TW200425744A - Quantization matrix adjusting method for quality improvement - Google Patents

Abstract

A quantization matrix adjusting method for quality improvement by scaling down the default quantization matrix. The adjusting method comprising the steps of: updating a used bits BB_X and an average quantizer scale Avg_Q; allocating the bit budget BB_C for the current picture; calculating an estimated quantizer scale Est_Q according to the used bits BB_X, average quantizer scale Avg_Q and the bit budget BB_C; deciding the operation mode, the encoder entering an adjusting mode when the estimated quantizer scale Est_Q is smaller than a threshold Th_Q, otherwise the encoder remaining at normal mode; scaling down the default quantization matrix as a new quantization matrix when the operation mode is the adjusting mode; and coding the current picture, the encoder coding the current using the new quantization matrix and default quantization matrix in the adjusting mode and normal mode, respectively.

Description

200425744

[Technical field to which the invention belongs] Matrix adjustment method, special-quality quantization matrix adjustment The present invention relates to quantization for improving image quality, not to mention reducing quantization matrix to improve image adjustment method.口 / 2. [Previous technology] In order to reduce the amount of image data, images are usually reduced in size, such as MPEG format image coding. The basic unit of image coding is a macro block (macr0 — bloc). If sampling

(sampling) is in 4: 2: 0 format, then each macro block has 6 blocks ′, which are 4 Y blocks, one cb block, and one cr block. Each block must first undergo a discrete cosine transform (hereinafter referred to as DCT), and then the DCT coefficients (coefficient) are quantized into integers. A zigzag scan or other alternative scanning method is used to arrange a two-dimension quantization coefficient array into one-dimension data. Finally, the so-called variable-length coding (VLC for short) is used to perform entropy coding. Generally, the quantization of the 'DCT coefficient F [v] [u] can be expressed as: QF [v] [u ] = 16 * F [v] [u] / (Q * W [v] [u]) ... (1) where v and u are indexes of the two-dimensional array, ranging from 0 to 7. Q is the quantizer scale of each block in each macro block, which can be changed with different macro blocks. W [v] [u] is the whole picture

Page 6 438 200425744 V. Description of the invention (2) The quantization matrix (qUantization matrix) defined by the picture is used to specify the weighting factor of each different DCT coefficient. Figure 1 shows the preset quantization matrix defined in the MPEG-2 intra block. As shown in the quantization matrix of Fig. 1, the value of the quantization matrix 耆 array index v, u increases gradually and becomes larger. This characteristic makes quantization coefficients with high frequencies easier to become 0, because human eyes are more sensitive to low-frequency signals and less sensitive to high-frequency signals. In addition, 'inverse-quantization' is used to recover DCT coefficients. The inverse quantization of the commonly quantized DCT coefficients QF [v] [u] is defined as: F '[v] [u] = QF [v] [u] * Q * W [vnu] / 16… (2) where F '[v] [u] is the DCT coefficient of the reply. The difference between the original DCT coefficient F [v] [u] and the returned DCT coefficient F, [v] [u] is called the quantization error and is defined as: E [v] [u] = F [v] [u]-Γ [v] [u]… (3) and the quantization scale factor Q is used to adjust the content of the quantization matrix W [v] [u]. The value of Q ranges from 1 ~ 31. The adaptive quantization applied to MPEG-1 and MPEG-2 can change the value of each macroblock at will, but it can only be slightly (31 丨 2) 11 at ^^ £ 6-4: 17) Adjust the Q value of the adjacent macro block. Furthermore, MPEG-2 has an actual quantization scale factor that corresponds to a non-linearity from (1 to 3 1) to (0.5 to 5 6). Adjusting the quantization scale factor Q can control the bit consumption and coding quality of the macro block. for

Page 7 43! 200425744 V. Explanation of the invention (3) A larger Q value will make the quantized DCT coefficient smaller, which will cause more quantified occ coefficient to be 0. Therefore, the coded bit stream (codded bi-seven stream) after VLC processing will be shortened. However, as a result, the quantization error becomes larger, resulting in poorer image quality after decoding. If we want to get better image quality, we must set a smaller Q value to reduce the quantization error, but the encoding bit stream will become longer. In general applications, the bit rate of an MPEG bit stream is constrained. For example, in the standard of Digital Versatile Disk (DVD), aMpEG-2 video bit rate cannot be higher than 98 intestines per second, so the bit consumption must be controlled during the encoding process. To meet this limitation. The content of each image in a series of visual glycoside pictures varies, and the complexity of the coding varies from part to part. The video buffer verifier (VideV buffer verifier (hereinafter referred to as VBV)) used in M ^ PEG-1 has the flexibility to provide bit rate consumption for different images (: ^ 4 乂; 11) 丨 1 "7) However, the overall bit rate for a long time is still a fixed bit " 1 rate ^ constant bit rate (hereinafter referred to as CBR). And MPEG-2 introduced a variable = variable bit rate (hereinafter referred to as VBR) operation mode ^ 丄 f provides more flexibility for the change of the bit rate consumption of each image. The VBV buffer is used to emulate the input buffer of the MpEG decoder and the bit stream generated by the MPEG encoder cannot be violated ( vi〇late) νπ 纒 the limiter of the buffer, otherwise the bit stream will not be able to be decoded properly. Second = the schematic diagram of the device in the fixed mode. If the _ ^ knife system table does not enter the buffer at the fixed bit rate The vertical line is 200425744. V. The description of the invention (4) part refers to the data of the decoded image that is removed from the buffer instantaneously. In order to make the bit stream meet the requirements of MPEG bit control, when the decoder needs to decode the image The image data must be stored in the buffer. And the decoder's buffer cannot be overfull (overf i 1 1). Referring to Figure 2, if an image consumes too many bits, the VBV buffer may be under-informed (underfl0w), so the bits can be allocated The upper bound of (all ocat able bits) is UB. Similarly, if an image consumes too few bits, the VBV buffer may overflow the data (overf l0w), so the bits can be allocated The lower limit of the (allocatable bits) is LB. Figure 3 shows a schematic diagram of the VBV buffer in VBR operating mode. The difference between vBR and CBR operating modes is that VBR operating mode is input at the specified maximum bit rate νβν The buffer is until the VBV buffer is fully loaded, so the average bit rate input rate of the VBr operation mode is variable. Therefore, VBR has no limit on the minimum bit rate, only the limit that can allocate the maximum bit rate UB. In order to meet the VBV The encoder must allocate a bit budget to each image, and then try to control the actual consumed bits close to the bit pre-exception. Generally, the virtual buffer mechanism (v 丨 r 丨 work) is used. buffer mechanism) Control bit consumption. When the buffer occupancy is low = 0, the quantization scale factor Q is corresponding to!; And when the buffer occupancy = amount is higher than a gate value R (threshold), the quantization scale factor q is corresponding. Is 31. The threshold R is generally called the response factor (reacti0n) Before encoding an image, the virtual buffer occupancy D is initially d 0 'The initial quantization scale factor q is q 〇, ! ^ = r * q 〇 / 31. The encoder uses the quantization scale factor q to perform the amount of the first macro block 200425744

If the amount of bits consumed by a macro block is greater than the average bit, the virtual buffer occupancy D increases, otherwise the virtual buffer occupancy increases if the virtual buffer occupancy D exceeds d 〇 + R / 31, for example, the factor Q is set Into q〇 + 1. This situation means that the bit is disappearing, so increase the quantization scale factor Q to try to reduce the bit. The virtual buffer occupancy D decreases below d0-R / 31, and the sub-Q is set to qO-1. This situation means bit consumption, so reducing the quantization scale factor Q to try to increase the bit buffering mechanism to control the bit consumption is how to refer to the initial occupation of d 0. When the initial occupancy of the edge virtual buffer is reduced and the Q budget value is encoded, the quantity D is reduced (then the quantization ratio consumption exceeds the consumption ratio. If the quantization ratio is lower than the expected value consumption ratio.

Use virtual dummy buffer

It can be regarded as the estimated value of the coding complexity of the current picture (e s t i m a t i 〇 n). For a fixed bit budget value, if the complexity of the current image is relatively high, you must choose a higher initial occupancy of the virtual buffer d0, so that the quantization scale factor Q of each macro block Relative improvement. However, the encoder does not know the complexity of the current image until the image is actually encoded. One solution is to inherit the virtual buffer occupancy of the previous image. This simple method can handle most situations without serious problems, but when the content of a series of video images changes rapidly, VBV data will be insufficient. The range of the quantization scale factor Q of each macroblock is from 1 to 31, and this range can be in most of the conditions of the video coding. However, when a series of videos are simple and the target bit rate is high, the quantization scale factor will become very small to generate more bits. Even when the video is really too simple,

Page 10 200425744 V. Description of the invention (6) Even if the quantization scale factor Q is set to 1, it cannot generate enough bits to increase the stuffing bits (rate). In addition, when the average quantization scale factor Avg-q is 丨 · 5 points, the most appropriate quantization scale factor should be j. 1 or 2 must be selected. Therefore, when the value of the quantization scale factor becomes coarser (c0 ar s er), the result will cause the quantization to change between 1 and 2, making the flat area of the image produce blocking artifacts. Enough bits to waste bits, for most of the five, but it must be smaller, its solution factor quickly generates a slightly visible purpose, listed steps: the previous image refers to the measurement using bit estimation The scaling factor is set to adjust the encoding mode. The coding is performed on the matrix. [Abstract] In view of the above problems, the purpose of the present invention is to provide a coarser scaling factor resolution, and at the same time can improve the amount of image quality Use budget value; calculate average quantization ratio; determine whether a quantization ratio is due to the factor factor of the whole method, meta-budgeting formula, and then the encoding quantization moment is reduced; after the little ones are preset to achieve the upper and lower inclusion; the calculation is based on the value The estimation measurement mode is set up, and the quantization update of the moment and the quantization update has positioned the element quantity and the scale factor are less than the mode, in order to adjust the Dangma Shima'dang encoding. Otherwise, when the mode is maintained, the weighting formula is “Mama J.” The “pre-adjustment mode” is a quantization matrix that can avoid qualitative quantization matrices and the average quantization ratio. The measurement scale factor and the bit entry. When adjusting the threshold of the mold, the general mode; when adjusting the set quantization matrix formula, according to the normal mode, the root is 200425744.

Because the invention reduces the preset quantization matrix when adjusting the mode, it can avoid the estimation scale factor being too low, thereby improving the negative image quality of the encoded image and avoiding the slight visible blocking effect in the flat image area. 4. [Embodiment] The method for adjusting the quantization array for improving the image quality of the present invention is described in detail below with reference to the drawings. ^ When the calculated estimated scaling factor EstQ-C of the current image becomes small and the current image is encoded with a preset quantization matrix, the Hiragaki image area will have a slightly visible blocking effect. In order to solve this problem, the encoder of the invention uses a reduction of the weight of the quantization matrix (heart; [handsome 1:] [call]) to adjust the estimated scale factor EstQ-C to a larger value. For example, as shown in the formula (1), when the preset quantization matrix w [v] [u] is reduced to 1/4, the average quantization scaling factor Avg_Q can be enlarged from 1.5 to 6 at this time. The specific factor Q can be increased to 7 or 8 to reduce the bit rate consumption, or 5 2 to increase the bit rate consumption. Therefore, the quantization of neighboring macroblocks = greatly reduced. Differences In the MPEG-1 and MPEG-2 specifications, a sequence of header data must be configured for each encoded bit stream, and the stick: is added to the subsequent bit stream. Once the decoder receives the header data, "the two-dimensionalization matrix W [v] [u] will be set to the default value. In MpEG-1, the quantization matrix W [v] [u] of the customized You can only set j j in the header to indicate that the header can only be inserted before a group of pictures (gr 0up 0 fp ^ c sevenures, ρ and GOP). This means that when the decoder wants to change the quantization matrix M Under Jane, Calcite Horse

Page 12 436 200425744 V. Description of the invention (8) The device must be restarted-a G0P 'inserts a header data before G0P, and a new quantization matrix is configured in the header lean material. And in MPEG_2, the new matrix can not only be configured in the header data, but also can be inserted-the epitaxial amount H ::::: If solution ... When changing the quantization matrix 'must be shown in each image Figure 4 shows the improved image of the present invention Quality quantization matrix adjustment method-a flowchart of the embodiment. The flow of this embodiment is that the encoder obliquely analyzes the image in advance to perform the encoding operation of the current image. Because there is no complexity of scanning and analyzing the image in advance, that is, the image belongs to the image, so = TEST model 5 (test m〇del 5, TM5) bit rate control = = rate control aig〇rithm, encoding The device can only use the previous load = image to estimate the complexity of the current image. Step ^ code of the first embodiment Step S402: Update the used bit amount Ββ —χ and the square factor Avg_Q. After a χ image (χ can be a! Image, ρ image, or encoding, the encoder obtains the average quantization ratio factor Avg_Q for the χ image, the final virtual buffer occupancy 乜 χ, and the ββ_χ Q Curse, step S404: Specify a bit budget value BB 乂 for the current image. In the situation, the initial virtual buffer occupancy of the next image ⑽: is set to the final virtual buffer occupancy d — x, And the initial quantization ratio is combined, and it is set to the average quantization scale factor Avg-Q. Moreover, the frame-levei rate contr0 refers to the budget value of 70 yuan BB_C. ⑺The position of the image

Page 13 437 200425744 V. Description of the invention (9). Step S406: Calculate the estimated scale factor EstQ_C. Because adjacent images of the same format have approximately the same degree of activity, the decoder uses the used bit amount BB_X of the previous image, the bit budget value BB_C of the current image, and the average quantization scale factor Avg_Q to Calculate the estimated scale factor EstQ_C of the current image:

EstQ—C = (BB—X / BB—C) * Avg—Q… (4) Step S408: Detect whether the adjustment mode is entered. When the measured scale factor EstQ-C is less than a scale factor threshold ThQ_X, for example, ThQ_X is 2, 3, or 4, it means that the image content of the video is quite simple for the given bit rate. If the used bit amount of the previous coded image and _ and the bit budget value assigned to the current image are approximately equal, it can be guessed that the bit budget value BB_c assigned to the current image will cause an estimated measurement ratio — The factor EstQ-C is too small. Therefore, when the estimated scale factor EstQ_c is smaller than the scale factor threshold value ThQ-X, the encoder enters the adjustment mode of step S4i0 '; otherwise, it is processed in the general mode of step S416. Step S4 1 0: Reduce the quantization matrix into a new quantization matrix. In the adjustment mode, the 'encoder will hope to increase the final average quantization scale factor of the current image to a safety value Sq —X, and the safety value SQ — χ may be greater than or equal to the threshold value ThQ —X. Therefore, the present invention will use the adjustment factor S to reduce the preset quantization matrix, and the true initial virtual buffer occupancy will be increased to do0j / s. The adjustment factor s is a value less than 1, and can be calculated by the following formula: s = EstQ—C / SQ—X… (5) and the new quantization matrix W, [v] [u] is:

200425744 V. Description of the invention (10) W ’[v] [u] = W0 [v] [u] * s (6) where v and u are indexes of 0 ~ 7. W0 [v] [u] can be a preset quantization matrix or a user-defined quantization matrix. Therefore, the larger the safety value § (0, the smaller the adjustment factor S. Since the minimum value of the quantization matrix is], the minimum value of the content of w '[v] [u] is limited to 1. In addition, the encoder Will use two quantization matrices, one is for the intra block, and the other is for the inter (^) block so that ::: Updating a quantization matrix requires 51 2 bits to encode the quantization The matrix itself and some extra bits, so in the case of limited bit usage: ::, the amount of quantization matrix updates is reduced. For example, Λ: only the internal quantization matrix, and the B image may only be the matrix ΐ μ 口 1 豕 J This is just enough to update the amount of interrelationship. For p-images, it is necessary to rely on the existence of a threshold #mega block for the image. If the internally encoded macro block is greater than one, otherwise It is only necessary to update the internal quantization matrix and the interrelated quantization matrix, and /, and update the interrelated quantization matrix. Edit i_12: Use the new quantization matrix [[v] [u] to develop the current image, ΞΓ 回 41 牛 4: Check whether the current image is encoded. End Step S V Step 3402 ′ If it has not been completed, go back to step 10. Perform i-code. 416: Use the preset quantization matrix w [v] [u] for the current image

If step S 4 1 D is completed, then skip to the second step to check whether the image is coded '. If the stone horse editing is completed at step 8402, if not, skip back to step S41. In the step, if the encoded X image enters the adjustment mode, the amount

d & Q 200425744 V. Description of the invention (11) 1 Dagger matrix: reduced to 1/8 times' Since the macro blocks are quantized with a new quantization matrix, the average Q ratio of these macro blocks is being calculated. Shrink / 丨, ·! / Q LU ". 丄 — Wuyi 3 will # 告, 告 斜 + Encoding complexity. — Tian Zi in the relative image of the adjustment factor S is not much different, then: mother: image change quantization matrix It is not a good way to reflect this small difference: because: the difference in the adjustment factor s can exceed the change of the row quantization matrix. In addition, there is a solution is to preset the adjustment factor into the range S and calculate Fortunately, the corresponding quantization adjustment factor S0 can be set to four U, U, 0.4, 0.2, etc.] The pre-amplification quantization matrices are W1, W2, W3, and W4, so when the adjustment factor is set to 0.67, At 0.62, 0.54, 0.61, or 0.58, the update amount is W2, and only the first image needs to encode the quantization matrix w2 in place: the stream, and the rest of the images also use the quantization matrix W2 Until the header changes. Figure 5 shows a flowchart of a second embodiment of a quantization matrix adjustment method for improving image quality according to the present invention. The flow of this embodiment is that after the encoder analyzes the image in advance, #the current image encoding operation is performed. The encoder can: pre-analyze the image to collect the f signal of the complexity estimate, This coding more accurately decides whether to change the quantization matrix and how to change it. In general, the complexity of the current image can be estimated by performing motion estimation (0101: 1011 and optimization). For the I image, you can take the difference of the internally encoded macro block. For p and b ^, each a 'must be based on the difference between the internal encoding and the correlation encoding mode. Mode decision. The sum of the differences is generally referred to as the activity of the pre-advance image. Before image coding, 3 edits

448 200425744 V. Description of the invention (12) The encoder can obtain the activity level Act-c of the image. After image coding, the average quantization scale factor Avg-Q and the actual bits BBj can be used to calculate the complexity Com of the image. When the encoder cannot analyze the current image in advance, the encoder only decides whether to change the quantization matrix and how to change it according to the previous decoded image. However, when the image content of two adjacent pictures changes rapidly, the decision of this method may be wrong. For example, when the horizontal quantization scale factor Avg-Q of the current image is only slightly lower than the scale factor threshold value ThQ-χ, FIG. 2 will not enter the adjustment mode at present. However, if the encoding complexity of the current image is equal to the complexity of an image, the bit rate of the current image after encoding will be far lower than the bit budget value, the bit rate will be too low. Conversely, if the complexity of the previous image is relatively low, but the complexity of the current image is quite high, an overly reduced quantization matrix will cause the image quality to be greatly attenuated. Here, the difference between the method of the second embodiment and the method of the first embodiment is that the embodiment estimates the activity level of the current image Act-C in advance, and then estimates it based on A / tongue movement Act-C. Calculation of quantized scale factor estQ c. The calculation of the estimated scale factor EstQ —C is as follows: — °

EstQ—C = (BB—X / BB—C) * Avg—Q * (Act—C / Act_X) ... (7) The difference between formula (7) and formula (4) is that formula (7) is calculated and measured In the case of the scale factor EstQ-C, in addition to the bit budget value and the average quantified scale factor Avg-Q, the activity level Act is also considered. That is, when the bit budget value is equal to the average quantization scale factor Avg-Q, if the current degree of activity of the image is higher than that of the previous image, the calculated estimated scale factor EstQ is calculated. A C will be relatively improved, so that the current image encoding

200425744 V. Description of invention (13) The bit rate after (13) will not exceed the bit budget value too much. Conversely, if the activity of the current image is lower than that of the previous image, the calculated estimated scaling factor EstQ_C will be relatively reduced, so that the bit rate of the current image after encoding will not be low. Too many budget values. After the estimated measurement scale factor Es tQ_C is calculated, the second embodiment is the same as the first embodiment, and both use the estimated measurement scale factor EstQ-C to determine whether to enter the adjustment mode. When the measured scale factor EstQ-C is less than the scale factor threshold ThQ-X, it enters the adjustment mode, otherwise it is processed in the general mode. After entering the adjustment mode, the encoder reduces the preset quantization matrix into a new quantization matrix. In the adjustment mode, the encoder will hope to increase the final average quantization scale factor of the current image to a safety value Sq — x, and the safety value SQ-X may be greater than or equal to the scale factor Π threshold ThQ-X. Because it will use an adjustment factor (ad〗 ustmentfac 70r) $ to reduce the preset quantization matrix, and the initial value of the initial virtual buffer occupancy is reduced to d0: X / S. This adjustment factor S is a value less than 1 and can be calculated by equation (5). The new quantization matrix w, [v] [u] can be calculated by equation (6). Since the minimum value of the quantization matrix is i, w, The minimum value of the content of [v] [u] is limited to!. The subsequent actions and the first embodiment will not be described repeatedly. However, the present invention is not limited by this. The scope of the present invention can be modified or changed without departing from the scope of the present invention.

Page 18 200425744 Brief description of the drawings ~ Figure 1 shows the preset quantization matrix defined for the internal blocks in the MPEG-2 specification. Figure 2 shows the VBV buffer in CBR operation mode. FIG. 3 shows the VBV buffer in VBR operation mode. FIG. 4 shows a flowchart of a first embodiment of a quantization matrix adjustment method for improving image quality according to the present invention. FIG. 5 shows a flowchart of a second embodiment of a quantization matrix adjustment method for improving image quality according to the present invention.

Page 19

Claims (1)

200425744 6. Scope of patent application • A quantization matrix adjustment method for improving image quality. Sun, 7, Encoder to a polygraph m Chu, 隹 — ^ ^ Ping Weng Fang Fang is a series of arrays, 1 仃, 扁Code, the encoder has a preset quantization moment, and the 5-week integer method includes the following steps: J new used bit quantity and average quantization scale factor; assigning a bit budget value to the current image; different estimates of the measurement scale The factor is based on the former 沭 P 佶 田 A-the above-mentioned average quantization scale factor, and Yu, X, and Z. According to f, we use the quantified and scale factor; Adjustment mode, when the aforementioned estimated scale factor looks like two = scale factor threshold value, the == sub-low form, otherwise it stays in the general mode; 僎 π- is again the adjustment modulus = quantity: matrix, when the encoding mode is adjustment e reduces the localization matrix; and encodes the other image in advance, when the encoding mode is the adjusted quantization matrix for image encoding, when the encoding mode: 'according to reduction, the image is performed according to a preset quantization matrix Encoding, ^ In the general mode, 2 · The method for adjusting the improvement matrix as described in item 丨 of the scope of patent application, in which the foregoing calculation estimates the measurement; in the qualitative quantity, the foregoing evaluation measures the scaling factor = (the step of using the New Year's Day factor) * Average quantization scale factor. Median budget value) 3. Adjust the quantization moment ^ and the step image quality adjustment factor as described in the modification matrix adjustment m described in item 2 of the patent application scope to reduce the preset ones. Quantization matrix. ", is based on 4. The quality of D-shirt image quality as described in item 3 of the scope of patent application, page 20 444 200425744 6. Method of adjusting the scope of patent application matrix, 1 The adjustment factor of Zhongyu is the aforementioned estimate & In the step of + 2 °° positive matrix, the aforementioned ratio. Determining scale factor and -safe quantization scale factor two; the amount of improvement of image quality by second loading 6. As described in item 4 of the scope of patent application, the improved image is a quantization matrix adjustment method, where the encoder is rounded :, And a plurality of corresponding reduced quantization matrices, when the centering is performed, the preset adjustment factor closest to the aforementioned tone is selected: and the two reduced quantization matrices are used as the reduced quantization matrix. Corresponding to 7 · — A kind of quantization matrix adjustment codec for improving image quality encodes a series of images. The adjustment method includes the following steps: · Flat code " has a preset Quantitative Moment 3 has used the bit amount, activity level Act_x, and average quantization scale factor to assign a bit budget value to the current image; calculate the activity level of the current image A ctc; 1 ί Γ: Γ scale factor Calculate estimates based on the amount of bits used, the average scaling factor described above, the top ten U — κ A-π Μ / the degree of tongue activity Act-X and Act-C, and the narration = pre-differentiation Measured scaling factor; ^ = 疋 No enters the adjustment mode. When the aforementioned estimated measurement scaling factor is low: ΤΓ 目 :: factor threshold value, then the coding mode is set to the adjustment mode, otherwise it remains the general mode; When the encoding mode is the adjustment mode, the aforementioned pre-p.21 445 200425744 VI. The quantization matrix set in the scope of the patent application is reduced; and the image encoding is performed. When the encoding mode is equal to the matrix for image encoding, When editing: horse ;: ", according to the reduction, the image is edited according to the preset quantization matrix, and the normal mode is used. 8. The method of adjusting the transformation matrix as described in item 7 of the scope of patent application, where In the calculation of the estimated quality of the shirt image, the above-mentioned estimated measurement scale factor = (the steps of the two case factors used * (activity level Act — C / activity level 杬 7 volume / bit budget value) 9. If the scope of patent application The scale factor of 20g in the 8th project. The adjustment method of the transformation matrix, in which the aforementioned adjustment of the quality of the image quality is an adjustment factor to reduce the preset quantization matrix. In the step of the dagger matrix, the 10 is used as the application The ninth itemization matrix adjustment method of the patent scope, the aforementioned adjustment in #: the quantity adjustment factor for good image quality is the aforementioned ratio in the step of estimating the measurement ratio J Q. The full quantization scale factor 11 for a child and a woman. The method of adjusting the quantization matrix of the 0th item of the patent, including the article described in the previous °. The image quality is 19 j. Vh ^ ^ The quantization scale factor of the queen is 6 〇12. Such as the 10th item of the patent application. 口 于 马 b Quantization matrix tuning This method has already been modified. The quality of the image, and multiple correspondences :: 'ί = Include; When there are several preset adjustment factors, choose the D array that is most connected to the preset adjustment factor. When entering The & small quantization matrix of the adjustment mode is used as the reduced quantization moment ^ R, and corresponds to 1 The activity level of the image in the category I 44fi Page 22 200425744 VI. In the step of applying for the scope of the patent Act_C, for the I image, the difference between each internally encoded macro block is obtained. For P and For the B image, the mode judgment is performed according to the difference between the internal coding and the interrelated coding modes, and the sum of the differences is the current degree of activity Act_C.