JP6357133B2 - Intra prediction processing device, intra prediction processing method, intra prediction processing program, image encoding device, and image decoding device - Google Patents

Description

  The present invention relates to a technique for encoding or decoding an image signal using intra prediction (intra prediction).

  In the international video coding standard HEVC (High Efficiency Video Coding, that is, ITU-T H.265 or ISO / IEC 23008-2), in order to efficiently compress a large amount of video information, two types of processing, prediction and conversion, are performed. Is given. The prediction is divided into two types: inter coding (also referred to as inter-frame prediction coding and inter-screen coding) and intra coding (also referred to as intra-frame prediction coding and intra-screen coding). Inter-frame predictive coding is a method for compressing information by using temporal correlation in a video. A typical example is inter-frame prediction using motion compensation.

  On the other hand, intra-frame predictive encoding is a method of generating a prediction signal using decoded pixels around a block to be encoded, such as adopted in HEVC, that is, information compression using intra-frame correlation. It is a technique to plan. Regarding the conversion, JPEG, which is an international standard for still image coding, and MPEG-2, which is an international standard for video coding, employ a method called Discrete Cosine Transform (DCT). JPEG2000, which is a successor to JPEG, employs a technique called discrete wavelet transform (DWT). After prediction between the frames and within the frame, the prediction residual signal (also referred to as a prediction error signal) undergoes the transformation and quantization, and finally becomes a binary signal (bit stream) by entropy coding. .

  In HEVC, encoding is performed in units of blocks called coding units (CU), and the size of 64 × 64 is normally set as the maximum CU size in HEVC reference software (HEVC test Model: HM). ing. A 64 × 64 CU is divided on a quadtree basis, and an 8 × 8 CU is allowed as a minimum unit. Each CU is divided into a block that performs prediction called a prediction unit (PU) and a block that performs transformation called a transform unit (TU). These PU and TU are defined independently in the CU. A PU is basically divided on a quadtree basis like a CU, but there are tools that apply divisions other than squares. A tool that allows non-square partitioning of a PU is called Asymmetric Motion Partition (AMP).

  FIG. 13 is a diagram showing the definition of each processing unit in HEVC. In AVC (Advanced Video Coding, that is, ITU-T H.264 or ISO / IEC 14496-10), which is a standard, 16 × 16 size called macro block (MB) is the maximum unit of encoding processing. However, in HEVC, encoding is performed in units of larger blocks, such as 64 × 64 blocks. Such processing of encoding, prediction, and conversion in a larger size greatly contributes to improvement of encoding efficiency particularly in high-resolution video.

  In intra coding in HEVC, a new prediction mode and the like are added in order to improve the performance of AVC intra coding. One is a prediction mode called Planar prediction, and the other is a prediction mode (Angular Prediction) that realizes a finer prediction direction. FIG. 14 is a diagram illustrating a correspondence relationship between intra prediction mode numbers and intra prediction modes in a HEVC prediction unit (common to all block sizes). FIG. 15 is a diagram illustrating a correspondence relationship between the angular prediction mode and the angle. While AVC has 8 directions, HEVC has 33 directions, which enables very flexible prediction and contributes to improved coding efficiency.

  In Planar prediction, the prediction signal is predSamples [x] [y] (x and y are the coordinates of the prediction target block, the pixel position on the upper left is equivalent to x = 0 and y = 0), the block size is nT, and the reference signal P [x] [y] (where x and y are the position of the reference pixel, and the pixel position at the upper left of the prediction target block is equivalent to x = 0 and y = 0).

  predSamples [x] [y] = ((nT-1-x) * p [-1] [y] + (x + 1) * p [nT] [-1] + (nT-1-y) * p [x ] [− 1] + (y + 1) × p [−1] [nT] + nT) >> (Log2 (nT) +1) (1)

  In particular, it is possible to generate a prediction signal flexibly using pixels located in the upper right (p [nT] [-1]) and lower left (p [-1] [nT]), and HEVC intra coding. , The selectivity is high.

The Angular prediction generates a prediction signal predSamples [x] [y] (x and y are the coordinates of the prediction target block, and the pixel position at the top left is x = 0 and y = 0) as follows.
(A) When the prediction mode number is 18 or more (hereinafter,... Indicates a numerical range)
1. Reference pixel array ref [x] (x = −nT 2 × nT, where nT is a block size)
ref [x] = p [-1 + x] [-1] (x = 0... nT)
When the angle (intraPredAngle) corresponding to the prediction mode number defined in FIG. 16 is smaller than 0 and (nT × intraPredAngle) >> 5 is smaller than −1:
ref [x] = p [−1] [− 1 + ((x × invAngle + 128) >> 8)] (x = −1... (nT × intraPredAngle) >> 5)
Here, the definition of invAngle is as shown in FIG.
FIG. 16 is a diagram illustrating a correspondence relationship between the prediction mode and the angle. FIG. 17 is a diagram illustrating the correspondence between prediction modes and parameters.
Other than the above:
ref [x] = p [−1 + x] [− 1] (x = nT + 1... 2 × nT)

2. Reference pixel array ref [x] (x = 0... NT−1)
(A) An index (iIdx) and a multiplier parameter (iFact) are defined below.
iIdx = ((y + 1) × intraPredAngle) >> 5
iFact = ((y + 1) × intraPredAngle) & 31
(B) The following processing is performed according to the value of iFact.
When iFact is non-zero:
predSamples [x] [y] = ((32−iFact) × ref [x + iIdx + 1] + iFact × ref [x + iIdx + 2] +16) >> 5
When iFact is 0:
predSamples [x] [y] = ref [x + iIdx + 1]
(C) When the prediction mode number is 26 (vertical prediction): (x = 0, y = 0... NT-1 when the signal to be processed is a luminance signal and satisfies nT <32)
predSamples [x] [y] = Clip1 _Y (p [x] [-1] + ((p [-1] [y] -p [-1] [-1]) >> 1))

(B) When the prediction mode number is less than 18 Reference pixel array ref [x] (x = −nT... 2 × nT)
ref [x] = p [-1] [-1 + x] (x = 0... nT)
When intraPredAngle defined in FIG. 16 is smaller than 0 and (nT × intraPredAngle) >> 5 is smaller than −1:
ref [x] = p [−1 + ((x × invAngle + 128) >> 8)] [− 1] (x = −1... (nT × intraPredAngle) >> 5)
Other than the above:
ref [x] = p [−1] [− 1 + x] (x = nT + 1... 2 × nT)

2. Reference pixel array ref [x] (x = 0... NT−1)
(A) An index (iIdx) and a multiplier parameter (iFact) are defined below.
iIdx = ((x + 1) × intraPredAngle) >> 5
iFact = ((x + 1) × intraPredAngle) & 31
(B) The following processing is performed according to the value of iFact.
When iFact is non-zero:
predSamples [x] [y] = ((32−iFact) × ref [y + iIdx + 1] + iFact × ref [y + iIdx + 2] +16) >> 5
When iFact is 0:
predSamples [x] [y] = ref [y + iIdx + 1]
(C) When the prediction mode number is 10 (horizontal prediction): (x = 0 to nT−1, y = 0 when the signal to be processed is a luminance signal and satisfies nT <32)
predSamples [x] [y] = Clip1Y (p [-1] [y] + ((p [x] [-1] -p [-1] [-1]) >> 1)

  By these processes, it is possible to generate prediction signals in 33 directions and create a flexible prediction signal, so that the intra coding performance of HEVC is improved over that of AVC.

  The above detailed the Angular prediction. Angular prediction has the property of effectively reducing the prediction residual energy for a directional texture, but is not suitable for prediction of a flat region. For a flat region, DC prediction in which the block is filled with the same prediction value is effective. The DC prediction mode is adopted for AVC, and the corrected DC prediction is adopted for HEVC.

ここで、ｋ＝Ｌｏｇ２（ｎＴ）となる。続いて、色差を示すフラグｃＩｄｘに応じて、下記の処理を実施する。 In DC prediction, a predicted signal is predSamples [x] [y] (x and y are the coordinates of the block to be predicted, the pixel position at the upper left is equivalent to x = 0 and y = 0), the block size is nT, and the reference signal P [x] [y] (where x and y are the position of the reference pixel, and the pixel position at the upper left of the prediction target block corresponds to x = 0 and y = 0), is first defined by equation (2) A predicted value dcVal is calculated.

Here, k = Log2 (nT). Subsequently, the following processing is performed according to the flag cIdx indicating the color difference.

(C) When cIdx = 0, that is, when the target block is a luminance signal and nT <32 predSamples [0] [0] = (p [−1] [0] + 2 × dcVal + p [0] [− 1] +2) >> 2
predSamples [x] [0] = (p [x] [− 1] + 3 × dcVal + 2) >> 2 (x = 1... nT−1)
predSamples [0] [y] = (p [−1] [y] + 3 × dcVal + 2) >> 2 (y = 1... nT−1)
predSamples [x] [y] = dcVal (x, y = 1... nT-1)

(D) When cIdx ≠ 0, that is, when the target block is a color difference signal predSamples [x] [y] = dcVal (x, y = 0... NT−1)
When the target block is a luminance signal, as described in (C), the prediction signal adjacent to the reference signal is not the average value dcVal, but a weighted sum with the reference pixel is obtained, and the value is used as the prediction value. . This calculation of the weighted sum is not introduced in AVC, and is a process specific to HEVC.

  In the above, each intra prediction mode of HEVC and its prediction processing method have been described in detail. Next, reference pixel smoothing (a process called intra smoothing) performed before intra prediction will be described. This process is a process of applying a smoothing filter to the reference signal before intra prediction is executed. In the following, the block size is nT, the reference signal is p [x] [y] (x and y are the position of the reference pixel, and the upper left pixel position of the prediction target block is equivalent to x = 0 and y = 0), and after the filter Let the reference signal be pF [x] [y].

  First, a flag filterFlag related to the application of a filter is defined and set by the following procedure.

-If any of the following conditions is satisfied, set filterFlag to 0.
The prediction mode is DC prediction.
The block size is 4.

• If not, apply the following:
Set the variable minDistVerHor as follows:
Min (Abs (predModeIntra-26), Abs (predModeIntra-10))
The variable intraHorVerDistThres [nT] is defined as follows:
When nT = 8, 7 is substituted into the variable.
When nT = 16, 1 is substituted into the variable.
When nT = 32, 0 is substituted into the variable.
filterFlag is derived below.
If minDistVerHor is greater than intraHorVerDistThres [nT], filterFlag is set to 1.
Otherwise, set filterFlag to 0.

When filterFlag is 1, the following processing is executed.
The variable biIntFlag is derived as follows. Here, biIntFlag is a flag indicating whether or not the reference pixel setting mechanism based on linear interpolation called strong intra smoothing introduced in HEVC can be used.
If all the following conditions are met, biIntFlag is set to 1.
strong_intra_smoothing_enabled_flag is 1.
nT is 32.
Abs (p [-1] [-1] + p [nT.times.2-1] [-1] -2.times.p [nT-1] [-1]) <(1 << (BitDepth _Y- 5))
Abs (p [-1] [-1] + p [-1] [nT × 2-1] -2 × p [-1] [nT-1]) <(1 << (BitDepth _Y- 5))
If not, set biIntFlag to 0.

-Filter processing is executed as follows.
If biIntFlag is 1, at x = -1, y = -1 ... 63, and x = 0 ... 63, y = -1,
pF [-1] [-1] = p [-1] [-1]
pF [−1] [y] = ((63−y) × p [−1] [− 1] + (y + 1) × p [−1] [63] +32) >> 6 (y = 0... 62)
pF [-1] [63] = p [-1] [63]
pF [x] [− 1] = ((63−x) × p [−1] [− 1] + (x + 1) × p [63] [− 1] +32) >> 6 (x = 0... 62)
pF [63] [-1] = p [63] [-1]

If biIntFlag is 0, then x = -1, y = -1 ... nT × 2-1, and x = 0 ... nT × 2-1, y = -1.
pF [-1] [-1] = (p [-1] [0] + 2 × p [-1] [-1] + p [0] [-1] +2) >> 2
pF [−1] [y] = (p [−1] [y + 1] + 2 × p [−1] [y] + p [−1] [y−1] +2) >> 2 (y = 0... nT × 2-2)
pF [−1] [nT × 2-1] = p [−1] [nT × 2-1]
pF [x] [− 1] = (p [x−1] [− 1] + 2 × p [x] [− 1] + p [x + 1] [− 1] +2) >> 2 (x = 0... nT × 2-2)
pF [nT × 2-1] [− 1] = p [nT × 2-1] [− 1]

  The process of deriving the reference pixel by linear interpolation applied when biIntFlag = 1 is a strong intra-smoothing process, and the process of applying a smoothing filter of [1, 2, 1] / 4 to the reference pixel when biIntFlag = 0. Each is generally called an intra-smoothing process.

  Each item described above is described in detail in Non-Patent Document 1, Non-Patent Document 2, Non-Patent Document 3, and Non-Patent Document 4.

  In this specification, an image includes a general moving image, video, and still image.

Satoshi Okubo (Director), Shinya Tsuno, Yoshihiro Kikuchi, Teruhiko Suzuki (ed.), “Revised Third Edition H.264 / AVC Textbook”, Impress R & D, Inc., pp. 110-116, 2009 ITU-T: “SERIES H: AUDIOVISUAL AND MULTIMEDIA SYSTEMS Infrastructure of audiovisual services-Coding of moving video, Advanced video coding for generic audiovisual services”, Recommendation ITU-T H. 264, pp.116-135, 2005 ITU-T: “SERIES H: AUDIOVISUAL AND MULTIMEDIA SYSTEMS Infrastructure of audiovisual services-Coding of moving video, High efficiency video coding”, Recommendation ITU-T H. 265, pp.101-111, 2013 Satoshi Okubo (Director), Teruhiko Suzuki, Noriyuki Takamura, Ken Nakatsuji (ed.), “H.265 / HEVC Textbook”, Impress Japan, pp. 115-124, 2013

  In HEVC intra prediction, similar to AVC, there is a process of applying a smoothing filter to a reference pixel signal before generating an intra prediction signal, but filter application is based on information such as the intra prediction mode number and block size. A filter is adaptively performed by determining the presence or absence, and noise such as encoding distortion of the reference pixel signal is efficiently removed to improve encoding performance.

  However, the application of the intra smoothing filter is not determined in consideration of the encoding parameters related to the accuracy of the decoded reference pixels in addition to the intra prediction mode number and the block size. Since the reference pixel used for intra prediction is a decoded signal, coding distortion is superimposed. In particular, when a QP (Quantization Parameter) value, which is a quantization parameter, is small (that is, at a high bit rate), a code amount is assigned and degradation of a decoded signal is small. Conversely, when the QP value is large (that is, at a low bit rate), the code amount is not allocated, and the degradation of the decoded signal increases.

  In other words, when the QP value is small, the coding distortion of the decoded signal used for intra prediction is small, so the processing of the intra-smoothing filter is omitted while maintaining the coding performance by weakening the intra-smoothing or narrowing the application range. There is a possibility that processing speed can be increased. On the contrary, when the QP value is large, the coding distortion of the decoded signal becomes large. Therefore, it is expected that the encoding performance can be maintained or further improved by increasing the intra-smoothing or expanding the application range.

  The present invention has been made in view of such circumstances, and in the current intra prediction smoothing filter, solves the problem that the degree of degradation of the decoded signal due to the QP value is not applied to the determination of the presence or absence of the application. Intra-prediction processing device, intra-prediction processing method, intra-prediction processing program, image encoding device, and image capable of realizing improvement in encoding performance and subjective image quality, and speeding up of encoding processing and decoding processing at a high bit rate An object is to provide a decoding device.

One aspect of the present invention is an intra prediction processing apparatus that performs intra prediction for generating a prediction signal within the same screen, the mode information acquiring unit that acquires intra prediction mode information for identifying an intra prediction mode, and the intra Size information acquisition means for acquiring block size information for identifying a block size used for prediction, position information acquisition means for acquiring decoded pixel position information for setting a decoded pixel signal as a reference pixel signal, and the intra prediction mode information And deriving means for deriving filter usage information by determining whether or not an intra-smoothing filter to be applied to a reference pixel signal is applied based on the block size information, and a processing target block specified by the decoded pixel position information A parameter acquisition means for acquiring a quantization parameter; Comparing a setting means for setting read thresholds for the quantization parameters from Buru, and said to have been acquired the quantization parameter threshold, the intra prediction mode number to be applied intra smoothing filter when the quantization parameter is less And means for correcting the filter usage information so as to increase the number of intra prediction modes when the quantization parameter is large, and means for implementing the intra-smoothing filter based on the corrected filter usage information Is an intra prediction processing apparatus.

  One aspect of the present invention is the intra prediction processing apparatus, wherein the correction unit includes an adjustment unit that adjusts the number of intra prediction modes to which the intra smoothing filter is applied, using a plurality of the threshold values.

  One aspect of the present invention is an image encoding device including the intra prediction processing device, comprising: threshold transmission means for selecting and encoding the threshold that minimizes a prediction error or encoding cost and transmitting the selected threshold to the image decoding device. An image encoding device provided.

  One aspect of the present invention is an image decoding apparatus having the intra prediction processing apparatus, wherein threshold values are set by decoding and setting the threshold values selected and encoded so as to minimize prediction errors or encoding costs. An image decoding device comprising means.

One aspect of the present invention is an intra prediction processing method performed by an intra prediction processing apparatus that performs intra prediction that generates a prediction signal within the same screen, and mode information that acquires intra prediction mode information that identifies an intra prediction mode. An obtaining step; a size information obtaining step for obtaining block size information for identifying a block size used for the intra prediction; and a position information obtaining step for obtaining decoded pixel position information for setting a decoded pixel signal as a reference pixel signal; , A derivation step for deriving filter use information by determining whether an intra smoothing filter to be applied to a reference pixel signal is applied based on the intra prediction mode information and the block size information, and a process specified by the decoded pixel position information Quantization parameters used for the target block A parameter acquiring step of, comparing a setting step of setting a predetermined table by reading the threshold value for the quantization parameter, and said to have been acquired the quantization parameter threshold, when the quantization parameter is small intra A correction step of correcting the filter usage information so as to reduce the number of intra prediction modes to which a smoothing filter is applied and to increase the number of intra prediction modes when the quantization parameter is large, and based on the corrected filter usage information This is an intra prediction processing method having an execution step of executing the intra smoothing filter.

  One aspect of the present invention is the intra prediction processing method, wherein the correction step includes an adjustment step of adjusting the number of intra prediction modes to which the intra smoothing filter is applied using a plurality of the threshold values.

  One aspect of the present invention is an intra prediction processing program for causing a computer to function as the intra prediction processing device.

  According to the present invention, in intra prediction, noise such as coding distortion superimposed on a decoded reference pixel signal is applied by applying the influence of the decoded reference pixel signal based on the QP value to the determination of performing the intra smoothing process. Can be efficiently deleted to improve coding efficiency and subjective image quality. In addition, the effect of speeding up the encoding process and the decoding process can be obtained by omitting the intra smoothing filter at a high bit rate.

It is a block diagram which shows the structure of the image coding apparatus by one Embodiment of this invention. It is a block diagram which shows the structure of the image decoding apparatus by one Embodiment of this invention. FIG. 3 is a block diagram illustrating a configuration example of an intra prediction processing unit 101 and an intra prediction processing unit 202 illustrated in FIGS. 1 and 2. It is a flowchart which shows the operation | movement of the intra prediction process which the intra prediction process part 101 shown in FIG. 3 performs. FIG. 4 is a block diagram illustrating a configuration of a reference pixel generation unit 302 illustrated in FIG. 3. It is a figure which shows the conditions where the intra smoothing filter in HEVC is applied. It is explanatory drawing which shows the correction method of the filter use information in the intra smoothing filter application range correction part 405. FIG. It is explanatory drawing which shows the correction method of the filter use information in the intra smoothing filter application range correction part 405. FIG. It is a block diagram which shows the structure of the reference pixel production | generation part 302 'by a prior art. It is a flowchart which shows the reference pixel production | generation processing operation which the reference pixel production | generation part 302 shown in FIG. 5 performs. It is a figure which shows the hardware structural example of a computer in the case of comprising the image coding apparatus 100 provided with the intra prediction process part 101 by a computer and a software program. It is a figure which shows the hardware structural example of a computer in the case of comprising the image decoding apparatus 200 provided with the intra prediction process part 202 with a computer and a software program. It is a figure which shows the definition of each processing unit in HEVC. It is a figure which shows the correspondence of the intra prediction mode number and intra prediction mode in the prediction unit of HEVC. It is a figure which shows the correspondence of Angular prediction mode and an angle. It is a figure which shows the correspondence of prediction mode and an angle. It is a figure which shows the correspondence of prediction mode and a parameter.

  Hereinafter, an image encoding device and an image decoding device according to an embodiment of the present invention will be described with reference to the drawings. FIG. 1 is a block diagram showing a configuration of an image encoding device according to the embodiment. The image encoding device 100 receives an input video signal to be encoded, divides each frame of the input video signal into a plurality of blocks, encodes each block, and outputs a bit stream of the encoding result as an encoded stream To do.

  As shown in FIG. 1, the image encoding device 100 includes an intra prediction processing unit 101, an inter prediction processing unit 102, a prediction residual signal generation unit 103, a transform processing unit 104, a quantization processing unit 105, and an inverse quantization processing unit. 106, an inverse transform processing unit 107, a decoded signal generation unit 108, a frame memory 109, an in-loop filter processing unit 110, an inter prediction information storage unit 111, an intra prediction information storage unit 112, and an entropy encoding processing unit 113.

  In the image encoding device 100 illustrated in FIG. 1, the intra prediction processing unit 101 has a characteristic configuration. The other functional units are equivalent to the functional units included in a general image encoding device used as an encoder such as HEVC. Note that a plurality of in-loop filters may be applied to the in-loop filter processing unit 110. For example, as in HEVC, a deblocking filter and a sample adaptive offset (SAO) are applied. A filter for removing coding distortion called an adaptive loop filter (ALF) proposed in HEVC standardization may be inserted.

  The intra prediction processing unit 101 receives an input video signal and generates a prediction signal based on the input video signal. Also, the intra prediction processing unit 101 generates intra prediction information including a prediction mode based on the input video signal, and stores the generated intra prediction information in the intra prediction information storage unit 112 for storage.

  The inter prediction processing unit 102 inputs an input video signal and a reference image output from the in-loop filter processing unit 110. The inter prediction processing unit 102 generates a prediction signal based on the input video signal and the reference image. Further, the inter prediction processing unit 102 generates inter prediction information including a motion vector based on the input video signal and the reference signal, and stores the generated inter prediction information in the inter prediction information storage unit 111 for storage.

  The prediction residual signal generation unit 103 calculates a difference between the input video signal and a prediction signal output from the intra prediction processing unit 101 or the inter prediction processing unit 102. The prediction residual signal generation unit 103 outputs the calculated difference to the conversion processing unit 104 as a prediction residual signal.

  The transform processing unit 104 performs orthogonal transform such as discrete cosine transform (DCT) on the prediction residual signal input from the prediction residual signal generation unit 103. The transform processing unit 104 outputs the transform coefficient obtained by the orthogonal transform to the quantization processing unit 105.

  The quantization processing unit 105 quantizes the transform coefficient input from the transform processing unit 104 and outputs the quantized transform coefficient to the inverse quantization processing unit 106 and the entropy coding processing unit 113.

  The inverse quantization processing unit 106 performs inverse quantization on the transform coefficient input from the quantization processing unit 105 and outputs the result to the inverse transform processing unit 107. The inverse transform processing unit 107 performs inverse orthogonal transform on the transform coefficient input from the inverse quantization processing unit 106. The inverse transform processing unit 107 outputs the prediction residual decoded signal obtained by the inverse orthogonal transform to the decoded signal generating unit 108.

  The decoded signal generation unit 108 adds the prediction residual decoded signal input from the inverse transform processing unit 107 and the prediction signal output from the intra prediction processing unit 101 or the inter prediction processing unit 102. The decoded signal generation unit 108 stores the addition result in the frame memory 109 as a decoded signal of an encoding target block obtained by encoding the addition result. This decoded signal is used as a reference image in the intra prediction processing unit 101 or the inter prediction processing unit 102.

  The frame memory 109 stores the decoded signal generated and output by the decoded signal generation unit 108. The in-loop filter processing unit 110 reads the decoded signal stored in the frame memory 109, and performs a deblocking filter, an SAO process, and the like on the read decoded signal. The in-loop filter processing unit 110 outputs the image after various in-loop filter processes to the inter prediction processing unit 102 as a reference image.

  The inter prediction information storage unit 111 stores the inter prediction information generated by the inter prediction processing unit 102. The intra prediction information storage unit 112 stores the intra prediction information generated by the intra prediction processing unit 101.

  The entropy encoding processing unit 113 includes the transform coefficient quantized by the quantization processing unit 105, the inter prediction information stored in the inter prediction information storage unit 111, and the intra stored in the intra prediction information storage unit 112. The prediction information is entropy encoded and output as an encoded stream.

  The image encoding apparatus 100 includes the above-described functional units, and is obtained by dividing each frame of the input video signal into a plurality of blocks, performing block-based predictive encoding, and encoding the input video signal. Output the encoded stream.

  FIG. 2 is a block diagram showing the configuration of the image decoding apparatus 200 in the embodiment. The image decoding apparatus 200 receives an encoded stream that is encoded and output by the image encoding apparatus 100 shown in FIG. 1, and outputs a decoded video signal that is a decoded image by decoding the encoded stream. .

  As shown in FIG. 2, the image decoding apparatus 200 includes an entropy decoding processing unit 201, an intra prediction processing unit 202, an inter prediction processing unit 203, an inverse quantization processing unit 204, an inverse transformation processing unit 205, a decoded signal generation unit 206, A frame memory 207, an in-loop filter processing unit 208, an inter prediction information storage unit 209, and an intra prediction information storage unit 210 are provided.

  In the image decoding apparatus 200 illustrated in FIG. 2, the intra prediction processing unit 202 has a characteristic configuration. The other functional units are equivalent to the functional units included in a general image decoding device used as a decoder such as HEVC.

  The entropy decoding processing unit 201 inputs an encoded stream, entropy-decodes the quantized transform coefficient of the decoding target block from the input encoded stream, and entropy-decodes intra prediction information related to intra prediction and inter prediction information related to inter prediction. To do. The entropy decoding processing unit 201 outputs the quantized transform coefficient to the inverse quantization processing unit 204, stores the inter prediction information in the inter prediction information storage unit 209, and stores the intra prediction information in the intra prediction information storage unit. It memorize | stores in 210 and stores.

  The intra prediction processing unit 202 reads the decoded signal stored in the frame memory 207 as a reference image. The intra prediction processing unit 202 reads the intra prediction information from the intra prediction information storage unit 210. Then, the intra prediction processing unit 202 generates a prediction signal based on the read reference image and the read intra prediction information.

  The inter prediction processing unit 203 reads inter prediction information from the inter prediction information storage unit 209. Then, the inter prediction processing unit 203 generates a prediction signal based on the inter prediction information and the reference image input from the in-loop filter processing unit 208.

  The inverse quantization processing unit 204 inversely quantizes the quantized transform coefficient input from the entropy decoding processing unit 201 to calculate a decoded transform coefficient, and outputs the calculated decoded transform coefficient to the inverse transform processing unit 205.

  The inverse transform processing unit 205 performs inverse orthogonal transform on the decoded transform coefficient input from the inverse quantization processing unit 204, calculates a predicted residual decoded signal, and outputs the calculated predicted residual decoded signal to the decoded signal generating unit 206. To do.

  The decoded signal generation unit 206 adds the prediction residual decoded signal input from the inverse transform processing unit 205 and the prediction signal output from the intra prediction processing unit 202 or the inter prediction processing unit 203. The decoded signal generation unit 206 stores the addition result in the frame memory 207 as a decoded signal of the decoding target block. The frame memory 207 stores the decoded signal calculated by the decoded signal generation unit 206.

  The in-loop filter processing unit 208 reads out the decoded signal from the frame memory 207 and performs various in-loop filter processes for reducing coding distortion on the image indicated by the read-out decoded signal. The in-loop filter processing unit 208 outputs the image after various in-loop filter processes as a decoded video signal. Further, the in-loop filter processing unit 208 outputs the image after filter processing to the inter prediction processing unit 203 as a reference image. Similar to the image encoding device 100, the in-loop filter processing unit 208 may include a filter processing unit such as an adaptive loop filter.

  The inter prediction information storage unit 209 stores the inter prediction information decoded by the entropy decoding processing unit 201. The intra prediction information storage unit 210 stores the intra prediction information decoded by the entropy decoding processing unit 201.

  The image decoding apparatus 200 includes the above-described functional units, thereby inputting an encoded stream of video encoded by block-based predictive encoding, decoding the video from the input encoded stream, and decoding video. Output a signal.

  Next, the configuration of the intra prediction processing unit 101 and the intra prediction processing unit 202 illustrated in FIGS. 1 and 2 will be described with reference to FIG. FIG. 3 is a block diagram illustrating a configuration example of the intra prediction processing unit 101 and the intra prediction processing unit 202 illustrated in FIGS. 1 and 2. Since the intra prediction processing unit 101 and the intra prediction processing unit 202 perform common processing and have the same configuration, the intra prediction processing unit 101 will be described below, and the description of the intra prediction processing unit 202 will be omitted.

  As illustrated in FIG. 3, the intra prediction processing unit 101 includes a block position identification unit 301, a reference pixel generation unit 302, and an intra prediction signal generation unit 303. The block position identifying unit 301 identifies the position of a block to be intra-predicted as a processing target. The block position identification unit 301 receives block position information (spatial coordinates and CU / PU / TU position) as a position in a frame or a slice in order to identify a block position, and acquires a reference pixel. The decoded pixel position information is output.

  The reference pixel generation unit 302 receives the decoded pixel position information output from the block position identification unit 301, the decoded pixel signal, and the intra prediction mode. Based on the decoded pixel position information, the reference pixel generation unit 302 performs storage of reference pixels necessary for prediction signal generation, necessary filter processing, and the like. The reference pixel storage described here means, for example, preparation of the above-mentioned p [x] [y] and generation of the reference pixel array ref [x]. The necessary filter processing includes intra smoothing, that is, processing of applying a smoothing filter of [1, 2, 1] / 4 to the reference pixel and applying the pixel value after the filtering to the predicted value, and the above-described (A ) And (B). The filtering process etc. which take the ratio of the position on 32 divisions shown in (b) are shown. That is, all the filtering processing of the decoded pixel signal necessary for the prediction signal generation is performed by the reference pixel generation unit 302. The reference pixel generation unit 302 outputs the decoded pixel signal after various filter processes to the intra prediction signal generation unit 303 as a reference pixel signal for copying to the array of prediction signals.

  The intra prediction signal generation unit 303 generates and outputs an intra prediction signal based on the reference pixel signal, the decoded pixel signal, and the intra prediction mode input from the reference pixel generation unit 302.

  As described above, the intra prediction processing unit 101 uses the block position identification unit 301, the reference pixel generation unit 302, and the intra prediction signal generation unit 303 based on the block position information, the decoded pixel signal, and the intra prediction mode that are input. Output a prediction signal. In the present invention, the reference pixel generation unit 302 is a characteristic part different from the conventional method.

  Next, with reference to FIG. 4, the processing operation of the intra prediction processing unit 101 shown in FIG. 3 will be described. FIG. 4 is a flowchart showing the operation of the intra prediction process performed by the intra prediction processing unit 101 shown in FIG. The block position identification unit 301 identifies the position of the block to be intra-predicted as a processing target based on the input block position information according to the procedure described in Non-Patent Document 3 and the like, and is necessary for prediction Information indicating the position of the decoded pixel for pixel generation is output (step S101).

  Next, based on the decoded pixel position information output from the block position identifying unit 301, the reference pixel generation unit 302 performs a process of storing the decoded pixel signal as an array, a filter process applied to the decoded pixel, and the like (step S102). . Subsequently, the intra prediction signal generation unit 303 generates and outputs an intra prediction signal based on the reference pixel signal output from the reference pixel generation unit 302 and the intra prediction mode and the decoded pixel signal that are separately input (step S103). ).

  Next, the configuration of the reference pixel generation unit 302 shown in FIG. 3 will be described with reference to FIG. FIG. 5 is a block diagram illustrating a configuration of the reference pixel generation unit 302 illustrated in FIG. As shown in FIG. 5, the reference pixel generation unit 302 includes an intra prediction mode storage unit 401, a decoded pixel position information setting unit 402, a block size acquisition unit 403, an intra smoothing filter use determination unit 404, and an intra smoothing filter application range correction unit. 405, an intra smoothing filter execution unit 406 and a QP threshold setting unit 407 are provided.

  The intra prediction mode storage unit 401 stores the prediction mode of the target block, and provides intra prediction mode information to the intra smoothing filter use determination unit 404.

  The decoded pixel position information setting unit 402 stores the decoded pixel position information given from the block position identifying unit 301 and uses the composite pixel position information for determining the position of the reference pixel to be used for intra prediction using an intra smoothing filter. Output to the determination unit 404.

  The block size acquisition unit 403 acquires the block size of the target block and outputs the block size to the intra smoothing filter use determination unit 404.

  Using the intra prediction mode information using the intra prediction mode information provided from the intra prediction mode storage unit 401, the decoded pixel position information provided from the decoded pixel position information setting unit 402, and the block size provided from the block size acquisition unit 403 The determination unit 404 determines whether to apply the intra-smoothing filter, that is, filterFlag. In addition, a necessary decoded pixel signal is acquired in order to perform the use determination of strong intra-smoothing, that is, the determination of biIntraFlag. The information on whether or not to use these smoothing filters and strong smoothing filters is defined as filter use information. The filter usage information is output to the intra smoothing filter determination correction unit 405.

  The QP threshold value setting unit 407 reads threshold information related to QP that is necessary for determining whether or not to use intra smoothing. The numerical value is output to the intra smoothing filter application range correction unit 405 as a QP threshold value.

  The intra smoothing filter application range correction unit 405 receives the filter usage information output from the intra smoothing filter use determination unit 404 and the QP threshold obtained from the QP threshold setting unit 407. Considering the coding distortion of the decoded signal from the QP threshold value, it is determined whether or not the QP threshold value is exceeded, and the necessity of intra-smoothing is determined again. A specific processing method will be described later. Here, the filter usage information is corrected, and the filter usage correction information is output to the intra smoothing filter execution unit 406. Here, if the result of the re-determination is not corrected, the input filter usage information can be output without change.

  The intra smoothing filter execution unit 406 executes a smoothing filter such as an intra smoothing filter based on the filter use correction information received from the intra smoothing filter determination correction unit 405, and removes reference pixel noise to use for intra prediction. For this purpose, a final reference pixel signal is output.

  FIG. 6 is a diagram illustrating conditions under which an intra smoothing filter is applied in HEVC. When the 4 × 4 size is selected, no intra smoothing filter is applied regardless of which prediction mode is selected. In the case of 8 × 8, an intra smoothing filter is applied to four types of modes 0, 2, 18, and 34. In 16 × 16 and 32 × 32, an intra smoothing filter is applied to more prediction modes. In the case of HEVC, the table shown in FIG. 6 to which this intra smoothing filter is applied is used even if the QP value varies.

  On the other hand, the conditions to which the intra smoothing filter in the present invention is applied are shown in FIG. Hereinafter, the correction method of the filter use information in the intra smoothing filter application range correction unit 405 will be described with reference to FIG. FIG. 7 is an explanatory diagram illustrating a method of correcting the filter use information in the intra smoothing filter application range correction unit 405. As described above, the application range of the intra smoothing filter is switched according to the QP value. For example, as shown in the upper table of FIG. 7, when the QP value is smaller than a specific threshold, it is assumed that a large amount of coding distortion is not superimposed on the reference pixel used for intra prediction, and the application range of the intra smoothing filter Reduce. Conversely, as shown in the lower table of FIG. 7, when the QP value is equal to or greater than a specific threshold, it is assumed that a large amount of encoding distortion is superimposed on the reference pixel, and the application range of the intra smoothing filter is Expanding. The table in FIG. 7 is merely an example, and it suffices if the number of intra prediction modes to which the intra smoothing filter is applied fluctuates if there is a tendency to expand and contract.

  Although the case where there is one threshold has been described in FIG. 7, there may be a plurality of thresholds. For example, a case where a plurality of thresholds are prepared will be described with reference to FIG. Two thresholds are assumed as T1 and T2, and each satisfies T1> T2. Considering the possible range of the QP value, in the case of HEVC, 51 ≧ T1> T2> 0. Here, if the QP value applied to the target block is equal to or greater than T1, a pattern that realizes the maximum applicable range is set. If the QP value is less than T1 and equal to or greater than T2, the same applicable range as in the conventional case is set. And if it is less than T2, the still smaller application range will be set than before. This application is also an example. For example, a setting example that does not depend on the prior art may be applied. Further, the threshold values may be set more finely, such as five types.

  As for the threshold value, whether it is one type or a plurality of types, it is not necessary to transmit the threshold value if a common table is held in the memory or the like by the encoder and the decoder. What is necessary is just to set in the range which QP value can take. Since an appropriate threshold varies depending on the nature of the video, the threshold may be changed, for example, in a specific scene unit of the video, a group of pictures (GOP) unit, a slice unit, or a frame unit. In that case, it is necessary to transmit in units of switching threshold values. Although the amount of code is generated as overhead, the threshold value can be switched in an appropriate unit, and there is a possibility that the coding performance and subjective image quality can be further improved as compared with the case where transmission is not performed. Since the effect due to the overhead generated and the granularity of switching is a trade-off relationship, the transmission unit may be determined in consideration of the relationship.

  The reference pixel generation unit 302 in the present embodiment is characterized by having an intra-smoothing filter application range correction unit 405 and a QP threshold setting unit 407, which is not in the related art.

  FIG. 9 is a block diagram illustrating a configuration of a reference pixel generation unit 302 ′ according to the related art. The intra prediction mode storage unit 501, the decoded pixel position information setting unit 502, the block size acquisition unit 503, the intra smoothing filter use determination unit 504, and the intra smoothing filter execution unit 505 are each an intra prediction mode storage unit illustrated in FIG. 401, the decoded pixel position information setting unit 402, the block size acquisition unit 403, the intra smoothing filter use determination unit 404, and the intra smoothing filter execution unit 406 have the same functions. 9 does not have the intra-smoothing filter application range correction unit 405 and the QP threshold setting unit 407 in FIG. 5, and the filter usage information output from the intra-smoothing filter usage determination unit 504 is output as it is without modification or modification. The

  Next, the operation of the reference pixel generation unit 302 shown in FIG. 5 will be described with reference to FIG. FIG. 10 is a flowchart showing the reference pixel generation processing operation performed by the reference pixel generation unit 302 shown in FIG.

  In order to generate reference pixels for intra prediction signal derivation processing in the reference pixel generation unit 302, first, the intra smoothing filter use determination unit 404 reads the intra prediction mode stored in the intra prediction mode storage unit 401, The decoded pixel position information is read from the decoded pixel position information setting unit 402, and the block size of the target block is read and acquired from the block size acquisition unit 403 (step S201). In the case of a strong intra-smoothing filter, a decoded pixel signal is also used, so that is also obtained. Then, the intra-smoothing filter use determination unit 404 determines whether or not to use the intra-smoothing filter for the intra prediction of the target block from the three types of read information (step S202).

  Next, the intra smoothing filter application range correction unit 405 reads the QP value used for the target block (step S203). Further, the threshold for QP is read from the QP threshold setting unit 407 (step S204). Hereinafter, an example in the case of the table on the lower side of FIG. 7 is shown. The correction of the filter use information is determined by comparing the QP value obtained in step S203 with the threshold value obtained in step S204 (step S205). If the QP value is equal to or greater than the threshold (YES in step S205), it is determined that the reference pixel signal includes strong coding distortion, and a modification is made to change the intra smoothing filter to be used in a wider range. The filter use correction information is output to the intra smoothing filter execution unit 406 (step S206). If the QP value is less than the threshold value (NO in step S205), the filter usage information is output to the intra smoothing filter execution unit 406 without changing the filter usage information. The intra smoothing filter execution unit 406 executes the intra smoothing filter based on the filter use correction information from the intra smoothing filter application range correction unit 405 and outputs reference pixels used for intra prediction (step S207).

  As described above, the intra prediction processing units 101 and 202 described in the embodiment of the present invention, the image encoding device 100 including the intra prediction processing unit 101, and the image decoding device 200 including the intra prediction processing unit 202 are configured using a computer and a software program. And can be realized.

  FIG. 11 is a diagram illustrating a hardware configuration example of a computer when the image encoding device 100 including the intra prediction processing unit 101 is configured by a computer and a software program. The computer includes a CPU 900, a memory 901, a video signal input unit 902, a program storage device 903, and an encoded stream output unit 905.

  The CPU 900 executes a program. The memory 901 is a RAM or the like that temporarily stores programs and data accessed by the CPU 900. The video signal input unit 902 inputs an input video signal to be encoded from a device such as a camera that generates a video signal. The video signal input unit 902 may be a storage device that stores an input video signal from a hard disk device or the like. The program storage device 903 stores an image encoding program 904 that is a software program that causes the CPU 900 to execute the encoding process described in the above embodiments. The encoded stream output unit 905 outputs an encoded stream generated when the CPU 900 executes the image encoding program 904 loaded in the memory 901. The encoded stream output unit 905 may output the encoded stream via a network. The encoded stream output unit 905 may be a storage device that stores an encoded stream such as a hard disk device. The CPU 900, the memory 901, the video signal input unit 902, the program storage device 903, and the encoded stream output unit 905 are connected to each other via a bus.

  FIG. 12 is a diagram illustrating a hardware configuration example of a computer when the image decoding device 200 including the intra prediction processing unit 202 is configured by a computer and a software program. The computer includes a CPU 1000, a memory 1001, an encoded stream input unit 1002, a program storage device 1003, and a decoded video output unit 1005.

  The CPU 1000 executes a program. The memory 1001 is a RAM that temporarily stores programs and data accessed by the CPU 1000. The encoded stream input unit 1002 inputs, for example, an encoded stream obtained by encoding the input video signal by the image encoding apparatus 100 through the above-described processing. The encoded stream input unit 1002 may be a storage device that stores an encoded stream by a hard disk device or the like. The program storage device 1003 stores an image decoding program 1004 that is a software program that causes the CPU 1000 to execute the decoding processing described in each of the above-described embodiments. The decoded video output unit 1005 outputs the decoded video obtained by decoding the encoded stream to a playback device or the like by the CPU 1000 executing the image decoding program 1004 loaded in the memory 1001. The decoded video output unit 1005 may be a storage device that stores a decoded video signal from a hard disk device or the like. The CPU 1000, the memory 1001, the encoded stream input unit 1002, the program storage device 1003, and the decoded video output unit 1005 are connected to each other via a bus.

  You may make it implement | achieve all or one part of the intra prediction process part in embodiment mentioned above with a computer. In that case, a program for realizing this function may be recorded on a computer-readable recording medium, and the program recorded on this recording medium may be read into a computer system and executed. Here, the “computer system” includes an OS and hardware such as peripheral devices. The “computer-readable recording medium” refers to a storage device such as a flexible medium, a magneto-optical disk, a portable medium such as a ROM and a CD-ROM, and a hard disk incorporated in a computer system. Furthermore, the “computer-readable recording medium” dynamically holds a program for a short time like a communication line when transmitting a program via a network such as the Internet or a communication line such as a telephone line. In this case, a volatile memory inside a computer system serving as a server or a client in that case may be included and a program held for a certain period of time. Further, the program may be a program for realizing a part of the above-described functions, and may be a program capable of realizing the functions described above in combination with a program already recorded in a computer system. It may be realized using hardware such as PLD (Programmable Logic Device) or FPGA (Field Programmable Gate Array).

  As mentioned above, although embodiment of this invention has been described with reference to drawings, the said embodiment is only the illustration of this invention, and it is clear that this invention is not limited to the said embodiment. is there. Therefore, additions, omissions, substitutions, and other modifications of the components may be made without departing from the technical idea and scope of the present invention.

  In the intra prediction process, the QP value is added to the determination of the application of the intra smoothing filter to determine the magnitude of the coding distortion superimposed on the reference pixel signal, and the intra smoothing filter is appropriately switched to obtain the coding efficiency and The present invention can be applied to applications where it is essential to improve the subjective image quality or to speed up the encoding and decoding processes at a high bit rate.

  DESCRIPTION OF SYMBOLS 100 ... Image coding apparatus, 101 ... Intra prediction process part, 102 ... Inter prediction process part, 103 ... Prediction residual signal generation part, 104 ... Transformation process part, 105 ... Quantization process part, 106 ... Inverse quantization process part , 107 ... Inverse transformation processing unit, 108 ... Decoded signal generation unit, 109 ... Frame memory, 110 ... In-loop filter processing unit, 111 ... Inter prediction information storage unit, 112 ... Intra prediction information storage unit, 113 ... Entropy encoding process , 200 ... image decoding apparatus, 201 ... entropy decoding processing part, 202 ... intra prediction processing part, 203 ... inter prediction processing part, 204 ... inverse quantization processing part, 205 ... inverse transformation processing part, 206 ... decoded signal generation part 207: Frame memory 208: In-loop filter processing unit 209 ... Inter prediction information storage unit 210: Intra prediction Information storage unit 301... Block position identification unit 302. Reference pixel generation unit 303. Intra prediction signal generation unit 401 and 501 Intra prediction mode storage unit 402 and 502 Decoded pixel position information setting unit 403 and 503 ... block size acquisition unit, 404, 504 ... intrasmoothing filter use determination unit, 405 ... intrasmoothing filter application range correction unit, 406,505 ... intrasmoothing filter execution unit, 407 ... QP threshold setting unit, 900, 1000 ... CPU, 901, 1001 ... Memory, 902 ... Video signal input unit, 903, 1003 ... Program storage device, 904 ... Image encoding program, 905 ... Encoded stream output unit, 1002 ... Encoded stream input unit, 1004 ... Image decoding program, 1005 ... Decoded video output unit

Claims (7)

An intra prediction processing device that performs intra prediction that generates a prediction signal within the same screen,
Mode information acquisition means for acquiring intra prediction mode information for identifying the intra prediction mode;
Size information acquisition means for acquiring block size information for identifying a block size used for the intra prediction;
Position information acquisition means for acquiring decoded pixel position information for setting the decoded pixel signal as a reference pixel signal;
Derivation means for determining applicability of an intra smoothing filter to be applied to a reference pixel signal based on the intra prediction mode information and the block size information and deriving filter use information;
Parameter acquisition means for acquiring a quantization parameter used for a processing target block specified by the decoded pixel position information;
Setting means for reading and setting a threshold value for the quantization parameter from a predetermined table;
Comparing the acquired and the quantization parameter and the threshold value, the reduced number of intra prediction modes that apply intra smoothing filter when the quantization parameter is small, the number of the intra-prediction mode when the quantization parameter is large A correction means for correcting the filter usage information so as to increase;
An intra-prediction processing device comprising: an implementing unit that implements the intra-smoothing filter based on the modified filter use information. The correcting means is
The intra prediction processing apparatus of Claim 1 provided with the adjustment means which adjusts the number of intra prediction modes to which an intra smoothing filter is applied using the said some threshold value. An image encoding device comprising the intra prediction processing device according to claim 1 or 2,
An image encoding apparatus comprising threshold transmission means for selecting, encoding, and transmitting to the image decoding apparatus the threshold that minimizes a prediction error or encoding cost. An image decoding apparatus having the intra prediction processing apparatus according to claim 1 or 2,
An image decoding apparatus comprising threshold setting means for decoding and setting the threshold selected and encoded so as to minimize a prediction error or encoding cost. An intra-prediction processing method performed by an intra-prediction processing device that performs intra prediction that generates a prediction signal within the same screen,
A mode information acquisition step of acquiring intra prediction mode information for identifying an intra prediction mode;
A size information acquisition step of acquiring block size information for identifying a block size used for the intra prediction;
A position information acquisition step of acquiring decoded pixel position information for setting the decoded pixel signal as a reference pixel signal;
A derivation step of deriving filter usage information by determining applicability of an intra smoothing filter applied to a reference pixel signal based on the intra prediction mode information and the block size information;
A parameter acquisition step of acquiring a quantization parameter used for the processing target block specified by the decoded pixel position information;
A setting step of reading and setting a threshold value for the quantization parameter from a predetermined table;
Comparing the acquired and the quantization parameter and the threshold value, the reduced number of intra prediction modes that apply intra smoothing filter when the quantization parameter is small, the number of the intra-prediction mode when the quantization parameter is large A modification step to modify the filter usage information to increase;
An intra prediction processing method comprising: performing the intra smoothing filter based on the modified filter usage information. In the correction step,
The intra prediction processing method according to claim 5, further comprising an adjustment step of adjusting the number of intra prediction modes to which the intra smoothing filter is applied using a plurality of the threshold values.   An intra prediction processing program for causing a computer to function as the intra prediction processing device according to claim 1.

Images