bit depth conversion prediction flag is “0”, the upper layer The prediction signal is output to the subtracter 79. The subtractor 79 subtracts the upper layer prediction signal from the input signal to generate a difference signal. The upper layer encoding unit 78 encodes the difference signal generated by the subtracter 79.
When the bit depth conversion prediction flag is “1”, the encoded stream multiplexing processing unit 80 uses the bit depth conversion amount, the lower layer signal, and the encoded data of the bit depth conversion prediction flag as one stream corresponding to the partial region. If the bit depth conversion prediction flag is “0”, the bit depth conversion amount, the lower layer signal, the bit depth conversion prediction flag, and the encoded data of the difference signal are encoded into one partial region. Performs multiplexing into a stream.

  5 and 6 are flowcharts for explaining the operation of the intra encoder according to the first embodiment. First, the intra coding method designation information coding unit 1 reads an input signal (step S21). Next, in steps S22 to S37, intraframe encoding is performed for each predetermined partial area (encoding unit).

  First, the intra coding scheme setting unit 2 uses any intra coding scheme (intra coding mode) among the intra coding units 3-1 to 3-M and the bit depth conversion prediction coding unit 60 described later. Is set (step S23). As an example of the setting method, there is a method of selecting an intra coding method that minimizes the cost function by using a weighted sum of code amount and coding distortion given in advance as a cost function. In step S23, the intra coding scheme setting unit 2 sets the bit depth conversion amount Δ and the bit depth conversion prediction flag that minimize the above cost function. The bit depth conversion amount Δ is any value from 1 to (N−1), and the bit depth conversion prediction flag is set to “1” when the difference signal is zero, and the difference signal is zero. When not, “0” is set. The bit depth conversion amount Δ may be a value determined in advance according to the input signal.

  Next, it is determined whether or not the intra coding mode is the bit depth conversion mode (step S24). When the intra coding mode is the bit depth conversion mode, the bit depth N of the input signal is read, written as encoded data (step S25), and a bit depth conversion algorithm is set (step S26). This bit depth conversion algorithm is given from the outside and is a predetermined algorithm. Examples of bit depth conversion algorithms include rounding, rounding down, and rounding up. Of course, other bit depth conversion algorithms can be used.

  Next, the bit depth conversion amount setting unit 71 reads the bit depth conversion amount Δ determined at the time of setting the intra coding mode (step S23) (step S27), and the bits determined at the time of setting the intra coding mode. A depth conversion prediction flag is read (step S28). More specifically, the bit depth conversion amount Δ and the bit depth conversion prediction flag are set in step S23 described above so that the cost calculated by the weighted sum of the code amount and the encoding distortion is minimized.

  Next, the bit depth conversion unit 72 performs a bit depth conversion process on the input signal (N bits), outputs a lower layer signal with a bit depth of N−Δ (step S29), and performs lower layer encoding. The unit 73 performs an encoding process on the lower layer signal and outputs the encoded data obtained by the encoding process to the encoded stream multiplexing processing unit 80 (step S30).

  Next, the bit depth conversion prediction determination unit 76 determines whether or not the bit depth conversion prediction flag is “1” (step S31). When the bit depth conversion prediction flag is “1”, The encoding process for the encoding unit is ended, and the encoding process for the next encoding unit is started.

  On the other hand, when the bit depth conversion prediction flag is “0”, the lower layer decoding unit 74 performs a decoding process on the encoded data output from the lower layer encoding unit 73, and N obtained by the decoding process. A -Δ bit decoded signal (lower layer decoded signal) is output (step S32). Then, the inverse bit depth conversion unit 75 performs inverse bit depth conversion of the lower layer decoded signal output from the lower layer decoding unit 74 into an N-bit decoded signal (upper layer prediction signal) (step S33).

  The bit depth conversion prediction determination unit 76 outputs the upper layer prediction signal generated by the inverse bit depth conversion unit 75 to the subtracter 79 via the switch unit 77, and the subtractor 79 receives the upper layer prediction signal from the input signal. A difference signal obtained by subtraction is generated (step S34). Next, the upper layer encoding unit 78 performs an encoding process on the difference signal generated by the subtractor 79 and outputs encoded data obtained by the encoding process to the encoded stream multiplexing processing unit 80 ( Step S35).

On the other hand, if the intra coding mode is not the bit depth conversion mode in step S24, the intra coding units 3-1 to 3-M perform preset intra coding (step S36). For example, H.M. H.264 is intra prediction. Also in this case, the above-described steps S22 to S37 are repeatedly executed for the next encoding unit.
In any case, the above-described steps S22 to S37 are repeatedly executed for the next encoding unit.

  Then, when intra coding is completed for all the coding units, the processing is finished.

  FIG. 7 is a block diagram showing the configuration of the intra decoder according to the first embodiment. In addition, the same code | symbol is attached | subjected about the part corresponding to FIG. 17, and description is abbreviate | omitted. In FIG. 7, a bit depth conversion prediction decoding unit 90 is provided in addition to the intra encoding units 12-1 to 12 -M. The bit depth conversion predictive decoding unit 90 decodes the bit depth conversion amount Δ determined at the time of setting the intra coding mode, performs a decoding process on the encoded data of the lower layer signal, and outputs an N−Δ bit decoded signal. (Lower layer decoded signal) is generated, inverse bit depth conversion is performed on the lower layer decoded signal to generate an N bit decoded signal (upper layer prediction signal), and the N bit decoded signal (upper layer prediction signal). Signal) and an upper layer decoded signal are generated and output as a decoded signal.

  FIG. 8 is a block diagram illustrating a configuration of the bit depth conversion prediction decoding unit 90 according to the first embodiment. In FIG. 8, the encoded stream separation processing unit 91 separates the encoded stream and supplies it to the bit depth conversion amount decoding unit 92, the lower layer decoding unit 93, and the bit depth conversion prediction bit decoding unit 95, respectively. The bit depth conversion amount decoding unit 92 decodes the bit depth conversion amount Δ determined when the intra coding mode is set. The lower layer decoding unit 93 performs a decoding process on the encoded data of the lower layer signal, and outputs an N-Δ bit decoded signal (lower layer decoded signal).

  The inverse bit depth conversion unit 94 performs inverse bit depth conversion on the lower layer decoded signal of N−Δ bits and outputs an N bit decoded signal (upper layer prediction signal). The bit depth conversion prediction bit decoding unit 95 decodes the bit depth conversion prediction flag determined when setting the intra coding mode. The bit depth conversion prediction determination unit 96 determines whether or not the bit depth conversion prediction flag is “1”.

The switch unit 97 switches whether to output the upper layer decoded signal output from the upper layer decoding unit 98 to the adder 99.
The upper layer decoding unit 98 decodes the difference signal input via the bit depth conversion prediction bit decoding unit 95 and the bit depth conversion prediction determination unit 95, and outputs the decoded difference signal to the adder 99 as an upper layer decoded signal. To do. The adder 99 adds the upper layer prediction signal and either the upper layer decoded signal or the zero value, and outputs the addition result as a decoded signal.

  9 and 10 are flowcharts for explaining the operation of the intra decoder according to the first embodiment. First, an encoded stream is read (step S41). Next, in steps S42 to S56, intra-frame decoding is performed for each predetermined partial region (encoding unit).

  First, the intra coding mode designation information decoding unit 10 decodes the intra coding mode (step S43), and the intra decoding mode setting unit 11 determines whether or not the intra coding mode is the bit depth conversion mode. (Step S44). If the intra coding mode is the bit depth conversion mode, the bit depth N of the input signal is decoded (step S45), and a bit depth conversion algorithm is set (step S46).

  In addition, when shared by an encoder / decoder, information for designating a bit depth conversion algorithm may be included in the encoded stream. In this case, the corresponding information is decoded and acquired. Examples of bit depth conversion algorithms include rounding, rounding down, and rounding up. Of course, other bit depth conversion algorithms can be used.

  Next, the bit depth conversion amount decoding unit 92 decodes the bit depth conversion amount Δ determined at the time of setting the intra coding mode (step S47), and the bit depth conversion prediction bit decoding unit 95 decodes the intra coding mode. The bit depth conversion prediction flag determined at the time of setting is decoded (step S48). Next, the lower layer decoding unit 93 performs decoding processing on the encoded data of the lower layer signal, outputs an N-Δ bit decoded signal (lower layer decoded signal) (step S49), and performs inverse bit depth conversion. The unit 94 performs inverse bit depth conversion on the lower layer decoded signal and outputs an N-bit decoded signal (upper layer prediction signal) (step S50).

  Next, the bit depth conversion prediction determination unit 96 determines whether or not the decoded bit depth conversion prediction flag is “1” (step S51). If the bit depth conversion prediction flag is “1”, The zero value is output to the adder 99, and the upper layer prediction signal from the inverse bit depth conversion unit 94 is output as a decoded signal (step S52).

  On the other hand, when the bit depth conversion prediction flag is “0”, the bit depth conversion prediction determination unit 96 converts the encoded data of the difference signal input via the bit depth conversion prediction bit decoding unit 95 into an upper layer decoding unit. The higher layer decoding unit 98 decodes the difference signal (step S53). Then, the adder 99 adds the upper layer prediction signal and the difference signal, and generates a decoded signal from the addition result (step S54).

On the other hand, if the intra coding mode is not the bit depth conversion mode in step S44, the intra decoding units 12-1 to 12-M perform a decoding process corresponding to the preset intra coding (step S55). ). For example, H.M. H.264 is intra prediction.
In any case, the above-described steps S42 to S56 are repeatedly executed for the next encoding unit.

  Then, when intra coding is completed for all the coding units, the processing is finished.

B. Second Embodiment Next, a second embodiment of the present invention will be described.
The second embodiment is characterized in that “bit shift” is used as an algorithm for extracting (N−Δ) bits from N bits in the first embodiment described above. In addition, since the structure of an intra encoder, a bit depth conversion prediction encoding part, an intra decoder, and a bit depth conversion prediction decoding part is the same as 1st Embodiment, description is abbreviate | omitted.

  11 and 12 are flowcharts for explaining the operation of the intra encoder according to the second embodiment. First, the intra coding method designation information encoding unit 1 reads an input signal (step S61). Next, in steps S62 to S76, intraframe encoding is performed for each predetermined partial region (encoding unit).

  First, the intra coding scheme setting unit 2 uses any intra coding scheme (intra coding mode) among the intra coding units 3-1 to 3-M and the bit depth conversion prediction coding unit 60 described later. Is set (step S63). As an example of the setting method, there is a method of selecting an intra coding mode that minimizes the cost function by using a weighted sum of code amount and coding distortion given in advance as a cost function.

  Next, it is determined whether or not the intra coding mode is the bit depth conversion mode (step S64). When the intra coding mode is the bit depth conversion mode, the bit depth N of the input signal is read and written as encoded data (step S65), and the bit depth conversion amount setting unit 71 is set when the intra coding mode is set. The bit depth conversion amount Δ determined in (Step S63) is read (Step S66), and the bit depth conversion prediction flag determined when the intra coding mode is set is read (Step S67). More specifically, the bit depth conversion amount Δ and the bit depth conversion prediction flag are set in step S63 described above so that the cost calculated by the weighted sum of the code amount and the coding distortion is minimized.

  Next, the bit depth conversion unit 72 performs a bit shift of Δ bits to the right with respect to the input signal (N bits), and outputs a lower layer signal with a bit depth of N−Δ (step S68). The layer encoding unit 73 performs an encoding process on the lower layer signal, and performs encoded stream multiplexing on the encoded data of the N-Δ bit decoded signal (lower layer decoded signal) obtained by the encoding process. The data is output to the processing unit 80 (step S69).

  Next, the bit depth conversion prediction determination unit 76 determines whether or not the bit depth conversion prediction flag is “1” (step S70). If the bit depth conversion prediction flag is “1”, The encoding process for the encoding unit is ended, and the encoding process for the next encoding unit is started.

  On the other hand, when the bit depth conversion prediction flag is “0”, the lower layer decoding unit 74 performs a decoding process on the encoded data output from the lower layer encoding unit 37, and N obtained by the decoding process. A -Δ bit decoded signal (lower layer decoded signal) is output (step S71). Then, the inverse bit depth conversion unit 75 shifts Δ bits to the left with respect to the lower layer decoded signal output from the lower layer decoding unit 74 to generate an N-bit decoded signal (upper layer prediction signal). (Step S72).

  Next, the bit depth conversion prediction determination unit 76 outputs the upper layer prediction signal generated by the inverse bit depth conversion unit 75 to the subtractor 79, and the subtracter 79 subtracts the upper layer prediction signal from the input signal to obtain a difference signal. Is generated (step S73). The upper layer encoding unit 78 performs an encoding process on the difference signal generated by the subtractor 79, and outputs encoded data obtained by the encoding process to the encoded stream multiplexing processing unit 80 (step S74). ).

On the other hand, if the intra coding mode is not the bit depth conversion mode in step S64, the intra coding units 3-1 to 3-M perform preset intra coding (step S75). For example, H.M. H.264 is intra prediction.
In any case, the above-described steps S62 to S76 are repeatedly executed for the next encoding unit.

  Then, when intra coding is completed for all the coding units, the processing is finished.

  13 and 14 are flowcharts for explaining the operation of the intra decoder according to the second embodiment. First, an encoded stream is read (step S81). Next, in steps S82 to S95, intra-frame decoding is performed for each predetermined partial region (encoding unit).

  First, the intra coding mode designation information decoding unit 10 decodes the intra coding mode (step S83), and the intra decoding mode setting unit 11 determines whether or not the intra coding mode is the bit depth conversion mode. (Step S84). When the intra coding mode is the bit depth conversion mode, the bit depth N of the input signal is decoded (step S85), and the bit depth conversion amount decoding unit 92 is determined when the intra coding mode is set. The bit depth conversion amount Δ is decoded (step S86).

  Next, the bit depth conversion prediction bit decoding unit 95 decodes the bit depth conversion prediction flag determined when the intra encoding mode is set (step S87). Next, the lower layer decoding unit 93 performs a decoding process on the encoded data of the lower layer signal, outputs an N-Δ bit decoded signal (lower layer decoded signal) (step S88), and performs inverse bit depth conversion. The unit 94 performs inverse bit depth conversion on the lower layer decoded signal and outputs an N-bit decoded signal (upper layer prediction signal) (step S89).

  Next, the bit depth conversion prediction determination unit 96 determines whether or not the bit depth conversion prediction flag is “1” (step S90). If the bit depth conversion prediction flag is “1”, a zero value is determined. Is output to the adder 99, and the adder 99 outputs the upper layer prediction signal output from the inverse bit depth conversion unit 94 as a decoded signal (step S91).

  On the other hand, when the bit depth conversion prediction flag is “0”, the bit depth change prediction immediate determination unit 96 decodes the encoded data of the difference signal input via the bit depth conversion prediction bit decoding unit 95 to the upper layer decoding The higher layer decoding unit 98 decodes the difference signal (step S92). Then, the adder 99 adds the upper layer prediction signal and the difference signal, and generates a decoded signal based on the addition result (step S93).

  If the intra coding mode is not the bit depth conversion mode in step S84, the intra decoding units 12-1 to 12-M perform a decoding process corresponding to preset intra coding (step S94). ). For example, H.M. H.264 is intra prediction. Also in this case, the above-described steps S82 to S95 are repeatedly executed for the next encoding unit.

  Then, when intra coding is completed for all the coding units, the processing is finished.

  In the operations of the intra encoder and intra decoder according to the second embodiment described above, the operations of steps S64 to S74 and steps S84 to S93 will be described. Hereinafter, the bit length of the input signal is N bits, and the bit length of the lower bits is Δ.

  In the image signal, the upper bit and the lower bit have different properties. In general, upper bits have a high correlation between spatial desired pixels, and redundancy can be removed by compression processing such as prediction. On the other hand, the lower order bits have low correlation, and it is difficult to remove redundancy by compression processing such as prediction. For this reason, it is difficult to reduce the prediction error (the difference signal between the input signal and the prediction signal) in the prediction process for the lower bits.

  Therefore, in the second embodiment, an N-Δ bit signal (corresponding to the upper N-Δ bit) is generated by shifting the input signal to the right by Δ bits by bit shifting, and the N-Δ bit signal is generated. A prediction signal of the input signal (N-bit signal) is generated by performing compression processing / decoding processing on the signal, and further inverse bit depth conversion processing (processing of expanding the N-Δ bit signal to the N-bit signal) A prediction error (difference signal) of the prediction signal with respect to is generated. Accordingly, it is possible to reduce the prediction error as compared with the spatial prediction, and as a result, it is possible to reduce the code amount.

According to the first and second embodiments described above, the following effects can be obtained.
Conventionally, in intra coding, due to the poor prediction performance of intra prediction, when the prediction signal itself is used as a decoded signal, it has been difficult to guarantee the decoded image quality. The system has been adopted. On the other hand, by introducing bit depth conversion prediction, even a prediction signal alone can be used as a decoded signal, and the bit depth conversion amount can be freely set according to the nature of the image. For this reason, according to the first and second embodiments described above, it is possible to improve the coding efficiency of intra coding even when the use of the code amount is restricted at a low rate.

Also, the bit depth conversion amount is the optimum bit depth conversion for the frame or partial region by using the bit depth reduction amount that minimizes the cost when coding using the bit depth conversion on the frame or partial region. The amount of encoding can be reduced by bit depth conversion.
Further, according to the first and second embodiments described above, the amount of encoding can be reduced by bit depth conversion by setting the bit depth conversion amount corresponding to the video of the frame for each frame. Further, by setting the bit depth conversion amount corresponding to the video for each predetermined partial region (encoding unit) of each frame, the encoding amount can be further reduced by bit depth conversion.
Further, according to the second embodiment described above, the amount of calculation can be reduced by performing bit depth conversion using a simple calculation called “bit shift”.

  Also, steps S32 to S35 in the encoding process in the first embodiment, steps S51 to S54 in the decoding process, steps S71 to S74 in the encoding process in the second embodiment, and steps S90 to S90 in the decoding process. In S93, a 1-bit determination flag (bit depth conversion prediction flag) is encoded as a zero value of the difference signal (= the difference signal between the upper layer prediction signal and the input signal). If this determination flag (bit depth conversion prediction flag) is not used, it is necessary to give, as encoded information, that the quantization value is zero for all the conversion coefficients of the difference signal. On the other hand, the present method using the determination flag (bit depth conversion prediction flag) can express the same in 1 bit. For this reason, the amount of codes at the time of encoding can be reduced.

  Note that the above-described video encoding device and video decoding device may have a computer system therein. In this case, the process performed by the above-described intra encoder and intra decoder is stored in a computer-readable recording medium in the form of a program, and the program is read out and executed by the computer. Will be done. 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.

DESCRIPTION OF SYMBOLS 1 Intra coding system designation | designated information encoding part 2 Intra coding system setting part 3-1 to 3-M Intra coding part 10 Intra coding system designation | designated information decoding part 11 Intra decoding system setting part 12-1 to 12-M Intra decoding unit 60 bit depth conversion prediction encoding unit 71 bit depth conversion amount setting unit 72 bit depth conversion unit 73 lower layer encoding unit 74 lower layer decoding unit 75 inverse bit depth conversion unit 76 bit depth conversion prediction determination unit 77 switch unit 78 Upper layer encoding unit 79 Subtractor 80 Encoded stream multiplexing processing unit 90 Bit depth conversion prediction decoding unit 91 Encoded stream separation processing unit 92 Bit depth conversion amount decoding unit 93 Lower layer decoding unit 94 Inverse bit depth conversion unit 95 Bit depth conversion prediction bit decoding unit 96 Bit depth conversion prediction determination unit 97 Switch unit 98 Hierarchical decoding unit 99 adder

Claims (21)

An input video signal to be image-encoded is read, and the input video signal is encoded using each of first intra coding using bit depth conversion and second intra coding not using bit depth conversion. A first step of calculating a cost at the time of using a cost function based on a weighted sum of a code amount and encoding distortion, and selecting an intra encoding corresponding to a cost with the lowest calculated cost;
A second step of encoding the input video signal by the first intra encoding;
A third step of encoding the input video signal by the second intra encoding;
In the first step, when the first intra encoding is selected, the input video signal encoded in the second step and information indicating the first intra encoding are output as an encoded stream, When the second intra coding is selected, a fourth step of outputting the input video signal encoded in the third step and information indicating the second intra coding as an encoded stream is included. An image encoding method characterized by the above. The second step includes
A fifth step of converting the input video signal based on a bit depth conversion amount determined according to the input video signal;
A sixth step of generating a first encoded signal by performing a first encoding on the input video signal subjected to the bit depth conversion;
A seventh step of generating a decoded signal from the first encoded signal by decoding corresponding to the first encoding;
An eighth step of converting the bit depth of the decoded signal to the same bit depth as the input video signal by inverse bit depth conversion;
The difference between the inverse bit depth converted decoded signal and the input video signal is calculated, and a second encoding is performed on the difference signal indicating the calculated difference to generate a second encoded signal, or the difference Selecting whether to omit the second encoding of the signal, and outputting the first encoded signal and the second encoded signal or the first encoded signal as the encoded stream. The image encoding method according to claim 1. The bit depth conversion amount is
It is a predetermined number of bits according to the input video signal, or a bit depth reduction bit number that minimizes the cost function when bit depth conversion is performed on the input video signal. The image encoding method according to claim 2, wherein: 4. The method according to claim 2, wherein the bit depth conversion amount and determination information indicating presence / absence of second encoding of the difference signal are set for each frame included in the input video signal. An image encoding method according to claim 1. The bit depth conversion amount and determination information indicating the presence or absence of the second encoding of the difference signal are further set for each predetermined region of each frame included in the input video signal. The image encoding method according to claim 4. The image according to any one of claims 2 to 5, wherein when the input video signal includes a plurality of color channels, the bit depth conversion amount is set for each color channel. Encoding method. A first step of reading an encoded stream to be decoded and decoding information indicating intra encoding obtained by encoding the encoded stream;
A second step of decoding the encoded stream by a first intra decoding method using inverse bit depth conversion;
A third step of decoding the encoded stream by a second intra decoding method that does not use inverse bit depth conversion;
And a fourth step of outputting the decoded signal decoded in one of the first step or the second step as a decoded image signal based on the information indicating the decoded intra coding. An image decoding method. The second step includes
A fifth step of separating data in the encoded stream;
A sixth step of decoding a bit depth conversion amount from the separated data;
A seventh step of performing a first decoding on the separated data to generate a first decoded signal;
An eighth step of performing inverse bit depth conversion for increasing the number of bits corresponding to the bit depth conversion amount for the first decoded signal;
A ninth step of decoding determination information indicating the presence or absence of a differential signal as a second encoded signal from the separated data;
Based on the determination information, it is selected whether to perform second decoding on the separated data to generate a second decoded signal, or to omit the second decoding, and after the inverse bit depth conversion And a tenth step of outputting either a difference signal between the decoded signal and the second decoded signal or the decoded signal after the inverse bit depth conversion as a decoded video signal. The image decoding method as described. The image decoding method according to claim 8, wherein the bit depth conversion amount and the determination information are decoded for each frame included in the decoded video signal. The image decoding method according to claim 9, wherein the bit depth conversion amount and the determination information are further decoded for each predetermined region of each frame included in the decoded video signal. When a plurality of color channels are included in the encoded stream to be decoded,
The image decoding method according to any one of claims 8 to 10, wherein the bit depth conversion amount is set for each color channel. An input video signal to be image-encoded is read, and the input video signal is encoded using each of first intra coding using bit depth conversion and second intra coding not using bit depth conversion. Calculating means using a cost function by a weighted sum of code amount and encoding distortion, and selecting means for selecting an intra encoding corresponding to a cost with the calculated cost being the smallest,
First encoding means for encoding the input video signal by the first intra encoding;
Second encoding means for encoding the input video signal by the second intra encoding;
When the selection unit selects the first intra encoding, the input video signal encoded by the first encoding unit and information indicating the first intra encoding are output as an encoded stream, When the second intra encoding is selected, the input video signal encoded by the second encoding means and output means for outputting information indicating the second intra encoding as an encoded stream. An image encoding device. The first encoding means includes
Bit depth conversion means for converting the input video signal based on a bit depth conversion amount set according to the input video signal;
First encoding means for generating a first encoded signal by performing first encoding on the input video signal having undergone the bit depth conversion;
Decoding means for generating a decoded signal from the first encoded signal by decoding corresponding to the first encoding;
Reverse bit depth conversion means for converting the bit depth of the decoded signal to the same bit depth as the input video signal by reverse bit depth conversion;
The difference between the inverse bit depth converted decoded signal and the input video signal is calculated, and a second encoding is performed on the difference signal indicating the calculated difference to generate a second encoded signal, or the difference Switching means for selecting whether to omit the second encoding for the signal, and outputting the first encoded signal and the second encoded signal, or the first encoded signal as the encoded stream. The image encoding device according to claim 12. The bit depth conversion amount and the determination information indicating the presence or absence of the second encoding of the difference signal are determined for each frame included in the input video signal or in a predetermined area in the frame included in the input video signal. The image encoding apparatus according to claim 13, further comprising a setting unit that sets the setting for each. The bit depth conversion means, when the input video signal includes a plurality of color channels,
The image coding apparatus according to claim 13, wherein the bit depth conversion amount is set for each color channel. An encoding method restoration unit that reads an encoded stream to be decoded and decodes information indicating intra encoding obtained by encoding the encoded stream;
First intra decoding means for decoding the encoded stream by a first intra decoding method using inverse bit depth conversion;
Second intra decoding means for decoding the encoded stream by a second intra decoding method that does not use inverse bit depth conversion;
Switching means for outputting, as a decoded image signal, a decoded signal decoded by either the first intra decoding means or the second intra decoding means based on the decoded intra-encoding information. An image decoding apparatus characterized by that. The first intra decoding means includes
Separating means for separating data in the encoded stream;
Bit depth conversion amount decoding means for decoding the bit depth conversion amount from the data separated by the separation means;
First decoding means for performing first decoding on the data separated by the separation means to generate a first decoded signal;
Reverse bit depth conversion means for performing reverse bit depth conversion for increasing the number of bits corresponding to the bit depth conversion amount for the first decoded signal;
Determination information decoding means for decoding determination information indicating the presence or absence of a differential signal as second encoded data from the separated data;
Based on the determination information, it is selected whether to perform second decoding on the separated data to generate a second decoded signal, or to omit the second decoding, and after the inverse bit depth conversion The output means which outputs either the difference signal of the decoded signal of this and the said 2nd decoded signal, or the decoded signal after the said reverse bit depth conversion as a decoded video signal is provided. Image decoding apparatus. The bit depth conversion amount and determination information indicating the presence or absence of the second decoding are set for each frame included in the decoded video signal or for each predetermined region in each frame included in the decoded video signal. The image decoding device according to claim 17, wherein: When a plurality of color channels are included in the encoded stream to be decoded,
The image decoding device according to claim 17 or 18, wherein the bit depth conversion amount is set for each color channel. In a computer provided in an image encoding device that encodes an input video signal,
An input video signal to be image-encoded is read, and the input video signal is encoded using each of first intra coding using bit depth conversion and second intra coding not using bit depth conversion. A first step of calculating a cost at the time of using a cost function based on a weighted sum of a code amount and encoding distortion, and selecting an intra encoding corresponding to a cost with the lowest calculated cost;
A second step of encoding the input video signal by the first intra encoding;
A third step of encoding the input video signal by the second intra encoding;
In the first step, when the first intra encoding is selected, the input video signal encoded in the second step and information indicating the first intra encoding are output as an encoded stream, When the second intra encoding is selected, the fourth step of executing the input video signal encoded in the third step and the information indicating the second intra encoding as an encoded stream is executed. Program for. In a computer provided in an image decoding apparatus for decoding an encoded stream,
A first step of reading an encoded stream to be decoded and decoding information indicating intra encoding obtained by encoding the encoded stream;
A second step of decoding the encoded stream by a first intra decoding method using inverse bit depth conversion;
A third step of decoding the encoded stream by a second intra decoding method that does not use inverse bit depth conversion;
And a fourth step of outputting the decoded signal decoded by either the first step or the second step as a decoded image signal based on the decoded intra-encoding information. program.

Images