JP6270293B1 - Moving picture coding apparatus and moving picture coding method - Google Patents

Abstract

Flicker generated in an encoded moving image is suppressed. A determination unit 241 for determining whether or not an encoding target block satisfies a first condition and a second condition, and for quantizing the encoding target block according to a result determined by the determination unit And a setting unit 242 that outputs a quantization parameter of the first, the first condition is that an index indicating a variation in luminance value of the encoding target block is smaller than a first threshold, and the second condition is The difference between the luminance value of the encoding target block and the luminance value of the block at the same position as the encoding target block in the picture at least one prior to the encoding target picture including the encoding target block is a second value. It is smaller than the threshold value. [Selection] Figure 2

Description

  The present invention relates to a moving image encoding apparatus and a moving image encoding method for encoding a moving image.

  MPEG (Moving Picture Expert Group) -1,2,4, H.264 / AVC, H.265 / HEVC, etc. have been established as international standards for moving picture coding technology. A video encoding device that employs these video encoding technologies divides an original image into blocks, which are the basic units of encoding, and performs motion compensated prediction and discrete cosine transform (DCT) for each block. By doing so, the code amount is compressed. A block that is a basic unit of encoding is called a macroblock in H.264 / AVC, and is called a coding unit (CU) in H.265 / HEVC.

  The amount of generated code when a moving image is encoded varies greatly depending on the characteristics of the moving image to be encoded. For example, when an encoded moving image is transmitted or stored in a recording medium, the moving image encoding device performs processing for suppressing the generated code amount to a desired value while maintaining good image quality, that is, the code amount It is necessary to control. As one of the code amount control algorithms, an MPEG-2 test model TM5 (Test Model 5) is known.

  In addition, a technique for improving the image quality in an area where deterioration is conspicuous visually by applying a TM5 code amount control algorithm has been proposed (Patent Documents 1 and 2).

JP 2009-200871 A JP 2011-172137 A

  In MT5, the quantization step is weighted based on the human visual characteristics to determine the final quantization step. When a moving image is encoded using a code amount control algorithm that employs such applied quantization, a flicker phenomenon called “FLICKER” occurs in a part of the reproduced video. May occur.

  For example, as shown in FIG. 1, when a non-fluctuating region A such as the ground exists in a part of an image obtained by shooting a landscape subject with movement such as a wave, a phenomenon in which the non-fluctuating region A appears to flicker occurs. There are things to do.

  The present invention has been made in view of the above problems, and an object of the present invention is to provide a moving image encoding apparatus and a moving image encoding method capable of suppressing flicker generated in an encoded moving image. .

In order to achieve the above object, the video encoding device of the present invention includes a determination unit that determines whether an encoding target block satisfies the first condition and the second condition, and a result determined by the determination unit. And a setting unit that outputs a quantization parameter for quantizing the encoding target block, and the first condition is that an index indicating a variation in luminance value of the encoding target block is a first threshold value The second condition is that the luminance value of the encoding target block and the same position as the encoding target block in the picture at least one prior to the encoding target picture including the encoding target block are The difference between the luminance value of the block is smaller than a second threshold value, the determination unit determines that the encoding target block satisfies the first condition and the second condition, and The encoding target picture containing-coding target block when it is determined that the I-picture, the setting unit outputs a quantization parameter of a predetermined fixed value.
The moving image encoding apparatus of the present invention includes: a determination unit that determines whether or not an encoding target block satisfies a first condition and a second condition; and the encoding target block according to a result determined by the determination unit And a setting unit that outputs a quantization parameter for quantizing the data, wherein the first condition is that an index indicating a variation in luminance value of the encoding target block is smaller than a first threshold, The second condition is that a difference between a luminance value of the encoding target block and a luminance value of a block at the same position as the encoding target block in a picture at least one before the encoding target picture including the encoding target block Is smaller than the second threshold, and when the determination unit determines that the encoding target block satisfies the first condition and the second condition, the setting unit determines that the encoding target block is And it outputs the quantization parameter of a predetermined fixed value in response to the picture type of the encoding target picture including a click.

In addition, the moving image encoding method of the present invention includes a determination step for determining whether or not an encoding target block satisfies the first condition and the second condition, and the encoding according to the determination result in the determination step. And a setting step for outputting a quantization parameter for quantizing the target block, wherein the first condition is that an index indicating a variation in luminance value of the encoding target block is smaller than a first threshold value. The second condition is that the luminance value of the encoding target block and the luminance value of the block at the same position as the encoding target block in the picture at least one prior to the encoding target picture including the encoding target block, And the determination step determines that the encoding target block satisfies the first condition and the second condition, and When the encoding target picture including the current block is determined to be an I-picture, the setting step for outputting a quantization parameter of a predetermined fixed value.
The moving image encoding method of the present invention includes a determination step for determining whether or not an encoding target block satisfies the first condition and the second condition, and the encoding target block according to a result determined by the determination step. And a setting step for outputting a quantization parameter for quantizing the image, wherein the first condition is that an index indicating a variation in luminance value of the encoding target block is smaller than a first threshold, The second condition is that a difference between a luminance value of the encoding target block and a luminance value of a block at the same position as the encoding target block in a picture at least one before the encoding target picture including the encoding target block Is smaller than the second threshold, and when the determination step determines that the encoding target block satisfies the first condition and the second condition, Constant step outputs the quantization parameter of a predetermined fixed value in response to the picture type of the encoding target picture including the current block.

  According to the present invention, it is possible to provide a moving image encoding apparatus and a moving image encoding method capable of suppressing flicker generated in an encoded moving image.

It is a figure which shows an example of the image | video which image | photographed the landscape subject with a motion. It is a figure which shows the structural example of the moving image encoder which concerns on embodiment of this invention. It is a flowchart which shows the determination process of the quantization parameter of 1st Embodiment. It is a flowchart explaining step S3 of FIG. It is another flowchart explaining step S3 of FIG. It is a figure which shows an example of the setting table of a quantization parameter. It is a flowchart which shows the determination process of the quantization parameter of 3rd Embodiment. It is a flowchart explaining step S5 of FIG. It is a flowchart which shows the determination process of the quantization parameter of 4th Embodiment. It is a figure which shows an example of an isolated point removal process. It is a figure which shows an example of a hole-filling process.

[TM5 code amount control algorithm]
First, the TM5 code amount control algorithm will be described. The TM5 code amount control algorithm consists of the following three steps.

  (Step 1) First, an initial quantization step Qint is set using a complexity index, which is an index indicating the difficulty of encoding the encoding target picture, and a target code amount. Then, the first encoding target block is encoded using Qint.

  (Step 2) Next, the quantization used for the next coding target block so that the generated code amount approaches the target code amount based on the difference between the generated code amount and the target code amount when encoding was performed in the past. Reset step Qstd.

  (Step 3) Weighting is applied to the quantization step Qstd set in Step 2 based on human visual characteristics to determine the final quantization step Qact. Such a weighting process is called “applied quantization”. Humans have a visual characteristic that, for example, an area with a flat luminance value in a picture is sensitive to even a slight deterioration in image quality. Therefore, TM5 applied quantization improves image quality by giving a small quantization step to flat areas where deterioration is noticeable visually and giving a large quantization step to other areas. I am trying.

Next, a method for deriving the quantization step Qact derived in step 3 will be described. A macroblock composed of 16 × 16 pixels is divided into four subblocks composed of 8 × 8 pixels, and the luminance value variances vblk ₁ , vblk ₂ , vblk ₃ , and vblk ₄ of each subblock are obtained by the following equations.

Here, P _i (i = 1 to 64) represents the luminance value of each pixel constituting the sub-block, and _Pavg represents the average value of the luminance values of each pixel constituting the sub-block.

The activity ACT _j of the macro block j is obtained from the variance of the luminance value of each sub block by the following equation.

Then, using the activity ACT _j derived from Equation 2, the normalized activity Nact _j of the macroblock _j is obtained by the following equation.

Here, AVG_ACT is an average value of activities ACT _j in all macroblocks in the encoding target picture. Finally, the final quantization step Qact _j is obtained by the following equation by accumulating the normalized activity Nact _j derived in Equation 3 to the quantization step Qstd _j derived in Step 2 above.

As described above, in the applied quantization process in TM5, the quantization step applied to each macroblock is corrected and derived in accordance with the activity ACT _j representing the variation in luminance value. For this reason, a smaller quantization step is given to a flat region in which deterioration is easily noticeable in a picture. That is, it is possible to suppress image quality deterioration in a flat region.

  MPEG-2 uses a quantization step, but H.264 / AVC and H.265 / HEVC use a quantization parameter proportional to the logarithm of the quantization step. The quantization parameter and the quantization step are related to each other when the quantization parameter is increased by 6 and the quantization step is doubled.

[Cause of flicker]
Next, the cause of occurrence of flicker will be described.

  Flicker is considered to occur because the degree of image quality deterioration of I pictures that arrive periodically varies from one I picture to another. When a certain I picture has deteriorated, the deterioration also affects the subsequent P picture and B picture, so the degree of deterioration continues over 1 GOP (Group of picture). When the degree of deterioration changes in the next I picture, the changed degree of deterioration continues over the next 1 GOP. Thus, it is considered that flicker occurs when the degree of image quality deterioration changes in one GOP cycle.

  In particular, when the quantization value (quantization step or quantization parameter) of the I picture is relatively large, flicker in a flat region becomes conspicuous. In TM5 that implements the code amount control algorithm, when obtaining the quantization step Qstd, the difference from the predicted image becomes large in the case of a video (moving image) whose luminance value varies greatly between consecutive pictures. The amount of generated code is suppressed by increasing the quantization step Qstd.

  Furthermore, TM5 determines the quantization step Qstd for each macroblock according to its complexity index, but if the region with large fluctuations in luminance value is dominant in the picture, the large Qstd of the entire picture As a result, Qstd of each macroblock also increases. That is, even if there is a macroblock with a small change in luminance value in the picture, it is dragged to a large quantization step Qstd of the entire picture, and a relatively large quantization step Qstd is assigned to the macroblock. . Then, the picture quality of the picture deteriorates, and flicker occurs if the degree of deterioration changes periodically. In particular, in a region where the luminance value changes little in the video, that is, a so-called flat region, visual image quality deterioration is conspicuous, and flicker is also noticeable.

  The present invention suppresses the flicker generated in the encoded moving image in consideration of the cause of the occurrence of such flicker. Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings.

[First Embodiment]
(Configuration of video encoding device)
FIG. 2 is a diagram illustrating a configuration example of the video encoding device 1 according to the present embodiment.

  The moving image encoding apparatus 1 according to the present embodiment generates a prediction image based on a local decoded image obtained by encoding and decoding an original image (input image) input from the input terminal 11, and the prediction image The prediction residual obtained by subtracting the image from the original image is encoded and output from the output terminal 17.

  In the present embodiment and the following embodiments, the H.264 / AVC and H.265 / HEVC quantization parameters will be described as examples, but the present invention is not limited to this. That is, the present invention can be applied to, for example, an MPEG-2 quantization step.

  An original image is input to the input terminal 11 of the moving image encoding device 1. The adder A12 subtracts a prediction image generated by the intra prediction unit 21 or the inter prediction unit 22 described later from the original image input from the input terminal 11 to generate a prediction residual.

  The orthogonal transform unit 13 performs orthogonal transform such as discrete cosine transform on the prediction residual that is the output of the adder A12. The quantization unit 14 quantizes the transform coefficient of the prediction residual that has been orthogonally transformed using the quantization parameter output from the control unit 24.

  The variable-length encoding unit 15 encodes the quantized prediction residual transform coefficient by a predetermined encoding method capable of reversible processing. The buffer 16 temporarily stores the encoded data output from the variable length encoding unit 15 and outputs the data to the output terminal 17 as a bit stream.

  The output of the quantization unit 14 is input to the variable length encoding unit 15 and also input to the inverse quantization unit 18. The inverse quantization unit 18 inversely quantizes the output of the quantization unit 14 and decodes the transform coefficient of the prediction residual. The inverse orthogonal transform unit 19 performs inverse orthogonal transform on the output of the inverse quantization unit 18 to decode the prediction residual. The adder B20 adds the prediction image generated by the intra prediction unit 21 or the inter prediction unit 22 and the prediction residual decoded by the inverse orthogonal transform unit 19 to generate a local decoded image.

  The intra prediction unit 21 uses the correlation between adjacent pixels, performs intra-screen prediction (intra prediction) using the pixel value of the locally decoded image, and generates a predicted image. The inter prediction unit 22 performs inter-screen prediction (inter prediction) using the motion information to be encoded, which is detected from the front or rear picture (frame) of the locally decoded image, and generates a predicted image. The switch 23 switches the output of the intra prediction unit 21 or the inter prediction unit 22 and selectively outputs it. The prediction image generated by the intra prediction unit 21 or the inter prediction unit 22 is output to the adder A12 and the adder B20.

  The control unit 24 acquires the generated code amount obtained from the data amount of the data stored in the buffer 16 and the original image input from the input terminal 11. Then, the control unit 24 determines a quantization parameter using the original image and the generated code amount, and outputs the quantization parameter to the quantization unit 14.

  The control unit 24 according to the present embodiment includes a determination unit 241, a setting unit 242, and a storage unit 243. The determination unit 241 determines whether or not the encoding target block satisfies a first condition and a second condition described later. The setting unit 242 outputs a quantization parameter for quantizing the encoding target block according to the determination result of the determination unit 241. Here, when the encoding target block satisfies the first condition and the second condition, the setting unit 242 has a predetermined value determined in advance according to the type (kind) of the encoding target picture including the encoding target block. Are set to the current block. That is, the setting unit 242 outputs the quantization parameter having the predetermined value to the quantization unit 14. The storage unit 243 stores a quantization parameter having a predetermined value corresponding to the type of the picture to be encoded.

  The moving image encoding apparatus 1 according to the present embodiment described above uses, for example, a general-purpose computer system including a CPU, a memory, an external storage device such as a hard disk, an input device, and an output device. it can. In this computer system, each function of the moving image encoding device 1 is realized by the CPU executing a program for the moving image encoding device 1 loaded on the memory. Further, the program for the moving image encoding apparatus 1 can be stored in a computer-readable recording medium such as a hard disk, a flexible disk, a CD-ROM, an MO, a DVD-ROM, or can be distributed via a network. .

(Quantization parameter determination process)
The quantization parameter determination process performed by the control unit 24 will be described below.

  The quantization parameter is determined for each encoding target block. The encoding target block used in the following description is equivalent to, for example, a “macroblock” in H.264 / AVC and a “coding unit (CU)” in H.265 / HEVC. It is.

  FIG. 3 is a flowchart illustrating quantization parameter determination processing according to the first embodiment. First, the setting unit 242 of the control unit 24 uses the complexity index, which is an index indicating the difficulty of encoding the encoding target picture (original image input from the input terminal 11), and the target code amount, to The quantization parameter BaseQP is calculated (step S1).

  Next, the setting unit 242 obtains a generated code amount obtained by encoding in the past, which is obtained from the data amount of the data accumulated in the buffer 16 or the like. Then, the setting unit 242 uses the next encoding target block so that the generated code amount approaches the target code amount based on the difference between the generated code amount when encoded in the past and the target code amount. The quantization parameter BaseQP ′ to be calculated is calculated (step S2).

  The processes in steps S1 and S2 may be the same as those in steps 1 and 2 of the TM5 code amount control algorithm described above.

  Finally, the control unit 24 determines a final quantization parameter QP in consideration of human visual characteristics (step S3).

  FIG. 4 is a flowchart for explaining the process of step S3 of FIG. 3 in more detail.

  First, the determination unit 241 of the control unit 24 divides the encoding target block into sub-blocks having a predetermined fixed size (step S3-1). In H.265 / HEVC, there are four types of 8 × 8 pixels, 16 × 16 pixels, 32 × 32 pixels, and 64 × 64 pixels as the size of the encoding target block (CU), and any one of the sizes is appropriately selected.

  For example, when 16 × 16 pixels are adopted as the fixed size of the sub-block, the encoding target block of 64 × 64 pixels is divided into 16 sub-blocks, and the encoding target block of 32 × 32 pixels is divided into four sub-blocks. In this case, the 16 × 16 pixel encoding target block and the 8 × 8 pixel encoding target block are not divided into subblocks, and the encoding target block and the subblock are the same. Of course, a sub-block having a size other than 16 × 16 pixels may be employed.

  Next, the determination unit 241 determines, for each subblock obtained by dividing the encoding target block, whether or not the subblock satisfies both [Condition 1] and [Condition 2] (Step S3-2). .

  [Condition 1] is a condition that an index (parameter) indicating a variation in luminance values of pixels constituting the target sub-block is smaller than a predetermined threshold (first threshold). In other words, [Condition 1] is a condition for determining whether or not the target sub-block is a flat region. As an index indicating the variation in the luminance value, for example, activity, dispersion, or the like can be used. When an activity is used, [Condition 1] can be described by the following equation.

Where x _i is the luminance value of each pixel constituting the sub-block, x _avg is the average value of the luminance values of all the pixels constituting the sub-block, n is the number of pixels in the sub-block, and Th1 is the flat area of the sub-block It is a threshold value (first threshold value) for determining whether or not.

  In [Condition 2], the difference between the luminance value of the target sub-block and the luminance value of the sub-block at the same position as the target sub-block in at least one picture before the encoding target picture including the target sub-block is The condition is smaller than a predetermined threshold (second threshold). In other words, [Condition 2] is a condition for determining whether or not the target sub-block is in a stationary state. [Condition 2] can be described by the following equation, for example.

Here, x _i represents the luminance value of each pixel constituting the sub-block, and y _i represents the luminance value of the pixel at the same position as the pixel x _i on the picture immediately preceding the encoding target picture. Th2 represents a threshold value (second threshold value) for determining whether or not the sub-block is in a stationary state.

  The at least one picture before the encoding target picture may be at least one previous picture regardless of the picture type (I, P, B), or at least one of the same picture type. It may be the previous picture.

  If the target sub-block does not satisfy both [Condition 1] and [Condition 2], that is, if at least one condition is not satisfied (step S3-2: NO), the process proceeds to step S3-3.

  In step S3-3, the determination unit 241 determines whether or not the process of step S3-2 has been performed on all subblocks of the encoding target block. When the process of step S3-2 is not performed for all the sub blocks (step S3-3: NO), the process returns to step S3-2, and the determination unit 241 performs the process of step S3-2 on the unprocessed subblock. Process.

  When the process of step S3-2 is performed for all the sub blocks (step S3-3: YES), the process proceeds to step S3-5. That is, if there is no sub-block satisfying both [Condition 1] and [Condition 2] in the encoding target block, the process proceeds to Step S3-5.

  In step S3-5, the setting unit 242 sets a quantization parameter (QP = BaseQP ′ + ΔQP) obtained by adding the weighting coefficient ΔQP to the quantization parameter BaseQP ′ set in step S2 in the encoding target block. The processing in step S3-5 may be performed in the same manner as in step 3 of the TM5 code amount control algorithm described above.

  On the other hand, if the target sub-block satisfies both [Condition 1] and [Condition 2] (step S3-2: YES), the process proceeds to step S3-4. That is, if even one sub-block satisfying both [Condition 1] and [Condition 2] exists in the encoding target block, the process proceeds to step S3-4.

  In step S3-4, the setting unit 242 sets a quantization parameter (QP = ForceQP) having a predetermined value that is predetermined according to the type of the picture, for the current block. The quantization parameter ForceQP having a predetermined value will be described later.

  In the process shown in FIG. 4, the control unit 24 divides the encoding target block into fixed-size sub-blocks (step S3-1), and at least one of the divided sub-blocks is [condition 1]. And [Condition 2] are satisfied (step S3-2: YES), the quantization parameter ForceQP having a predetermined value is set to the current block to be encoded (step S3-4).

  However, as illustrated in FIG. 5, the determination unit 241 does not divide the encoding target block in Step S3-1 of FIG. 4 into sub-blocks, and the encoding target block itself is [condition 1] and [condition 2]. It may be determined whether or not both are satisfied (step S3-2).

  In this case, [Condition 1] is a condition that the index indicating the variation in the luminance value of the pixels constituting the encoding target block is smaller than a predetermined threshold (first threshold). [Condition 2] is that the difference between the luminance value of the encoding target block and the luminance value of the block at the same position as the encoding target block in the picture at least one prior to the encoding target picture including the encoding target block is: The condition is that it is smaller than a predetermined threshold (second threshold).

  Step S3-2: In the case of NO, the process directly proceeds to Step S3-5 without performing the process of Step S3-3 in FIG. Step S3-5 is the same processing as step S3-5 in FIG. Step S3-2: If YES, proceed to step S3-4. Step S3-4 is the same processing as step S3-4 in FIG.

(Setting example of quantization parameter ForceQP)
Next, the quantization parameter ForceQP having a predetermined value set for the encoding target block in step S3-4 of FIG. 4 or FIG. 5 will be described. In step S3-4, the setting unit 242 sets a quantization parameter ForceQP having a predetermined value determined in advance according to the picture type, for the current block.

  FIG. 6 shows an example of a setting table for the quantization parameter ForceQP. In the illustrated example, “26” is set as the quantization parameter ForceQP in the case of an I picture, and “40” is set as the quantization parameter ForceQP in the case of a P picture or a B picture. The setting table is stored in, for example, the storage unit 243 provided in the control unit 24.

  When the picture type of the picture (or slice) including the encoding target block is an I picture, the setting unit 242 refers to the setting table and sets a predetermined value (for example, “24”) that is a fixed value for the I picture. The quantization parameter ForceQP is set to the current encoding target block.

  This is because flicker is considered to occur when the degree of image quality deterioration of an I-picture that arrives periodically continues over 1 GOP and the degree of image quality deterioration in this 1 GOP period changes (fluctuates). is there. Therefore, in the present embodiment, the degree of image quality degradation in one GOP cycle is obtained by uniformly assigning a predetermined quantization parameter ForceQP, which is a fixed value, to the encoding target blocks satisfying the conditions 1 and 2. Fluctuations, thereby suppressing the occurrence of flicker.

  Here, as the “predetermined value” for the I picture, for example, from the quantization parameter value BaseQP ′ + ΔQP (FIG. 4, FIG. 5: Step 3-5) set by a general applied quantization algorithm It is desirable to make the value small.

  On the other hand, when the picture type of the picture (or slice) including the encoding target block is a P picture or a B picture, the setting unit 242 refers to the setting table and sets a predetermined value that is a fixed value for the P picture or the B picture. A value quantization parameter ForceQP is set in the current block to be encoded. As the “predetermined value” for the P picture or the B picture, a value such that the transform coefficient of the prediction residual after quantization by the quantization unit 14 (hereinafter referred to as “quantization coefficient”) becomes zero ( In the example shown, it is “40”). The quantization unit 14 quantizes the transform coefficient of the prediction residual that has been orthogonally transformed using the quantization parameter output from the control unit 24.

  When the quantization coefficient of the encoding target block becomes 0, the quantization unit 14 does not output the quantization coefficient of the encoding target block, and thus the variable length encoding unit 15 converts the quantization coefficient to the variable length code. Do not output converted codes. Such an encoding target block is called a “skip block”. When a moving image decoding apparatus (not shown) that decodes an encoded moving image recognizes that the block to be decoded is a skip block, the skip block refers to the skip block. Decoding is performed without adding a prediction residual to a reference block in a picture.

  In FIG. 6, when the encoding target block is included in the P picture or B picture, the quantization parameter is set such that the quantization coefficient is 0 in the encoding target block. However, since the encoding target block included in the P picture or B picture only needs to be a “skip block”, the setting unit 242 sets a predetermined value of the quantization parameter ForceQP, instead of skipping with a syntax. The flag may be set to “1”. In this case, the setting unit 242 outputs “skip flag = 1” indicating that the block is a skip block to the quantization unit 14 without outputting the quantization parameter to the quantization unit 14. That is, the setting unit 242 forcibly sets the encoding target block as a skip block. When “skip flag = 1” is input, the quantization unit 14 does not perform the quantization process, and therefore does not output the quantization coefficient to the variable length encoding unit 15.

  As described above, in this embodiment, the quantization parameter for quantizing the encoding target block is output according to the determination results of [Condition 1] and [Condition 2]. Specifically, it is an encoding target block that satisfies both [Condition 1] and [Condition 2] (step S3-2 in FIG. 4 and FIG. 5: YES encoding target block), and the encoding target block If the encoding target picture including the I picture is an I picture, a quantization parameter having a predetermined value is set in the encoding target block and output.

  It should be noted that both [Condition 1] (the index indicating the variation in luminance value is smaller than the first threshold value) and [Condition 2] (the difference between the luminance value of at least the previous picture is smaller than the second threshold value) The encoding target block to be satisfied is a flat region with little variation in luminance value and a still state region with little motion, and flicker is likely to occur.

  In the present embodiment, since the quantization parameter having the same value is always set for the encoding target block in which flicker is likely to occur in the I picture, the degree of image quality deterioration does not vary in one GOP cycle. It is possible to suppress the occurrence of flicker occurring in the encoded moving image. In addition, since the degradation of the I picture also affects the subsequent P picture and B picture, the quantization parameter of a predetermined value is uniformly applied to the encoding target block in which I picture flicker is likely to occur. The occurrence of flicker can be more effectively suppressed.

  Further, in the present embodiment, when a coding target block that satisfies both [Condition 1] and [Condition 2] is included in a P picture or B picture, it is determined as a skip block. As a result, the same image quality as that of the I picture continues during 1 GOP, and the degree of image quality deterioration does not change within 1 GOP. Therefore, the occurrence of flicker occurring in the encoded moving image can be more effectively suppressed.

[Second Embodiment]
In the first embodiment, whether or not the current block satisfies both [Condition 1] and [Condition 2] is used as a criterion for determining whether or not to apply the quantization parameter ForceQP having a predetermined value. (Step S3-2 in FIGS. 4 and 5).

  The present embodiment is different from the first embodiment in that [Condition 1A] is used instead of [Condition 1] as a determination criterion, and the rest is the same as the first embodiment. Hereinafter, [Condition 1A], which is a difference from the first embodiment, will be described in detail.

  [Condition 1] described in the first embodiment is a condition that the index indicating the variation in the luminance value of the encoding target block is smaller than the predetermined first threshold. On the other hand, in [Condition 1A] used in the present embodiment, the index indicating the variation in the luminance value of the block to be encoded is smaller than a predetermined first threshold value, and is smaller than a predetermined third threshold value (third threshold value <first threshold value). It is also a condition that is large.

  Here, the condition that the index indicating the variation in the luminance value is larger than the third threshold is, in other words, a condition for eliminating a sub-block that has no change in luminance. In an artificial image that does not include noise generated when captured by a camera, there are not a few regions in the picture where the luminance and color tone do not change at all. Artificial images include, for example, images from signal generators such as animation images, CG images, and color bars.

  When the first embodiment is applied to an artificial image, the quantization parameter of an I picture may be too larger than the quantization parameter value BaseQP ′ + ΔQP determined by normal code amount control. In this case, a minute change of a signal that is not noise cannot be reproduced (decoded), and there is a disadvantage that the image is deteriorated as compared with normal code amount control.

  Therefore, in the present embodiment, [Condition 1A] is used in which the condition that the index indicating the variation in luminance value is larger than the third threshold is further added to [Condition 1]. As a result, the encoding target block of the artificial image is excluded from the application target of the quantization parameter ForceQP having a predetermined value, and only the encoding target block of the natural image obtained by capturing a person or a landscape with the camera is applied. be able to. That is, it is possible to avoid an adverse effect on image quality caused by applying a predetermined quantization parameter ForceQP to an artificial image. Natural images always have analog fluctuations in luminance values. Therefore, in [Condition 1A], a third threshold value that is smaller than the first threshold value is provided for variations in luminance values, and an artificial image is excluded.

  For example, an activity may be used as an index indicating the variation in luminance value in [Condition 1A]. In this case, [Condition 1A] can be described by the following equation.

  Here, Th3 represents a predetermined threshold (third threshold) for determining whether or not the encoding target block is a natural image.

  In the case of FIG. 4, the determination unit 241 of the video encoding device 1 according to the present embodiment, in the case of FIG. 4, determines whether there is a sub-block that satisfies both [Condition 1A] and [Condition 2] in the target block. It is determined whether or not (step S3-2). When there is a sub-block that satisfies both [Condition 1A] and [Condition 2], the setting unit 242 sets a quantization parameter ForceQP having a predetermined value according to the picture type for the current block. (Step S3-4).

  In the case of FIG. 5, the determination unit 241 determines whether or not the encoding target block satisfies both [Condition 1A] and [Condition 2] (Step S3-2). When the encoding target block satisfies both [Condition 1A] and [Condition 2], the setting unit 242 sets a quantization parameter ForceQP having a predetermined value according to the picture type for the encoding target block. (Step S3-4).

  Note that the value of the quantization parameter ForceQP may be set to the same value as in the first embodiment.

  As described above, in this embodiment, instead of [Condition 1], [Condition 1A] is adopted in which the index indicating the variation in the luminance value of the encoding target block is smaller than the first threshold and larger than the third threshold. did. Thereby, in this embodiment, setting the quantization parameter ForceQP of a predetermined value for an artificial image that may induce image quality degradation is avoided, and only the natural image is set to the quantization parameter ForceQP. Can be set.

[Third Embodiment]
As described above, in the case of a picture in which an area with a large luminance value fluctuation is dominant, a relatively large quantization parameter is given even to a block with a small luminance value change, dragged by a large quantization parameter of the entire picture. It will be. On the contrary, when the region where the fluctuation of the luminance value is large is not so dominant, it is not necessary to apply the quantization parameter ForceQP having the predetermined value described in the first and second embodiments.

  Therefore, in the present embodiment, the encoding target for which it has been determined that the quantization parameter ForceQP having a predetermined value should be set in the encoding target picture while adopting the processing based on the first or second embodiment. Only when the ratio of blocks (hereinafter referred to as “forced QP target block”) is less than a predetermined ratio, a quantization parameter ForceQP having a predetermined value is set in the forced QP target block.

  Hereinafter, the method for determining the quantization parameter of the present embodiment will be described with reference to FIGS.

  FIG. 7 is a flowchart illustrating a quantization parameter determination method according to this embodiment. The encoding target block used in the following description is called “macroblock” in H.264 / AVC, for example, as in the first and second embodiments. 265 / HEVC is equivalent to what is called an “encoding unit”.

  Steps S1 and S2 in FIG. 7 are the same as steps S1 and S2 described in FIG. That is, the setting unit 242 of the video encoding device 1 calculates the initial quantization parameter BaseQP from the complexity index that is an index indicating the difficulty of encoding the encoding target picture and the target code amount (step S1). Then, the setting unit 242 is used for each encoding target block so that the generated code amount approaches the target code amount based on the difference between the generated code amount and the target code amount when encoded in the past. A quantization parameter BaseQP ′ is calculated (step S2).

  Then, the control unit 24 determines a final quantization parameter QP in consideration of human visual characteristics (step S3A). Here, step 3A performs the same processing as step 3 described in the first or second embodiment. However, in S3-4 and S3-5 (FIGS. 4 and 5) in step S3 of the first or second embodiment, the setting unit 242 actually sets the quantization parameter for each encoding target block. In contrast, in step S3A of the present embodiment, the setting unit 242 sets each encoding target block as “a flag indicating whether or not it is a forced QP target block” (hereinafter referred to as “ForceQP flag”). Set.

  That is, in the case of FIG. 4, when at least one sub-block satisfies both [Condition 1] (or [Condition 1A]) and [Condition 2] (step S3-2: YES), the setting unit 242 “ForceQP flag = 1” is set in the target block (step S3-4). On the other hand, when all the sub-blocks do not satisfy the condition of step S3-2 (step S3-2: NO), the setting unit 242 sets “ForceQP flag = 0” to the encoding target block (step S3-5). ).

  In the case of FIG. 5, when the encoding target block satisfies both [Condition 1] (or [Condition 1A]) and [Condition 2] (step S3-2: YES), the setting unit 242 sets the encoding target block as the encoding target block. “ForceQP flag = 1” is set (step S3-4). On the other hand, when the encoding target block does not satisfy the condition of step S3-2 (step S3-2: NO), the setting unit 242 sets “ForceQP flag = 0” to the encoding target block (step S3-5). ).

  The encoding target block set with “ForceQP flag = 1” is a forced QP target block. The encoding target block set with “ForceQP flag = 0” is a non-forced QP target block other than the forced QP target block.

  Next, the setting unit 242 determines whether or not the ForceQP flag is set for all the encoding target blocks included in the encoding target picture (step S4). When the ForceQP flag is not set for all the encoding target blocks (step S4: NO), the control unit 24 performs the processes of step S2 and step S3A on the next encoding target block. On the other hand, when the ForceQP flag is set for all the encoding target blocks (step S4: YES), the process proceeds to step S5.

  In step S5, the setting unit 242 determines the quantization parameter of each encoding target block based on the ratio of the number of compulsory QP target blocks for which ForceQP flag = 1 among the total number of encoding target blocks of the encoding target picture. Determine and set QP.

  FIG. 8 is a flowchart for explaining the process of step S5 in FIG. 7 in more detail.

  First, the setting unit 242 calculates the ratio of the number of forced QP target blocks to the total number of blocks to be encoded of the encoding target picture (step S5-1). When this ratio is less than the predetermined ratio (step S5-1: YES), the process proceeds to step S5-2, and when it is equal to or greater than the predetermined ratio (step S5-1: NO), the process proceeds to step S5-3.

  The “predetermined ratio” varies depending on the nature of the image to be encoded, the mounting form of the moving image encoding apparatus 1, etc., but a value of about 20% may be used, for example.

  In step S5-2, the setting unit 242 sets a quantization parameter in each encoding target block according to the value of the ForceQP flag. Specifically, the setting unit 242 sets a predetermined quantization parameter ForceQP for the encoding target block with the ForceQP flag = 1, and applies the applied quantum to the encoding target block with the ForceQP flag = 0. The quantization parameter BaseQP ′ + ΔQP calculated based on the quantization algorithm is set.

  On the other hand, in step S5-3, the setting unit 242 sets the quantization parameter BaseQP ′ + ΔQP calculated based on the applied quantization algorithm for all the encoding target blocks regardless of the value of the ForceQP flag. .

  7 and 8, the setting unit 242 counts the number of blocks to be coded with ForceQP flag = 1 for each coding target picture, thereby compulsory QP targets in the coding target picture. Although the ratio of the number of blocks is obtained, the present invention is not limited to this.

  In general, moving pictures are very similar to each other and have a strong temporal correlation. Therefore, for example, the ratio of the number of forced QP target blocks in the picture immediately before the encoding target picture may be used for the encoding target picture. In this case, the setting unit 242 performs the process of step S5-1 in FIG. 8 on the first picture for the consecutive n pictures, and the process of step S5-1 for the subsequent n−1 pictures. Without performing step S5-2, the process of step S5-2 or step S5-3 is performed using the ratio calculated for the first picture.

  As described above, in the present embodiment, the compulsory QP target block is used only when the ratio of the number of compulsory QP target blocks (specific blocks) is less than a predetermined ratio among the total number of blocks to be encoded of the encoding target picture. A predetermined quantization parameter ForceQP is set for and output.

  As a result, in the present embodiment, since the quantization parameter ForceQP having a predetermined value is set only for a picture in which flicker is expected to occur, both improvement in image quality and high-accuracy code amount control are achieved. However, the occurrence of flicker can be suppressed.

[Fourth Embodiment]
In this embodiment, at least one process of “isolated point removal” and “hole filling” is added to the third embodiment.

  “Isolated point removal” is a case where the encoding target block is determined to be a forced QP target block, but if many of the surrounding blocks are determined not to be a forced QP target block, the encoding target block is not forced QP. This is processing to change to the target block.

  If “Block filling” is determined that the encoding target block is a non-mandatory QP target block, but many of the surrounding blocks are determined to be compulsory QP target blocks, the encoding target block is determined to be a compulsory QP target. It is a process to change to a block.

  Hereinafter, a method for determining a quantization parameter according to the present embodiment will be described with reference to FIGS. 9 to 11.

  FIG. 9 is a flowchart illustrating a quantization parameter determination method according to this embodiment. The encoding target block used in the following description is called “macroblock” in H.264 / AVC, for example, as in the first and second embodiments. 265 / HEVC is equivalent to what is called an “encoding unit”.

  Steps S1, S2, S3A, and S4 in FIG. 9 are the same processes as steps S1, S2, S3A, and S4 in FIG. 7 of the third embodiment, and thus description thereof is omitted here. In step S5, the setting unit 242 refers to the ForceQP flag of each block constituting the encoding target picture, and performs “isolated point removal” and / or “hole filling” processing. Note that these processes may be performed according to the characteristics of the picture (moving picture) or from the viewpoint of mounting the moving picture coding apparatus 1, both “isolated point removal” and “hole filling”. Alternatively, only one of the processes may be performed.

  FIG. 10 is a diagram illustrating an example of the isolated point removal process in step S5. In the example illustrated in FIG. 10, the encoding target block is a block described as “Cur” and is arranged at the center. In addition, the blocks determined as the compulsory QP target blocks are hatched. Further, the eight adjacent blocks present around the encoding target block are hatched and described as “force” when determined to be a forced QP target block.

  In the example shown in FIG. 10A, the encoding target block itself is a “forced QP target block”, and the number of “forced QP target blocks” is one among eight adjacent blocks.

  When the encoding target block is a “forced QP target block” and the number of “forced QP target blocks” among the adjacent blocks is less than N, the setting unit 242 determines that the encoding target block is “non-forced QP It is regarded as “target block”, and ForceQP flag = 0 is set to the encoding target block. N (first predetermined number) may be N = 2, for example.

  In the example of FIG. 10A, since the number of “forced QP target blocks” in the adjacent blocks is one, the setting unit 242 changes the ForceQP flag to change the encoding target block to “forced QP target block”. Change from "Target block" to "Non-mandatory QP target block". FIG. 10B is a diagram illustrating the encoding target block after the change.

  FIG. 11 is a diagram illustrating an example of the hole filling process in step S5. In the example shown in FIG. 11A, the encoding target block itself is a “non-compulsory QP target block”, and the number of “forced QP target blocks” is six among eight adjacent blocks.

  When the encoding target block is a “non-mandatory QP target block” and the number of “forced QP target blocks” among the adjacent blocks is M or more, the setting unit 242 It is regarded as “target block”, and ForceQP flag = 1 is set to the encoding target block. M (second predetermined number) may be M = 6, for example.

  In the example shown in FIG. 11A, since the number of “forced QP target blocks” in the adjacent blocks is six, the setting unit 242 changes the ForceQP flag to change the encoding target block to “non- Change from "Forced QP block" to "Forced QP block". FIG. 11B is a diagram showing the encoding target block after the change.

  Finally, the setting unit 242 performs the process of step S6. Step S6 is the same as the process of step S5 (FIGS. 7 and 8) in the third embodiment. In step S <b> 6, the setting unit 242 assigns each encoding target block based on the ratio of the number of “forced QP target blocks” in which the ForceQP flag = 1 among the total number of encoding target blocks in the encoding target picture. Sets the value of the quantization parameter QP to be set.

  As described above, in this embodiment, in addition to the processing of the third embodiment, at least one of “isolation point removal” or “hole filling” is performed. Thereby, in this embodiment, it can be judged appropriately according to the condition of the surrounding adjacent block whether an encoding object block is made into a forced QP object block.

  In the present embodiment, the quantization parameter ForceQP having a predetermined value is set only for a picture in which flicker is expected to occur, so that both improvement in image quality and high-accuracy code amount control are achieved. The occurrence of flicker can be suppressed.

  In addition, this invention is not limited to the said embodiment, Many deformation | transformation are possible within the range of the summary. For example, in the above embodiment, a quantization parameter having a predetermined value for an I picture is assigned to an encoding target block that satisfies [Condition 1] (or [Condition 1A]) and [Condition 2] in the I picture. However, a quantization parameter having a predetermined value for the base layer may be set for a base layer picture in temporal layer coding as in the case of the I picture of the above embodiment. By uniformly setting a predetermined value of the quantization parameter to the picture in the base layer, it is possible to more effectively suppress the occurrence of flicker occurring in the encoded moving image.

  In addition, at least two arbitrary embodiments of the first to fourth embodiments described above may be combined.

1: Moving picture encoding device 11: Input terminal 12: Adder A
13: Orthogonal transformation unit 14: Quantization unit 15: Variable length coding unit 16: Buffer 17: Output terminal 18: Inverse quantization unit 19: Inverse orthogonal transformation unit 20: Adder B
21: Intra prediction unit 22: Inter prediction unit 23: Switch 24: Control unit 241: Determination unit 242: Setting unit 243: Storage unit

Claims (12)

A determination unit that determines whether the encoding target block satisfies the first condition and the second condition;
A setting unit that outputs a quantization parameter for quantizing the encoding target block according to the determination result of the determination unit, and
The first condition is that an index indicating a variation in luminance value of the encoding target block is smaller than a first threshold,
The second condition is that a luminance value of the encoding target block and a luminance value of a block at the same position as the encoding target block in a picture at least one prior to the encoding target picture including the encoding target block The difference is less than the second threshold,
When the determination unit determines that the encoding target block satisfies the first condition and the second condition, and determines that the encoding target picture including the encoding target block is an I picture, The moving picture coding apparatus, wherein the setting unit outputs a quantization parameter having a predetermined fixed value. A determination unit that determines whether the encoding target block satisfies the first condition and the second condition;
A setting unit that outputs a quantization parameter for quantizing the encoding target block according to the determination result of the determination unit, and
The first condition is that an index indicating a variation in luminance value of the encoding target block is smaller than a first threshold,
The second condition is that a luminance value of the encoding target block and a luminance value of a block at the same position as the encoding target block in a picture at least one prior to the encoding target picture including the encoding target block The difference is less than the second threshold,
When the determination unit determines that the encoding target block satisfies the first condition and the second condition, the setting unit determines in advance according to a picture type of the encoding target picture including the encoding target block. Output a fixed parameter with a fixed value
A video encoding apparatus characterized by the above. The first condition, the index is less than the first threshold value, the moving picture encoding apparatus according to claim 1 or 2, wherein the greater than the third threshold value. The setting unit is configured to specify the specific block when a ratio of a specific block determined to satisfy the first condition and the second condition is less than a predetermined ratio among all the encoding target blocks of the encoding target picture. moving picture coding apparatus according to any one of claims 1 and outputs the quantization parameter of the fixed value for the encoding target block which is determined to block 3. The setting unit
The encoding target block is a specific block that is determined to satisfy the first condition and the second condition, and among the adjacent blocks of the encoding target block, less than a first predetermined number of adjacent blocks In the case of a specific block, the coding target block is changed to a non-specific block determined not to satisfy at least one of the first condition and the second condition, or
When the encoding target block is the non-specific block, and the second predetermined number of adjacent blocks among the adjacent blocks of the encoding target block are the specific block, the encoding target block is changed to the specific block. The moving picture coding apparatus according to any one of claims 1 to 4 , wherein at least one of the changing is performed. The setting unit skips the encoding target block when the encoding target block satisfies the first condition and the second condition, and the encoding target picture is a P picture or a B picture. moving picture coding apparatus according to any one of claims 1 to 5, characterized in that a. A determination step of determining whether the encoding target block satisfies the first condition and the second condition;
A setting step for outputting a quantization parameter for quantizing the encoding target block according to the result determined in the determination step,
The first condition is that an index indicating a variation in luminance value of the encoding target block is smaller than a first threshold,
The second condition is that a luminance value of the encoding target block and a luminance value of a block at the same position as the encoding target block in a picture at least one prior to the encoding target picture including the encoding target block The difference is less than the second threshold,
When the determination step determines that the encoding target block satisfies the first condition and the second condition, and determines that the encoding target picture including the encoding target block is an I picture, The moving picture coding method, wherein the setting step outputs a quantization parameter having a predetermined fixed value. A determination step of determining whether the encoding target block satisfies the first condition and the second condition;
A setting step for outputting a quantization parameter for quantizing the encoding target block according to the determination result of the determination step, and
The first condition is that an index indicating a variation in luminance value of the encoding target block is smaller than a first threshold,
The second condition is that a luminance value of the encoding target block and a luminance value of a block at the same position as the encoding target block in a picture at least one prior to the encoding target picture including the encoding target block The difference is less than the second threshold,
When the determination step determines that the encoding target block satisfies the first condition and the second condition, the setting step is performed in advance according to a picture type of the encoding target picture including the encoding target block. Output a fixed parameter with a fixed value
  A video encoding method characterized by the above. The moving image encoding method according to claim 7 or 8 , wherein the first condition is that the index is smaller than the first threshold and larger than a third threshold. In the setting step, when the ratio of specific blocks determined to satisfy the first condition and the second condition among all the encoding target blocks of the encoding target picture is less than a predetermined ratio, the specifying step 10. The moving picture encoding method according to claim 7 , wherein the quantization parameter having the fixed value is output for an encoding target block determined to be a block. 11. The encoding target block is a specific block that is determined to satisfy the first condition and the second condition, and among the adjacent blocks of the encoding target block, less than a first predetermined number of adjacent blocks In the case of a specific block, changing the encoding target block to a non-specific block determined not to satisfy at least one of the first condition and the second condition, or
When the encoding target block is the non-specific block, and the second predetermined number of adjacent blocks among the adjacent blocks of the encoding target block are the specific block, the encoding target block is changed to the specific block. The moving picture coding method according to claim 7 , further comprising at least one of changing steps. In the setting step, when the encoding target block satisfies the first condition and the second condition, and the encoding target picture is a P picture or a B picture, the encoding target block is skipped The moving picture coding method according to any one of claims 7 to 11 , wherein:

Images