JP4343667B2 - Image coding apparatus and image coding method - Google Patents

Description

  The present invention relates to a moving image encoding apparatus and a moving image encoding method for inputting moving images and outputting encoded data, and in particular, it is possible to obtain good image quality even at a low bit rate. The present invention relates to an image encoding device, an image encoding method, and the like.

  Due to dramatic progress in digital signal processing technology in recent years, recording of moving images on storage media and transmission of moving images via transmission paths, which has been difficult in the past, are performed. In this case, each frame constituting the moving image is subjected to compression processing, and the amount of data is greatly reduced. A typical technique for this compression processing is known as, for example, MPEG (Moving Picture Experts Group). When compressing and encoding an image in accordance with the MPEG system, the amount of code often varies greatly depending on the spatial frequency characteristics, the scene, and the quantization scale value, which are characteristics of the image itself. An important technique for realizing a decoded image with good image quality in realizing an encoding apparatus having such encoding characteristics is code amount control.

  TM5 (Test Model 5) (see Non-Patent Document 1) is known as one of the code amount control algorithms. The code amount control algorithm based on TM5 is composed of the three steps outlined below, and the code amount is controlled so that the bit rate is constant for each GOP (Group of Picture).

[STEP1: Target bit allocation]
In step 1, the target code amount of the next picture to be encoded is set. Here, in the processing of STEP1, the code amount Rgop allowed in the current GOP is calculated by the following equation (1) (in the following equations, “*” means multiplication). .

Rgop = (ni + np + nb) * (bits_rate / picture_rate) (1)
Here, ni, np, and nb are the number of remaining pictures in the current GOP of I, P, and B pictures, bits_rate represents the target bit rate, and picture_rate represents the picture rate. Further, the complexity of the picture is obtained from the encoding result for each of the I, P, and B pictures by the following equation (2).

Xi = Ri * Qi
Xp = Rp * Qp (2)
Xb = Rb * Qb
Here, Ri, Rp, and Rb are code amounts obtained as a result of encoding I, P, and B pictures, respectively, and Qi, Qp, and Qb are Q in all macroblocks in the I, P, and B pictures, respectively. Reference value of the scale. From the expressions (1) and (2), the target code amounts Ti, Tp, and Tb for the I, P, and B pictures can be obtained by the following expression (3).

Ti = max {(Rgop / (1 + ((Np * Xp) / (Xi * Kp)) + ((Nb * Xb) / (Xi * Kb)))), (bit_rate / (8 * picture_rate))}
Tp = max {(Rgop / (Np + (Nb * Kp * Xb) / (Kb * Xp))), (bit_rate / (8 * picture_rate))}
Tb = max {(Rgop / (Nb + (Np * Kb * Xp) / (Kp * Xb))), (bit_rate / (8 * picture_rate))}
... (3)
Np and Nb are the remaining number of P and B pictures in the current GOP, respectively, and constants Kp = 1.0 and Kb = 1.4.

[STEP2: Rate control]
In STEP2, three virtual buffers are used for each of I, P, and B pictures, and the difference between the target code amount obtained by Expression (3) and the generated code amount is managed. The data accumulation amount of the virtual buffer is fed back, and the reference value of the Q scale is set for the macro block to be encoded next so that the actual generated code amount approaches the target code amount based on the data accumulation amount. For example, when the current picture type is a P picture, the difference between the target code amount and the generated code amount can be obtained by arithmetic processing according to the following equation (4).

dp, j = dp, 0 + Bp, j−1 − ((Tp * (j−1)) / MB_cnt) (4)
Here, the subscript j is the number of the macroblock in the picture, dp, 0 indicates the initial fullness of the virtual buffer, Bp, j is the total code amount up to the jth macroblock, and MB_cnt is in the picture Indicates the number of macro blocks. The relationship of equation (4) is shown in a graph, for example, as shown in Fig. 2. In Fig. 2, the horizontal axis is the number of macro blocks (MB_cnt) in the picture, and the vertical axis is the p picture. Dp, j in the figure is a difference value obtained by equation (4).

  Next, when the reference value of the Q scale in the j-th macro block is obtained using dp, j (hereinafter referred to as “dj”), equation (5) is obtained.

Qj = (dj * 31) / r (5)
Where r = 2 * bits_rate / picture_rate (6)
It is.

[STEP3: Adaptive Quantization]
In the process of STEP 3, a process of finally determining the quantization scale is executed based on the spatial activity of the macro block to be encoded so that the visual characteristics, that is, the image quality of the decoded image becomes good.

ACTj = 1+ min (vblk1, vblk2, ..., vblk8) (7)
In equation (7), vblk1 to vblk4 indicate the spatial activity in the 8 × 8 sub-block in the frame-structured macro block, and vblk5 to vblk8 indicate the spatial activity in the 8 × 8 sub-block in the field-structured macro block. Show. Here, the calculation of the space activity can be obtained by the following equations (8) and (9).

vblk = Σ (Pi−Pbar) ² (8)
Pbar = (1/64) * ΣPi (9)
Here, Pi is a pixel value in the i-th macro block, and Σ in equations (8) and (9) is an operation of i = 1 to 64. Next, ACTj obtained by the equation (7) is normalized by the following equation (10).

N_ACTj = (2 * ACTj + AVG_ACT) / (ACTj + AVG_ACT) (10)
Here, AVG_ACT is a reference value of ACTj in a previously encoded picture, and finally a quantization scale (Q scale value) MQUANTj is obtained by the following equation (11).

MQUANTj = Qj * N_ACTj (11)
According to the above TM5 algorithm, a large amount of code is assigned to the I picture by the processing of STEP1, and the amount of code is in a flat portion (low spatial activity) that is visually noticeable in the picture. A lot will be allocated.

  In Patent Document 1, which is a technique for the problem of TM5, in order to obtain a balance point of a cutoff frequency of a low-pass filter (LPF) as shown in FIG. 3, a balance function is defined and quantized. Attempting to solve the problem in TM5 by matching the distortion and the degradation of image sharpness. This is realized by reducing the spatial frequency of each picture input to the encoding device with LPF in order to suppress quantization distortion. The two curves in FIG. 3 correspond to the following two functions F1 and F2.

F1 (motion amount, filter coefficient, quantization scale, code amount)
F2 (filter coefficient, quantization scale)
The intersection of the functions F1 and F2 is set as the balance point, and the quantization scale and the LPF filter coefficient with the best matching of the code amount and the image quality are set.

Patent Document 2 discloses a technique for avoiding a rapid change when changing a filter coefficient as a moving image encoding method.
Test Model 5 (Test Model Editing Commitee: "Test Model 5", ISO / IEC JTC / SC29 / WG11 / N0400 (Apr. 1993)) Japanese Patent No. 2894137, “Prefilter Control Method and Apparatus in Video Coding” JP 2002-247576 A

  However, the above TM5 algorithm has the following problems. That is, as the determination information for obtaining the final MQUANTj, except for the difference (divergence) between the target code amount and the code amount in equation (4), the Q scale reference value of the previous picture encoding result in equation (5) ( Qj) and the spatial activity (ACTj) in STEP3 processing are used, and the quantitative degradation of image quality and human visual characteristics are not fully considered in the code amount control in TM5. Accordingly, it is difficult to control the amount of code in accordance with human visual characteristics.

  Also, in the technology in Patent Document 1 that compensates for the problem with the TM5 algorithm, a large-scale circuit is required to calculate the “motion amount” that is an argument in the function F1. Therefore, in the case of a scene change or the like, there is a problem that the filter characteristics change abruptly and the image quality blur is conspicuous.

  According to Patent Document 2 that discloses a method for solving a problem of avoiding a sudden change when changing a filter coefficient as a moving image encoding method, first, encoding difficulty level Y is set as I, The function F (12) is used for each P and B picture.

Y = F (accumulated code amount, average Q scale) (12)
Next, a filter coefficient parameter Z is obtained from the encoding difficulty levels Yi, Yp, and Yb obtained for I, P, and B, respectively, using equation (13) (equation (13)).

Z = (Yi + Yp + Yb) / (bits_rate) (13)
Depending on the value of Z obtained by equation (13), selection is made from preset filter coefficients S0, S1, and S3 as shown in FIG. That is, a sudden change in the filter coefficient is avoided by giving a width to the value of Z corresponding to each filter coefficient.

  However, since the method according to Patent Document 2 simply predicts from information of only the accumulated code amount and the average Q scale, it is still difficult to say that the degradation of image quality and human visual characteristics are still taken into consideration. .

  In view of the above problems, an object of the present invention is to provide a moving image encoding apparatus and a moving image encoding method that take into consideration the degradation of image quality and human visual characteristics. The moving picture coding apparatus according to the present invention that solves the above-described problems and achieves the object mainly has the following configuration.

That is, an image encoding apparatus having a block encoding unit that divides an image into blocks and performs compression encoding, and a decoding unit that decodes the compression encoded data,
Filter means for varying the spatial filter characteristics by the filter coefficient;
Includes all pixels in the block except for the boundary pixel including the boundary pixel of the divided block from the image output by filtering the input image by the filtering unit and the decoded image output from the decoding unit And calculating means for calculating a difference between the images, and calculating a distortion degree of the decoded image based on the difference,
The encoding means receives an image filtered by the filter means, and a quantization means for varying a quantization characteristic according to a quantization scale;
Determining means for determining the filter coefficient and the quantization scale for controlling the spatial filter characteristic of the filter means and the quantization characteristic of the quantization means, respectively, according to the calculation result of the calculating means; With
Said determining means, in response to the distortion degree calculated by said calculating means, the greater the strain rate, cut-off frequency determines the filter coefficients such that a low,
Further, the determining means determines a value for correcting the value of the quantization scale in accordance with the degree of distortion calculated by the calculating means.

  According to the present invention, when controlling the code amount using the prefilter, the setting of at least one of the filter coefficient and the quantization scale is changed based on the level of block distortion obtained for each previous block. Thus, it is possible to realize filter control and code amount control for obtaining a high-quality decoded image without distortion reflecting human visual characteristics.

  Preferred embodiments of the moving picture coding apparatus and method of the present invention will be described below with reference to the accompanying drawings.

[First embodiment]
FIG. 1 is a block diagram of a moving picture coding apparatus 200 showing the configuration of a device that implements the moving picture coding method according to the present invention, and shows an MPEG coding unit 100 capable of executing the above-described TM5 algorithm. It is configured to use. The MPEG encoding unit 100 specifically corresponds to the MPEG-1, MPEG-2, MPEG-3, or MPEG-4 standards, and is not limited to a specific encoding standard. That is, the moving picture coding technique according to the present invention can be applied to any coding standard that includes a configuration for quantizing an input image (a configuration corresponding to QTZ 104 as a quantizer).

  The moving image encoding apparatus 200 further includes a block distortion degree calculation unit 109 and an encoding parameter determination unit 110. Based on the distortion degree of the image obtained by the block distortion degree calculation unit 109, the encoding parameter is determined. The determination unit 110 performs calculation for determining the quantization scale MQUANT for each macro block (MB), for each picture, or for a plurality of times within one picture. Hereinafter, the flow of the moving image encoding process will be described in detail with reference to the block diagram of FIG. 1 and the flowchart of FIG.

<Description of overall operation flow>
FIG. 5 is a flowchart for explaining processing in the coding parameter determination unit of FIG. First, in step S501, in order to use STEP1 of the above-described TM5 algorithm, initial values of a quantization scale (Q scale value) MQUANT and a filter coefficient value (C_LPF) are set.

  Next, the process proceeds to step S502, and target code amounts Ti, Tp, and Tb for the picture types (I, P, and B pictures) are calculated according to equation (3). In calculating the target code amount in this step, the target code amount of the next picture to be encoded is set. FIG. 12 is a diagram illustrating an encoding target picture. Here, images of each picture type (I, P, B picture) are represented as Xi, Xp, and Xb, respectively. Assuming that the P2 picture constituting the current GOP is the next encoding target, the target code amount for this picture is calculated.

  Next, the process proceeds to step S503, where a macro block (MB in FIG. 1) is input to the moving picture coding apparatus 200, and spatial filter processing is performed. Here, as a method for realizing the pre-filter 101, for example, a spatial filter of a 3 × 3 square matrix as shown in FIG. 6 can be used. FIG. 7 shows the relationship between the value of C_LPF and the filter coefficient of the 3 × 3 square matrix in this case. When C_LPF = 0, the filter operation is turned OFF, and the input image (MB) = the output image. In the example of the filter coefficient in FIG. 7, the cutoff frequency is set to be lower as the value of C_LPF is larger. By changing the filter coefficient that gives the characteristics of the pre-filter, it is possible to control the spatial filtering processing for the input macro block so as to match a predetermined photographing mode. For example, when the encoding parameter determination unit 110 determines that block distortion occurs from the obtained block distortion degree, the difference between the target code amount and the accumulated code amount up to the present time, the encoding parameter determination unit 110 sets the passband of the prefilter characteristics as the current Set to a lower frequency region.

  Here, the input pre-coding macroblock (MB) is subjected to spatial filter processing for block distortion calculation described later, and then input to the block distortion degree calculation unit 109. The processing in the block distortion degree calculation unit 109 will be described in detail later.

  In step S504, the MPEG coding unit 100 uses the quantization scale (MQUANT) value set as the initial value in the previous step S501 to quantize (104) the macro block subjected to the discrete cosine transform (103). ) To generate variable length encoded data (105). Note that the MPEG encoding unit 100 can be realized by processing conforming to the MPEG encoding standard, and therefore includes units relating to motion prediction (102) and motion compensation (108). The specific description about is omitted.

  Next, the process proceeds to step S505, and the encoded data generated by the process of step S504 is input to the local decoding unit 111, and reverse conversion processing is performed by IQTZ 106 and IDCT 107 to generate decoded data. Note that the local decoding unit 111 can be realized by processing conforming to the MPEG encoding standard, and therefore a specific description of each unit is omitted here.

  FIG. 13 is a block diagram showing a configuration of the block distortion degree calculation unit 109. In step S506, the block distortion degree calculation unit 109 stores the macro block before encoding (in step S503, the macro block after spatial filter processing is stored in the pre-encoding data storage unit 109a of the block distortion degree calculation unit 109. And the macro block decoded by the local decoding unit 111 input via the decoded data input unit 109b, the block distortion degree calculation unit 109c compares the image generated by the MPEG processing. The block distortion degree is calculated as a parameter for evaluating distortion. As a block distortion calculation method, for example, the following two methods can be used.

<Method 1>
Method 1 is a method for obtaining a PSNR (Peak Signal to Noise Ratio) for two images before encoding and after decoding. The luminance component of the input image to the MPEG encoding unit 100 is Pj, and the local decoding unit 111 is used. If the luminance component of the output image is Rj, PNSR can be obtained by the following equation (14).

SUM = Σ (Pj−Rj) ² (j = 0 to 255)
PSNR = 20 × log10 (255 / sqrt (SUM / 256)) (14)
By evaluating the PSNR obtained from the equation (14), the degree of distortion between the input image and the output image can be obtained relatively.

<Method 2>
In Method 2, regarding the two images before encoding and after decoding, each of them is an 8 × 8 block, and a difference sum calculation is performed according to Equation (15) for each pixel at the boundary of the 8 × 8 block. FIG. 8 is a diagram exemplarily showing a macro block (FIG. 8 (b)) and boundary pixels (FIG. 8 (a)) of the 8 × 8 pixel blocks constituting the macro block (FIG. 8 (a)). For each boundary pixel of the four 8 × 8 blocks constituting the macro block, the luminance components (P0j, P1j, P2j, P3j) of the input image to the MPEG encoding unit 100 and the output of the local decoding unit 111 As the luminance components (R0j, R1j, R2j, R3j) of the image, the difference sum (BN) between them can be obtained by the equation (15).

BN = Σ (P0j−R0j) + Σ (P1j−R1j) + Σ (P2j−R2j) + Σ (P3j−R3j)
(15)
(J takes the shaded boundary pixel number in FIG. 8)
The block distortion degree calculation unit 109c can calculate the block distortion degree by the method 1 or the method 2 described above.

  Returning to the flowchart of FIG. 5, in step S507, from the block distortion degree (BN) obtained by the equation (15) and the code amount output from the VLC (variable length encoder) 105 in the MPEG encoding unit 100, The filter coefficient value (C_LPF) and Q scale value (MQUANT) values to be used for the next macro block are determined. This determination process is executed by the encoding parameter determination unit 110 of the moving image encoding apparatus 200, and this specific process will be described in detail later with reference to FIGS.

  By repeating the above-described steps S502 to S507 for all the macro blocks in the picture (S508, S509), the code amount can be controlled. Further, the above processing may be repeated for each macro block (MB), for each picture, or for a plurality of times within one picture.

<Description of operation of encoding parameter determination means>
Next, regarding step S507 in FIG. 5, the flow of processing will be specifically described with reference to the flowchart in FIG. The encoding parameter determination unit 110 that executes step S507 includes an encoding parameter calculation unit 110a, a filter coefficient determination unit 110b, and a quantization scale determination unit 110c as illustrated in FIG. 110a calculates coding parameters used for moving picture coding, using the calculation result of the block distortion degree, the code amount, and the target bit rate as input values. Based on the calculation result, the filter coefficient determination unit 110b and the quantization scale determination unit 110c respectively set the filter coefficient (C_LPF) set in the prefilter 101 and the quantization scale set in the quantizer (QTZ) 104. (Q scale) Determine MQUANT.

  In step S901 in FIG. 9, the encoding parameter calculation unit 110a acquires the AREA variables (0 to 2) corresponding to the three areas as shown in FIG. 10 based on the block distortion degree value (BN). Here, the value of the block distortion degree (BN) takes a value from 0 to 28560 when one pixel is expressed by 8 bits from (15) (expression). C_BN0 and C_BN1 in FIG. 10 are setting values for dividing the distortion degree (BN) into three AREAs, and the magnitude relationship between these setting values and the distortion degree (BN) obtained by calculation is as follows. As shown, the encoding parameter calculation unit 110a determines which AREA the distortion degree (BN) corresponds to (S1 to S6).

IF (BN <C_BN0) THEN (S1)
AREA = 0 (S2)
ELSE IF (BN <C_BN1) THEN (S3)
AREA = 1 (S4)
ELSE (S5)
AREA = 2 (C_BN1 ≧ C_BN0) (S6)
At this time, the reference values C_BN0 and C_BN1 for dividing the AREA are set in advance before the input image is input to the moving image encoding apparatus 200 according to the embodiment of the present invention. Next, the encoding parameter calculation unit 110a executes the following processing according to each obtained AREA.

  In step S902, if AREA = 0 (S902-Yes), the process proceeds to step S906. At this time, since the filter coefficient determination unit 110b and the quantization scale determination unit 110c have a small value of the block distortion degree BN of the immediately preceding macro block, the quantization scale obtained by STEP2 of the spatial filter processing by the prefilter and the TM5 algorithm The value of the reference value Qj (see equation (6)) is used as it is to obtain the quantization scale MQUANT.

  In step S903, if AREA = 1 (S903-Yes), the process proceeds to step S904, and AREA1 is further divided into two regions by parameter C_BN2 (see FIG. 10B). The encoding parameter calculation unit 110a predicts whether or not the block distortion is visually noticeable by the following method (S7 to S14).

IF (BN <C_BN2) THEN (S7)
WARN_BN = 0 (S8)
ELSE (S9)
IF (WRAN_BN_COUNT> C_BN_COUNT) THEN (S10)
WARN_BN = 1 (S11)
ELSE (S12)
WARN_BN = 0 (S13)
(C_BN1 ≧ C_BN2 ≧ C_BN0) (S14)
Here, the parameter C_BN2 for further dividing AREA = 1 into two regions is set in advance in the same manner as the parameters C_BN0 and C_BN1. WARN_BN is a parameter for specifying that the degree of block distortion is large and is in a warning state. In step S905 of FIG. 9, the encoding parameter calculation unit 110a determines the value of this parameter.

  If WARN_BN = 0 (S905-No), the same processing as in AREA = 0 is performed (S906). If WARN_BN = 1, whether to perform the same processing as in AREA = 0 or not. Make a decision.

  Here, the number of times that the BN value in one horizontal scan for the macro block in the past is larger than the value of C_BN2 is set as “WARN_BN_COUNT”, and the determination is made based on whether or not it is larger than a preset constant C_BN_COUNT. If WARN_BN = 0 (S905-No), the process proceeds to step S906, and if WARN_BN = 1 (S905-Yes), the filter coefficient of the pre-filter 101 is determined to be in a warning state because the degree of block distortion is large. Coefficients are set so as to be changed (S907). In step S907, processing for changing the value of the filter coefficient (C_LPF) so as to reduce the block distortion is performed, and the quantization scale (MQUANT) is set to the quantization scale reference value Qj in the same manner as in step S906. The value is used as it is to obtain the quantization scale MQUANT.

  The filter coefficient determination unit 110b uses the function GET_F (BN) for obtaining the filter coefficient (C_LPF) and the relationship between the block distortion degree (BN) and the parameter C_BNi (i = 2 to 5) based on the data table shown in FIG. And the corresponding filter coefficient (C_LPF) is specified. Here, the specified filter coefficient (C_LPF) is set in the pre-filter 101 (S907).

  Note that when the degree of block distortion (BN) ≦ C_BN2, the processing is the same as that in step S906, and in this case, the filter coefficient C_LPF = 0.

  If C_BN1 <block distortion (BN), the processing is the same as that of AREA2 in step S908 described later.

  By determining whether or not the degree of block distortion should be warned in this way, it is possible to avoid in advance occurrence of visually noticeable block distortion.

  On the other hand, if the block distortion degree (BN) corresponds to AREA = 2 in the process of step S903 (S903-No), the process proceeds to step S908. Since this AREA is an area where the degree of block distortion (BN) is large, the quantization scale (MQUANT) is changed here in addition to the setting of the filter coefficient for changing the spatial filter processing. As with the processing in step S907, the filter coefficient setting unit 110b uses the data table as shown in FIG. 11B to set the filter coefficient value (C_LPF) corresponding to the corresponding block distortion degree (BN). ) = Ci (i = 1 to 4) is specified and set in the pre-filter 101. The filter coefficient set here is set so as to realize the characteristics of the pass band in the low frequency region as the calculated block distortion level increases.

  Further, the quantization scale determining unit 110c further adds a constant ADD_Qi (i = 1 to 4) as shown in FIG. 11B to the quantization scale reference value Qj obtained in STEP2 of the TM5 algorithm. It is specified according to the block distortion (BN). The constant ADD_Qi specified here is added to the reference value Qj of the quantization scale and set in the quantizer (QTZ) 104. Here, the reference value Qj of the quantization scale is given by the equations (4) and (5) in STEP 2 in the TM5 algorithm, and is output from the target code amount (target bit rate) and the VLC 105 in the MPEG encoding device 100. It is calculated based on the code amount. The quantization scale is changed by adding a value (ADD_Qi (i = 1 to 4)) corresponding to the level of the block distortion value to this reference value Qj, and the block distortion is controlled in the code amount control for the next picture. Information is reflected.

  Note that the parameters C_BN3 to C_BN8 in FIGS. 11A and 11B are set in advance in the same manner as the parameters C_BN0 and C_BN1 for dividing the AREA. In this way, when the block distortion degree is in a large region, it is possible to effectively avoid the occurrence of visually noticeable block distortion in advance by changing the filter coefficient and the quantization scale.

  As described above, according to the present embodiment, when controlling the code amount using the pre-filter, based on the level of block distortion obtained for each previous block, the filter coefficient and the quantization scale are used. By changing at least one of the settings, it is possible to realize filter control and code amount control for obtaining a high-quality decoded image without distortion that reflects human visual characteristics.

[Other Embodiments]
In addition, as described above, the object of the present invention is to provide a storage medium storing a program code of software that realizes the functions of the embodiment to a system or apparatus, and the computer of the system or apparatus (or CPU or MPU) stores it. Needless to say, this can also be achieved by reading and executing the program code stored in the medium. In this case, the program code itself read from the storage medium realizes the functions of the above-described embodiments, and the storage medium storing the program code constitutes the present invention.

  As a storage medium for supplying the program code, for example, a floppy (registered trademark) disk, hard disk, optical disk, magneto-optical disk, CD-ROM, CD-R, magnetic tape, nonvolatile memory card, ROM or the like is used. be able to.

  Further, by executing the program code read by the computer, not only the functions of the above-described embodiments are realized, but also an OS (operating system) running on the computer based on the instruction of the program code. Needless to say, some or all of the actual processing is performed and the functions of the above-described embodiments are realized by the processing.

  Furthermore, after the program code read from the storage medium is written to the memory provided in the function expansion board inserted into the computer or the function expansion unit connected to the computer, the function is based on the instruction of the program code. It goes without saying that the CPU or the like provided in the expansion board or the function expansion unit performs part or all of the actual processing, and the functions of the above-described embodiments are realized by the processing.

It is a block diagram of the moving image encoder 200 which shows the structure of the apparatus which implement | achieves the moving image encoding method concerning this invention. It is a figure explaining STEP2 of a code amount control algorithm (TM5). It is a figure explaining the balance function shown in patent documents 1. It is a figure explaining the process for determining the filter coefficient in a prior art. It is a flowchart explaining the flow of the moving image encoding process concerning embodiment of this invention. It is a figure which illustrates the spatial filter of 3x3 square matrix. It is a figure which shows the relationship between the value of a filter coefficient (C_LPF), and the filter coefficient of a 3x3 square matrix. It is a figure which shows illustratively the boundary pixel of a macroblock and the 8x8 pixel block which comprises this. It is a flowchart explaining the determination process of the encoding parameter concerning embodiment of this invention. It is a figure explaining three area | regions (AREA) which classify | categorize block distortion degree BN. It is a figure which illustrates the structure of the data table which shows the relationship of the constant (ADD_Q) for adding to a filter coefficient (C_LPF) and quantization scale value (Q) corresponding to a block distortion degree (BN). It is a figure which illustrates the encoding object picture. 3 is a block diagram showing a configuration of a block distortion degree calculation unit 109. FIG. 2 is a block diagram showing a configuration of an encoding parameter determination unit 110. FIG.

Claims (4)

In an image encoding apparatus having block encoding means for dividing an image into blocks and compression encoding, and decoding means for decoding the compression encoded data,
Filter means for varying the spatial filter characteristics by the filter coefficient;
Includes all pixels in the block except for the boundary pixel including the boundary pixel of the divided block from the image output by filtering the input image by the filtering unit and the decoded image output from the decoding unit And calculating means for calculating a difference between the images, and calculating a distortion degree of the decoded image based on the difference,
The encoding means receives an image filtered by the filter means, and a quantization means for varying a quantization characteristic according to a quantization scale;
Determining means for determining the filter coefficient and the quantization scale for controlling the spatial filter characteristic of the filter means and the quantization characteristic of the quantization means, respectively, according to the calculation result of the calculating means; With
Said determining means, in response to the distortion degree calculated by said calculating means, the greater the strain rate, cut-off frequency determines the filter coefficients such that a low,
Furthermore, the determination means determines a value for correcting the value of the quantization scale according to the degree of distortion calculated by the calculation means. An image encoding method comprising a block encoding step of dividing an image into blocks and compression encoding, and a decoding step of decoding the compression encoded data,
A filter step for varying the spatial filter characteristics by the filter coefficient;
From the image output by filtering the input image by the filtering step and the decoded image output from the decoding step, all pixels in the block including the boundary pixels of the divided block and excluding the boundary pixels are obtained. Calculating a difference between images not included, and calculating a distortion degree of the decoded image based on the difference,
The encoding step takes the image filtered in the filter step as an input, and a quantization step for varying a quantization characteristic by a quantization scale;
A determination step for determining the filter coefficient and the quantization scale for controlling the spatial filter characteristic of the filter step and the quantization characteristic in the quantization step, respectively, according to the calculation result of the calculation step; Have
The determination step in response to the distortion degree calculated in said calculation step, the greater the strain rate, cut-off frequency determines the filter coefficients such that a low,
Furthermore, the determination step determines a value for correcting the value of the quantization scale in accordance with the degree of distortion calculated in the calculation step.   A program causing a computer to execute the image encoding method according to claim 2.   A computer-readable storage medium storing a program for causing a computer to execute the image encoding method according to claim 2.

Images