TW201138477A - Image decoding method, image encoding method, and devices, programs, and integrated circuits therefor - Google Patents

Abstract

An image decoding method which achieves a reduction in the amount of codes is an image decoding method for decoding an encoded image included in an encoded stream, said image decoding method comprising: analyzing an offset update flag in a header included in the encoded stream (S10); determining whether or not the analyzed offset update flag indicates a predetermined value (S11); calculating a quantization offset matrix used for the encoded image using an offset update parameter included in the header when it is determined that the analyzed offset update flag indicates the predetermined value (S12); and decoding the encoded image by inversely quantizing the encoded image using the calculated quantization offset matrix (S13).

Description

201138477 VI. INSTRUCTIONS: C TECHNICAL FIELD OF THE INVENTION FIELD OF THE INVENTION The present invention relates to a method of decoding an encoded image, and a method of encoding an image, in particular, Decoding and encoding of quantized and quantized parameters. [Prior Art 3] In various image or image encoding methods, quantization is an important step of compressing data by removing several pieces of information in an image or image. In the quantization process, in order to make the image or data more compressed by losing the information, the quantization is usually performed in the frequency-converted region. In most image or image coding methods, the quantization process can be controlled by quantization parameters. At this time, the larger the value of the quantization parameter, the larger the compression amount, and thus the more information is lost. And, vice versa. In order to control the quantization process, among the coding methods of the image or the video, in addition to the quantization parameter, the quantization and inverse quantization process can be controlled by the set of quantization offset values (quantization offset matrix). Here, each of the frequency coefficients (conversion coefficients) in the two-dimensional conversion block can be quantized using both the quantization parameter and the corresponding one of the quantization offset values. An example of the inverse quantization process can be expressed by the following equation.

AbsCoeff[i][j]=((abs(QuantizedCoeff[i][j])«7)-Offset[i][j])*

LevelScale»(QShift+7) Here, [i] and [j] indicate the position of the frequency coefficient of the 2-dimensional conversion block, 201138477

AbsCoeff[i][j] is the absolute value of the frequency coefficient at the position (i, j) and the inverse quantized value. QuantizedCoeff[i][j] represents the quantized value of the frequency coefficient in position (y). Also, 'and' indicates the left and right displacement of the bit, and the * indicates the multiplication. In addition, LevelScale and QShift are controlled by quantization parameters. Offset[i][j] is the quantized offset value for the frequency coefficient in the position (i, j). There are several prior art notes that use quantized offset springs for quantization processing. However, the quantization offset parameters as described in the prior art are used only for quantization processing and not for inverse quantization processing. Such a quantization offset parameter only acts as a rounding calculation in a typical division or right displacement operation. The prior art, which is a quantization offset value transmitted to the decoder by the inverse quantization process, is only a few (see, for example, Non-Patent Document 1). Moreover, the dynamic range of the quantization offset value is usually It is limited to the quantization scale step size. For example, it is briefly explained below. The inverse quantization process is usually proportionally increased or decreased (scaUng), that is, multiplicative sound. The inverse quantization value is usually based on the quantization ratio. The magnitude of the order is proportionally increased or decreased.: The inverse quantized value can be adjusted upward or downward by performing addition or subtraction of the quantized horizontal offset value corresponding to the quantized offset value. The adjustment performed by the offset value is usually limited to one step size. Figure 1 shows the inverse quantized value adjusted by the quantized horizontal offset value. The quantitative quantifier red scale is the eigenvalue (the absolute value of the quantized value) multiplied by the quantization scale order size QSS. In 201138477, the 'inverse quantized value is subtracted (or added) from the inverse quantized value. Move the value and go up or Adjusted below and set as the final inverse quantized value. Prior Technical Documents Non-Patent Literature Non-Patent Document 1:

http://wftp3.itu.int/av-arch/jvt-site/2004_03_Munich/JVT -K026.zip

C ^ '明内 J SUMMARY OF THE INVENTION Problem to be Solved by the Invention The problem of the prior art is that the number of bits required to transmit the quantization offset matrix including the quantization offset matrix in the element to the decoder is very good. When there are many quantization offset matrices transmitted to the decoder, the prior financial system separately transmits the respective offset values, so that it is necessary to have a large number of bits for presenting the matrices transmitted to the decoder. For the purpose of the present invention, the object of the present invention is to provide an image decoding method and an image encoding method for the number of money codes. In order to achieve the above object, a method for decoding an image according to an aspect of the present invention, which is the image-encoding image of the present invention, has the following processing: analysis and analysis of the two The offset update flag in the header; the determined value; whether the offset update flag displays the previously set value, and the aforementioned offset update flag is determined by the display. The offset update parameter contained in the header will be suitable for 201138477

Moreover, in the image decoding method of the present invention, the offset stored in the header or the sequence header can be adjusted by adjusting the inverse quantization value of the range of the inverse quantization value of the offset matrix. More flexible, and the quality of the picture. In addition, the ΜIt θ It匕' improves the main character of the decoded image, and the header includes an update type identifier indicating the update side of the quantization offset matrix. When the quantization offset matrix is calculated, the update type identifier included in the header is parsed, and the densification offset matrix is calculated in accordance with an update method indicated by the updated update type identifier. Thereby, since the quantization offset matrix is extracted by the update method corresponding to the update type identifier, an appropriate update method can be selected in the inverse quantization of the amount 彳匕i for the image coding apparatus, so as to correspond to the The offset update parameter and the update type identifier of the selected update method are included in the header of the encoded stream. Therefore, the update or calculation of the quantization offset matrix can be made more flexible. 6 201138477 Further, when calculating the quantization offset matrix, the offset update parameter is applied to another quantization offset matrix stored in the memory, thereby calculating the quantization offset matrix of the calculation target. Thereby, the smaller the difference between the quantization offset matrix to be calculated and the other quantization offset matrix, the more the number of bits of the offset update parameter can be reduced, and the number of codes of the encoded stream can be further reduced. Further, when calculating the quantization offset matrix, the previously calculated quantization offset matrix is used as the other quantization offset matrix, and the offset update parameter is applied. Thereby, since the quantization offset matrix calculated last has a tendency to strongly calculate the quantization offset matrix of the target, the number of bits of the offset update parameter can be surely reduced. Further, the header includes a matrix identifier for specifying the other quantization offset matrix. When calculating the quantization offset matrix, the matrix identifier included in the header is parsed, and the other quantization offset matrix identified by the parsed matrix identifier is retrieved from the memory. Thereby, in the aspect of the image coding apparatus, another quantization offset matrix having a small difference from the quantization offset matrix of the calculation target can be appropriately selected, and the matrix identification for specifying the other quantization offset matrix selected thereby can be appropriately selected. Symbol, you can include the header. Therefore, the difference between the quantization offset matrix of the calculation target and the other quantization offset matrix can be surely reduced, and the number of codes of the encoded stream can be surely reduced. Further, the header includes the first and second correction parameters and the offset value of 5, and the δ offset value of 201138477 corresponds to each element included in the quantization offset matrix to be calculated. When calculating the quantization offset matrix, the first and second coefficients are calculated based on the first and second correction parameters included in the header; and each of the elements of the other quantization offset matrix is known. The quantization offset value 'multiplies the first coefficient by the known quantized offset value' to calculate a first value, and multiplies the second coefficient by a difference between the known quantized offset value and a fixed value, thereby calculating a second value. And adding the first value to the second value to calculate a predicted quantization offset value, and adding the predicted offset offset value to the 5 offset corresponding to an element of the quantization offset matrix to be calculated. The value is used to calculate the quantized offset value as the aforementioned element. For example, the aforementioned header additionally contains an offset update denominator. When the first coefficient is calculated, the first correction parameter is added to the offset update denominator, and the value obtained by the addition is divided by the offset update denominator to calculate the first coefficient. Further, when calculating the second coefficient, the second correction coefficient is divided by the offset update denominator to calculate the second coefficient. Here, the fixed value is a value set in advance for the quantization offset matrix to be calculated. Further, the fixed value may be included as an element and included in the other quantization offset matrix, and may be a quantization offset value corresponding to the position of the DC component. By using the second value calculated by multiplying the second coefficient by the difference between the known quantized offset value and the fixed value of the other quantization offset matrix, the AC component of the other quantization offset matrix can be expressed and used. The AC component is calculated as an appropriate predicted quantized offset value, and thus the delta offset value can be reduced. Therefore, the number of codes can be further reduced. 201138477 Further, the fixed value is an element and is included in an average value of the complex quantized offset values of the other quantization offset matrices. Thereby, the second value calculated by multiplying the second coefficient by the difference between the known quantized offset value and the fixed value of the other quantization offset matrix can respectively indicate the deviation of each element of the other quantization offset matrix, and The appropriate predicted quantization offset value can be calculated using the deviation, and thus, the offset value can be reduced (5. Therefore, the number of codes can be further reduced. Further, the offset update parameter includes a complex δ offset value, and the complex number 5 is biased The shift value system displays a difference between the complex quantized offset values which are elements of the quantized offset matrix of the object. When calculating the quantized offset matrix, the following processing is repeated: the complex δ is repeated One of the offset values is added to the quantized offset value calculated in the quantized offset matrix, thereby calculating a new quantized offset value. Thereby, among the quantized offset matrices of the calculated object The less the difference between the complex quantized offset values, the more the number of bits of the offset update parameter can be reduced, and the number of codes of the encoded stream can be further reduced. Further, the offset update parameter includes only one As a quantization offset value of a single quantization offset value, the quantization offset value is an element of a quantization offset matrix to be calculated. When calculating the quantization offset matrix, the pair is included in the quantization offset matrix of the calculation target. For all the elements, the aforementioned single quantization offset value included in the offset update parameter is set. Thereby, if the complex quantized offset value in the quantization offset matrix of the calculated object is the same value, only one quantization offset value is used as The offset update parameter is included in the header, so the number of bits of the offset update parameter can be reduced, and the number of codes of the encoded stream can be reduced by step 201138477. However, the present invention can be implemented not only as such an image decoding method. And an image encoding method corresponding to the image decoding method, an image decoding device that decodes an image according to the image decoding method, an image encoding device that encodes an image according to the image encoding method, and the like The integrated circuit corresponding to the device, the program for causing the computer to perform image processing of the methods, and the recording medium on which the program is recorded are implemented. The method can be implemented as a method of writing a quantized offset matrix into a coded stream, a method of analyzing a quantized offset matrix from a coded stream, a device, a program, a recording medium, and an integrated circuit corresponding to the methods. Advantageous Effects of Invention The image decoding method and the image encoding method of the present invention have an effect of reducing the number of codes, that is, the number of bits required for encoding the quantization offset matrix becomes smaller, and inverse quantization in inverse quantization processing The processing of the value becomes more flexible, so that the coding efficiency can be improved. BRIEF DESCRIPTION OF THE DRAWINGS Fig. 1 is a diagram showing the inverse quantized value adjusted by the quantized horizontal offset value. Fig. 2A is a view showing an embodiment of the present invention. 1 is a block diagram showing the configuration of an image decoding apparatus. Fig. 2B is a flowchart showing an image decoding method in the first embodiment of the present invention. Fig. 3A is a diagram showing a first embodiment of the present invention. Fig. 3B is a flow chart showing an image encoding method in the first embodiment of the present invention. Fig. 4 is a block diagram showing an example of a detailed configuration of an image decoding apparatus in the first embodiment of the present invention. Fig. 5 is a block diagram showing an example of a detailed configuration of an image coding apparatus in the first embodiment of the present invention. Fig. 6A is a view showing the construction of a sequence header in the first embodiment of the present invention. Fig. 6B is a view showing the construction of a picture header in the first embodiment of the present invention. Fig. 7 is a view showing the configuration of an offset parameter set in the first embodiment of the present invention. Fig. 8 is a flow chart showing the process of decoding the quantization offset matrix in the first embodiment of the present invention. Fig. 9 is a flow chart showing the process of encoding the quantization offset matrix in the first embodiment of the present invention. Fig. 10 is a flow chart showing the process of calculating a new quantization offset matrix corresponding to one conversion block in the first embodiment of the present invention. Fig. 11 is a flow chart showing the process of calculating and writing an offset parameter corresponding to one conversion block in the first embodiment of the present invention. Fig. 12 is a flowchart showing a process in which the image decoding apparatus analyzes the written offset update parameter by the first update method in the first embodiment of the present invention. Fig. 13 is a flowchart showing a process of calculating a quantization offset value by the first updating method in the image decoding device 201138477 in the first embodiment of the present invention. Fig. 14 is a flowchart showing a process of calculating an offset update parameter (5 offset value) by the image encoding apparatus according to the first embodiment of the present invention. Fig. 15 is a view showing the implementation of the present invention. In the first aspect, the image encoding apparatus writes the offset update parameter into the header by the first update method. Fig. 16 shows a second update of the image decoding apparatus according to the first embodiment of the present invention. The flowchart of the process of calculating the quantized offset value is shown in Fig. 17. Fig. 17 is a flowchart showing the process of writing the 5 offset value into the header by the image encoding apparatus according to the second update method in the first embodiment of the present invention. Fig. 18 is a flowchart showing a process of adjusting the inverse quantization value by the quantization offset value by the inverse quantization unit of the image decoding device and the inverse quantization unit of the image coding device according to the first embodiment of the present invention. Fig. 20 is a view showing a configuration of an offset parameter set in a first modification of the first embodiment of the present invention. Fig. 20 is a view showing a modification 1 corresponding to a conversion block in the first modification of the first embodiment of the present invention. Flowchart for processing the new quantized offset matrix Fig. 21 is a flowchart showing a process of calculating and writing an offset parameter corresponding to one conversion block in the first modification of the first embodiment of the present invention. Fig. 22 is a view showing an embodiment of the present invention. Fig. 23 is a flowchart showing a process of analyzing the offset update parameter by the image decoding device in the second modification of the first embodiment of the present invention. Fig. 24 is a flowchart showing the processing of the specific matrix of the decoding device of Fig. 12 in the second modification of the first embodiment of the present invention. Fig. 25 is a view showing a modification 2 of the first embodiment of the present invention. A flowchart of a process in which an image encoding device writes an offset update parameter into a header. Fig. 26 is a schematic diagram showing an example of the overall configuration of a content supply system that implements a content delivery service. Fig. 28 is a block diagram showing an example of a configuration of a mobile phone. Fig. 29 is a view showing an example of a whole configuration of a digital broadcasting system. Fig. 30 is a block showing a configuration example of a television set. Fig. 31 is a block diagram showing an example of a configuration of an information reproduction recording unit that reads and writes information to and from a recording medium. Fig. 32 is a view showing a configuration example of a recording medium of the optical disk. Fig. 33 is a view showing the realization of each embodiment. (Embodiment 3) Embodiments of the present invention will be described below with reference to the drawings. (Embodiment 1) Fig. 2A is a block diagram showing the configuration of the image decoding device of the embodiment. The image decoding device 1 of the present embodiment is a device for decoding a coded image included in a coded stream, and includes flag analysis. The unit 10 analyzes the offset update flag included in the header of the encoded stream; the determining unit 11, 13 201138477 determines whether the offset update flag analyzed has a value set in advance. When the estimated offset update flag indicates that the previously set value is displayed, the quantization offset matrix calculation unit 12 calculates an offset map parameter included in the header to calculate a code map. And the inverse quantization decoding unit 13 performs inverse quantization on the coded image by using the calculated quantization offset matrix, thereby decoding the coded image. Fig. 2B is a flow chart showing the image decoding method of this embodiment. The image decoding method according to the present embodiment is a method of decoding a coded image included in a coded stream, and has the following steps: parsing an offset update flag included in the coded stream (S10) After that, it is determined whether the analyzed offset update flag has a value set in advance (S11), and when it is determined that the analyzed offset update flag has a value set in advance, it is included in ' The offset update parameter is calculated, and the quantized offset 12) applied to the encoded image is calculated, and the encoded image is decoded by the encoded quantized offset matrix (S13). ❿, in step sn, = is determined that the resolved offset is less (4) phase set first "μ is the existing quantization offset matrix, the material code image performs inverse quantization, the wrong solution is called code II In (S14), in the image decoding device 1 of the present embodiment, the quantization offset matrix is calculated with the offset unnecessary = number, so that the matrix encoding device shifts the matrix itself. For the encoded stream, it is only necessary to quantize the compression of the original shape, so that the compressed quantization offset matrix can be used as the offset update parameter (4) header of the human coded turbulence 14 201138477. Therefore, The number of bits required to quantize the offset matrix can be reduced, and the number of codes of the encoded stream can be reduced. Fig. 3A is a block diagram showing the configuration of the image coding apparatus of the present embodiment. The encoding device 2 is a device that encodes image data to generate a coded stream, and includes a determination unit 20 that determines whether or not the quantization offset matrix should be updated; and the flag writing unit 21 displays the determination. Result offset update flag write code The header of the stream; the offset update parameter calculation unit 22 calculates the offset update parameter based on the updated new quantization offset matrix when it is determined that the update is to be performed; the parameter write unit 23 The offset update parameter is written to the header; and the quantization coding unit 24 performs quantization on the image data by using the new quantization offset matrix, thereby encoding the image data. Fig. 3B is shown in the embodiment. A flowchart of an image encoding method. The image encoding method of the present embodiment is a method of encoding image data to generate a encoded stream, and has the following steps: determining whether the quantization offset matrix should be updated (S20) After that, the offset update flag indicating the result of the determination is written in the header of the encoded stream (S21), and when it is determined in the determination that it should be updated, the offset is calculated based on the updated new quantized offset matrix. After updating the parameter (S22) and writing the offset update parameter to the header (S23), quantization is performed on the image data with its new quantization offset matrix, thereby encoding the image data (S24). If it is determined that it should not be updated When the offset matrix is obtained, the offset update flag indicating the result of the determination is written to the header 15 of the code_stream (201125477) (S25), and the image data (10) is encoded by the existing quantization offset moment. Perform quantization on the image tribute. In this way, in the quantization offset moment (4) of the present embodiment, the new quantization MCU is placed in the encoded stream, = Without the need for compression, the information used to quantize its compressed offset moment is included as an offset update parameter. (4) The transfer matrix is restored to the original shape, which reduces the number of strings needed to quantize the offset matrix. The number of codes of the stream is 70. The number of codes is reduced, and the code 1 is reduced. The second embodiment of the present invention will be described in more detail. Ma Ling's 5 sheep fine image decoding device 1000 is provided with: pseudo offset value calculation, > number analysis unit 1500, quantization 15.6, two: quantized offset value storage unit 15〇4, said The body part 508 1508, the inverse conversion unit 151 sample prediction unit 1514, and the 帛 sample weight (4) (5) 2, the analysis unit 15 〇〇, and the denier. Heart ^ 516. Further, the constituent element group composed of the CT 506 is obtained by the offset parameter, and the constituent element group inverse inverse conversion unit 1510 and the sample composed of the flag analysis unit 1 and the matrix calculation unit 12 are not included in the map. In the recombination and refinement unit 150, the sample prediction unit (5) 4 and the second inverse quantization decoding unit of the memory unit (5) 6 constitute an element group, and the corresponding multi-analysis unit 15 is shown in the first embodiment. Streaming header 16 201138477

Dl5〇l (phase mark paste 5〇la or ^header〇1亀), and the offset parameter footer contained in D1501 is parsed (decoded), and the ^offset parameter D15G3 is output to the quantized offset value. The calculation unit 15Q2. The 曰 quantized offset value calculation unit (10) measures the offset parameter D, and the offset value D15G5. At this time, the quantization offset value calculation unit 1502 calculates the new quantization offset value D1505 as necessary after the calculated calculation of the first memory unit 1506. In addition, the ^it offset value it 匕 juice 'I ̄ 偏移 俏 150 1502 will output its calculated new quantized offset value 至 to the quantized shift value storage correction 04. Further, the amount (4) shift (10) memory portion offset value is stored in the first memory portion 1506. D15r: r°8 incorporates the new quantized offset value stored in the code block D1511'i body 1506 of the picture containing the encoded stream, and performs inverse quantization. Further, the inverse quantization unit 15Q quantizes the generated inverse quantized value D1515, and the value, ., 5, 曰 is reversed by the inverse component (10). Reverse conversion inside. The inverse conversion side will be outputted to the sample recombination section (10) by its complex image; into = decoding residual such as η. The block is taken as the sample weight (4) (5) 2 and its decoding ❹m5i7 measuring unit 1514 obtains the prediction sample D1521; ^, i predicts that she 521 adds the decoding residual as 17 and borrows the 1512 system to output D(10). Reorganization _519 is stored, :cheng =:: 1516. In the sample prediction unit 2nd memory unit 1516: 2' is referred to as the reference figure 523, and the image stored in the temple is generated to generate the predicted sample m52 of the above-mentioned 17 201138477 block corresponding to the (4) object and output. To the sample recombination unit 1512. Fig. 5 is a block diagram showing an example of a detailed configuration of an image coding apparatus according to the present embodiment. The image and the code name are set to 2: an offset parameter calculation unit 丄_, an offset parameter writing unit 16G2, a first recording unit, a subtraction unit 16G6, a conversion unit 1608, a 161 unit, and a counter The quantization unit (6), the inverse conversion unit (6)*, the second memory unit 1618, the addition unit 162G, and the sample prediction unit i6i6 are composed of the offset parameter calculation unit 1_, the offset parameter write unit, and the first memory. The constituents of the body 1604 are configured to correspond to the shape constituent elements of the determination unit 20, the flag write unit 21, the offset update parameter calculation unit 22, and the parameter write unit 23 indicated by the collapse. Further, the constituent elements of the Φ subtraction unit 16G6, the conversion unit 1608, the quantization unit 1610, the inverse quantization unit 16u, the inverse conversion unit 1614, the second memory unit 1618, the addition unit 1620, and the sample prediction unit 1616 are equivalent. The quantization coding unit 24 shown in Fig. 3A. The offset parameter calculation unit 1600 acquires the new quantized offset value D1505, and acquires the old quantized offset value D1509 from the first memory unit 1604. The offset parameter calculation unit 1600 calculates the offset parameter D1503 based on the quantized offset values D1505 and D1509, and outputs the offset parameter D1503 to the offset parameter writing unit 602. The offset parameter writing unit 1602 writes the calculated offset parameter D1503 to the header D1501 (picture header D1501b or sequence header D1501a), and outputs its header D1501. The subtraction unit 1606 incorporates the uncompressed block D1609 and the predicted sample D1615 of the picture. The uncompressed block D1609 is a block (b 1 〇 c k ) of the encoding object formed by the complex pixels included in the picture, and the predicted sample D1615 is a block displaying the predicted picture corresponding to the block of the encoded object. The subtraction unit 1606 subtracts the prediction sample D1615 from its non-compressed 18 201138477 constricted block D1609, thereby generating a residual block D1611 and outputting it. The conversion unit 1608 generates the coefficient block D1613' formed by the complex frequency coefficient (conversion coefficient) by performing orthogonal conversion on the residual block D1611 and outputs the coefficient block D1613 to the quantization unit 1610. The quantization unit 1610 obtains the coefficient block D1613 and reads the new quantization offset value D1607 from the first memory unit 1604. Further, the quantization unit 1610 generates a coded block D1621 of the picture by performing quantization using the new quantized offset value D1607 on the coefficient block D1613 and outputs it. The inverse quantization unit 1612 acquires the coded block D1621 from the quantization unit 1610, and reads the new quantization offset value D1607 from the first memory unit 1604. Further, the inverse quantization unit 1612 generates inverse quantization value 〇 1619 by performing inverse quantization using the new quantization offset value D1607 on the coding block D1621, and rotates to the inverse conversion unit 1614. The inverse transform unit 1614 obtains the inverse quantized value D1619, performs inverse orthogonal transform on the block formed by the complex inverse quantized value D1619, and converts the block into a recombined residual block 〇1627 formed by the complex pixels. And output it. The addition unit 1620 obtains its recombination residual block D 丨 62 7 and derives a prediction sample D1615 corresponding to its recombination residual block D1627 from the sample prediction 4 1616. Further, the addition unit 162 calls the recombination block which generates the picture by adding the prediction sample m6l5 to the reassembly residual block D1627' and outputs it. The recombination fine 625 is stored in the second memory portion 1618. The sample prediction unit _6_ multiplexes the image stored in the second memory portion as a reference image (10) and refers to 'generates the prediction sample D1615 corresponding to the above-mentioned block of the encoding target and outputs it. Fig. 6A is a diagram showing the sequence header of the embodiment shown in Fig. 19 201138477. The sequence header D1501a includes an update offset parameter flag 〇3〇2, an offset update denominator D304, and an offset parameter set D306. From the front side, the order is: update offset parameter flag D302, offset update denominator D304, and offset parameter set D306. The offset parameter set D306 is used to calculate the data of the customizable quantization offset matrix corresponding to each transform block. However, the conversion block performs blocks of orthogonal transform, inverse orthogonal transform, and quantization or inverse quantization units. The update offset parameter flag D302 is used to transmit the flag of the offset parameter set 〇6, which is encoded. When the update offset parameter flag D] 〇 2 shows 1 'the offset parameter set D 306 is configured after updating the offset parameter flag D3 〇 2 . However, when the update offset parameter flag D302 displays 〇, the non-customizable quantization offset matrix corresponding to each transform block size is configured after the update offset parameter flag D302. Or, when the update offset parameter flag 〇3〇2 is displayed ’, the above-mentioned non-customizable quantization offset matrix and the offset parameter set D3〇6 are not configured. The offset update denominator D304 is used to determine the denominator of the rate of change when the present invention updates the quantized offset value (quantization matrix). The details of the offset update denominator D304 will be described later. Fig. 6B is a view showing the configuration of the picture header d 15 in the present embodiment. The picture header D1501b contains: an update offset parameter flag D3〇8' offset update denominator D310, and an offset parameter set D312. From the front side, the offset parameter flag D308, the offset update denominator D31〇, and the offset parameter 20 201138477 set D312 are updated. The update offset parameter flag D308, the offset update denominator D310, and the offset parameter set D 312 are the same as the update offset parameter flag D302, the offset update denominator D304, and the offset parameter shown in FIG. 6A. Set D306. Further, in the present embodiment, at least one of the sequence header D1501a and the picture header D1501b may be constructed as described above (as shown in Figs. 6A and 6B). Hereinafter, when the header is merely described as a header, the header indicates the sequence header D1501a or the picture header D1501b. Fig. 7 is a view showing the configuration of the offset parameter sets D306 and D3 12. The offset parameter sets D306 and D312 are composed of a complex offset parameter D1503. The offset parameter D1503 is used to calculate a customizable quantization offset matrix corresponding to the conversion block. When there are complex transform blocks of different sizes in the sequence or picture, the offset parameters D1503 are included in the offset parameter sets 〇 306 and 1) 312 according to the respective sizes. For example, as shown in Fig. 7, the offset parameter sets D3〇6 and D312 include one offset parameter D1503 corresponding to the size of the [~Mth transform block]. The offset parameter D1503 includes an offset update flag D1〇〇 and an offset update parameter D1〇2 for calculating the quantization offset matrix. The offset update flag Di is used to indicate whether the flag of the quantization offset value of the conversion block corresponding to its offset parameter D15G3 should be updated. That is, the offset update flag (1) (10) indicates whether or not the new degenerate offset matrix should be calculated, and thereby the flag of each quantized offset value included in the quantized offset matrix before the update is updated. For example, when the offset update flag D100 is 1, its offset update flag D100 is displayed to be updated, and when the offset update flag D100 is ,, its offset update flag surface is not 21 201138477 offset The update parameter D102 is arranged next to the offset update flag D100, and has a different configuration depending on the update method. In the present embodiment, there are three update methods (first to third update methods), and each update method is assigned an update type identifier for identifying the update method. In the first update method, the 'offset update parameter D102 includes: an update type identifier D104 for displaying 〇, a first and second correction parameters D106 and D108, a number of offset values D110, and a plurality of consecutive (5 offsets). The value D112. The first and second correction parameters D106 and D108 are respectively used to calculate a value used for the quantization offset value. The 5 offset value is added to the values calculated by the first and second correction parameters D106 and D108. In this way, by adding 5 offset values, the quantization offset value corresponding to one conversion coefficient (frequency coefficient) can be calculated. Further, if the number of offset values D110 is Ν', there is continuous N. The 5 offset value D112 is stored in the header D1501. In the second update method, the offset update parameter 〇1〇2 includes: displaying the updated type identifier D104 of the UI, according to the number of conversion coefficients included in the conversion block. The continuous 5 offset value D114. In the second update method, when calculating the quantization offset value corresponding to the predetermined conversion coefficient, the 5 offset value D114 is added to the scan order and the previous conversion coefficient. Quantizing the offset value, thereby calculating a corresponding to its predetermined conversion coefficient In the third update method, the offset update parameter 〇1〇2 includes: an update type identifier D104 of display 2, and a single quantization offset value DU6. In the third update method, 'and a conversion block' The quantization offset value corresponding to all the conversion coefficients is equal to the single quantization offset value D116. However, in each of the above updating methods, the data contained in the header D1501 is all 22 201138477 Department = variable length coding, dable Cut h eGding) and so on. *1 is a flowchart of the process of decoding the quantization offset matrix of the present embodiment. The soft center Φ offset parameter analysis unit 15GG analyzes the update header of the flag, and the step S4GQ)°, in the present embodiment, when the analysis is included in the standard shift table 1 Decoding and other decoding. Next, the offset analyzing unit 15 G G determines whether or not the update offset parameter flag is 丨 (step S402). / μ_, the official offset parameter analysis unit 1500 analyzes the offset update denominator within the target when it is determined to be 1 (step S402) (step S4〇4). On the other hand, if the multi-update denominator is a value that is not included in the header and is previously discriminated, the processing of step S404 will be skipped. Next, the offset parameter analysis unit 1500 analyzes the offset parameter set of the header (vS4〇6)', and calculates the new quantization offset matrix by the quantized offset value calculation unit (step S408). Fig. 9 is a flow chart showing the process of encoding the quantization offset matrix of the present embodiment. The offset parameter calculation unit 16 determines to update the offset parameter flag, and the offset parameter write unit 16G2 writes the update offset parameter flag to the header (step S500). Further, in the present embodiment, when data is written in the header, encoding of variable length coding or the like can be performed. Next, the offset parameter calculation unit 16 determines whether or not the update offset parameter flag is 1 (step S502). That is, the offset parameter calculation unit 16 determines whether or not the quantization offset matrix is calculated to update each quantization offset value. Here, the offset parameter calculation unit 1600 determines that the update offset parameter flag is 丨 (step § 5〇2 23 201138477)

Yes) ’ determines the offset update denominator, and the offset parameter writing unit 1602 writes its offset update denominator to the header (step S504). However, the offset update denominator may be a value that is not previously set in the header and is set in advance. At this time, the processing of step S5〇4 is skipped. Next, the offset parameter calculation unit 1600 calculates an offset parameter set based on the third to third update methods (step S508), and the offset parameter write unit 1602 writes the offset parameter set to the header (step S510). . Fig. 10 is a flow chart showing the processing of calculating a new quantization offset matrix corresponding to one conversion block in the present embodiment. On the other hand, the first drawing shows the detailed processing of steps S406 and S410 of Fig. 8. First, the offset parameter analysis unit 1500 analyzes the offset update flag in the header (step S600). The offset parameter analysis unit 15 determines whether or not the offset update flag is 1 (step S601). Here, if the offset parameter analysis unit 15 determines that the offset update flag is 丨 (Yes in step S601), the updated type identifier in the header is parsed (step S602). Next, the offset parameter analysis unit 1500 determines whether or not the update type identifier is displayed (step S6〇4). Here, if the offset parameter analysis unit _determines the update type identifier display 〇 (Yes in step S604), the offset update parameter of the update type metric is immediately parsed in the header (step s (4). The quantization offset value calculation unit 15_ calculates a new quantization offset value (quantization offset matrix) from the offset update parameter by the βth new method (step S6i). Further, the offset parameter analysis unit 1500 determines the step. (10) The medium update type identifier is not displayed 〇 (No in step S604), and the other is determined to be (4) Qiu Bu job 6). Here, if the parameter analysis unit 靡24 201138477 determines to update the type identifier display 1 (Yes in step S606), the offset update parameter in the header immediately following the update type identifier is analyzed, and the offset value is quantized. The calculation unit 1502 calculates a new quantization offset value (quantization offset matrix) from the offset update parameter by the second update method (step S616). Further, the offset parameter analysis unit 1500 determines that the type identifier is updated in step S606. If 1 is not displayed (No in step S606), the offset update parameter of the update type identifier in the header is analyzed by a single quantization offset value (step S612). The quantization offset value calculation unit 15〇2 calculates a new quantization offset value (quantization offset matrix) from the single quantization offset value by the third update method (step S614). In other words, the quantization offset value calculation unit 15〇2 sets the quantization offset value corresponding to all the conversion coefficients (the positions of the frequency components) in the conversion block to the single quantization offset value that has been analyzed. Fig. 11 is a flow chart showing the process of calculating and writing an offset parameter corresponding to one conversion block in the present embodiment. The figure u shows the detailed processing of steps S5〇8 and S510 of Fig. 9. First, the offset parameter calculation unit 16 determines the offset update flag, and the offset parameter write unit 16G2 writes the offset update flag determined by it to the header (step (10)). Next, the offset parameter calculation unit 1600 determines whether or not the new quantization offset matrix is used based on the determined offset update flag ((4)). Here, the offset parameter calculation unit 16_defines the quantization offset matrix (Yes in step S701), that is, the shirt update flag, and the offset parameter writing unit (10) 2 writes the update_identifier determined by the offset parameter writing unit (10)2. The human header (step π is called. Next, the offset parameter calculation unit 16 determines whether the quantization offset matrix should be encoded by the third update method according to the determined update type identifier ' (step 25 201138477 to step S704). The offset parameter calculation unit 1 deletes whether or not the quantization offset matrix should be compiled by the first update method (addition of step 4), and calculates the offset update parameter according to the new quantization offset matrix m (step s7).偏移8) The offset parameter writing unit 16 〇2 writes the human error header as the shift update parameter by the first update method (step s7i). The right offset parameter calculation unit 16 In step S704, it is determined whether the quantization is that the shift matrix should not be encoded by the first update method (N of step S704), and whether the quantization matrix values of the new quantization offset matrix are all the same (step S7G6) Here, if the offset parameter calculation unit 丨_ determines that they are not all the same ( Step S706.) That is, the offset update parameter is calculated by the second update side based on the new quantization offset matrix. Further, the offset parameter writing unit 16〇2 calculates the calculated offset by the second update method. The update parameter is written to the header (step S7116). On the other hand, if the offset parameter calculation unit 1600 determines that all are the same in step S706 (Yes in step S706), that is, according to the new quantization offset matrix, the single quantization offset is determined. Further, the offset parameter writing unit 16〇2 writes the determined single-quantization offset value as the offset update parameter into the header by the third updating method (step S712). The image decoding apparatus 1 of the present embodiment analyzes the processing of the offset update parameter written by the first update method. The 12th figure shows the step 86 of FIG. First, the offset parameter analysis unit 1500 analyzes the first correction parameter included in the offset update parameter of the header (step S800), and analyzes the second correction parameter (step S802). The parameter analysis unit 15 analyzes the offset update parameter of the header. It has the number of offset values (step S8〇6). Moreover, 26 201138477 offset, number analysis unit! initializes the count value to q (step (10) 8), and '', there is a satisfaction count value smaller than the step Na The condition of the number of offset values analyzed is repeated, and the processing of the loop A is repeated. The offset parameter analysis unit 15_ parses out from the complex W offset value in the offset update parameter stored in the header. The offset value selected according to the scanning material (step S8H)). Further, the 'offset parameter analysis unit' counts the count value (i_nent) (step S812). That is, the offset parameter analysis unit measures: repeats the processing of steps S8 and 8812 until the number of the 5 offset values parsed by the step coffee clock is "at the count value. The complex number stored in the offset update parameter The continuous 5 offset values are arranged in a scan order in a 2-dimensional matrix corresponding to the transform block. For example, the 5 offset values are in a zigzagsening order in a 2-dimensional matrix corresponding to the transform block. From the low frequency component to the high frequency component. ❿ 'Depending on the size of the conversion block, the scanning order will be different. For example, the scanning order of the block (8 X2 block) formed by the frequency component (conversion coefficient) of 8X2, It is different from the scanning order of the block (2x8 block) formed by the frequency component of 2χ8. Fig. 13 is a view showing the process of calculating the quantized offset value by the first updating method in the image decoding apparatus of the embodiment. In the above, the detailed processing of step 8610 of Fig. 10 is displayed. First, the quantization offset value calculation unit (10) obtains the quantization offset value before the update (step_). That is, the quantization offset value is calculated (4) 2 Read from the first memory unit 1506 (example) The previous one is the quantization offset value at the same position as the quantization offset value of the calculation target among the quantization offsets stored in the memory unit. 27 201138477 Next, the quantization offset value calculation unit 1502 is first. The correction parameter D106 determines the first coefficient (Factor1), and increases or decreases the quantization offset value before the update by the first coefficient to calculate the first value (step S1202). The first coefficient (Factorl) is as follows. (Formula 1): In (Formula 1), "First Modification Parameter" displays the first correction parameter, and "update Denominator" displays the offset update denominator.

Factorl=(First Modification Parameter+Update Denominator)/ Update Denominator (Formula 1) and 'As shown in Figures 6A and 6B, the offset update denominator may be included in the header' or the offset update denominator may not be included. The settings have pre-defined values. For example, the value defined in advance is 16 or 128. The first value (First Value) is calculated by the following (Formula 2). In (Formula 2), "〇 id Offset Value" displays the quantization offset value before the update.

First Value = 01d Offset ValuexFactorl (Expression 2) Next, the quantized offset value calculation unit 1502 calculates a difference between the pre-updated quantization offset value (Old Offset Value) and the fixed value (Fixed Value) (step S1204). The difference system is displayed by "Old Offset Value-Fixed Value". For example, a fixed value is 16', that is, a DC offset value or an average offset value of a block. The DC offset value is the maximum value of the DC component of the quantized offset matrix before the update, and the average offset value is the average of all the quantized offset values in the quantized offset matrix before the update. . Further, the 'quantization offset value calculation unit 15〇2 calculates the second coefficient from the second correction parameter D108 shown in FIG. 7 and increases or decreases the calculated difference by the second coefficient, thereby calculating the second coefficient. Value (step S1206). 28 201138477 The second coefficient is calculated by the following (Formula 3). In (Formula 3), the "Second Modification Parameter" displays the second correction parameter.

Factor 2 = Second Modification Parameter / Update Denominator (Form 3) The second value (Second Value) is calculated by the following (Formula 4).

Second Value=(01d Offset Value-Fixed Value) xFactor2 (Expression 4) Next, the quantization offset value calculation unit 1502 adds the first value and the second value, thereby calculating the predicted quantization offset value (calculated value) ( Step S1208). Further, the quantization offset value calculation unit 1502 adds the predicted offset offset value to the 5 offset value (Delta Value) corresponding to the quantization offset value of the calculation target, that is, "Fist Value + Second Value + Delta Value" calculates a new quantized offset value to be calculated (step S1210). The equations of the steps S1202 to S1210 are shown by the following (Formula 5). In the case of 'Expression 5', "New Offset Value" displays the new quantized offset value of the calculation target.

New Offset Value = (01d Offset ValuexFactorl)+((〇ld Offset Value-Fixed Value)xFactor2)+Delta Value (Expression 5) Further, in the present embodiment, steps S1200 to S1210 in accordance with FIG. 13 are employed. The sequence shown 'executes the operation shown by (Equation 5), but its operation can also be performed in other orders. Fig. 14 is a flow chart showing the process of calculating the offset value of the offset update parameter by the image encoding device 2 of the present embodiment by the first update method 29 201138477. However, this Fig. 4 shows the detailed processing of step S708 of Fig. 11. First, the offset parameter calculation unit 1600 acquires the quantization offset value before the update (step S1300). In other words, the 'offset parameter calculation unit 1600 reads out from the first memory unit 1604' that the previous (for example, the previous) singed offset matrix stored in the first memory unit 16〇4 is out of the object (5 offset offset value is located at the same position as the quantized offset value. Next, the offset parameter calculation unit 1600 determines the first coefficient (Factor1)' with the first correction parameter d 1 〇6 and updates the first coefficient by the first coefficient The first offset value is calculated by increasing or decreasing the quantized offset value (step S1302). The first coefficient (Factorl) is calculated by the following equation (Expression 6). In (Expression 6), the "First Modification Parameter" is displayed. 1 Correct the parameter, and rUpdate Denominator shows the offset update denominator.

Factorl=(First Modification Parameter+Update Denominator)/

Update Denominator ·.. (Formula 6) and 'As shown in Figures 6A and 6B' may include an offset update denominator or a pre-defined value without including an offset update denominator. For example, the value defined in advance is 16 or 128. The first value (First Value) is calculated by the following (Formula 7). In (Equation 7), "Old Offset Value" displays the quantization offset value before the update.

First Value=01d Offset ValuexFactorl (Expression 7) Next, the offset parameter calculation unit 1600 calculates a difference between the quantization offset value (Old Offset Value) and the fixed value (Fixed Value) before the update (step S1304). The difference is displayed as "Old Offset Value-Fixed Value". For example, the fixed value is 16, which is the DC offset value or the average offset value of the block. In addition, the offset parameter calculation unit 1600 calculates the second coefficient ' by the second correction parameter D108 shown in FIG. 7 and increases or decreases the calculated difference by the second coefficient, thereby calculating the second. Value (step S1306). The second coefficient is calculated by the following (Equation 8). In (Equation 8), the "Second Modification Parameter" displays the second correction parameter.

Factor2=Second Modification Parameter/Update Denominator (Formula 8) The second value (Second Value) is calculated by the following (Formula 9).

Second Value=(01d Offset Value-Fixed Value) xFactor2 (Expression 9) Next, the offset parameter calculation unit 1600 adds the first value and the second value to calculate a predicted quantization offset value (calculated value) (step S1308). In addition, the offset parameter nose portion 1600 subtracts the predicted quantization offset value from the updated New Offset Value, that is, "New Offset Value-(Fist Value+Second Value)" The calculation (5 offset value is calculated (step S1310). The equations of the steps S1302 to S1310 are displayed as follows (Formula 1).

Delta Value = New Offset Value-(01d Offset ValuexFactorl)+((〇ld Offset

Value-Fixed Value) xFactor2) (Expression 10) In the present embodiment, the calculation shown in (Formula 10) is executed in the order shown in steps S1300 to S1310 of FIG. But you can also perform its operations in other orders. Fig. 15 is a flow chart showing the process of writing the offset update parameter to the header in the image encoding apparatus 2 of the present embodiment by the 31 201138477 1 update method. However, the fifteenth figure shows the detailed processing of step S710 of Fig. 11. First, the offset parameter writing unit 1602 writes the first correction parameter used by the offset parameter calculation unit 1600 into the header (step S900), and then the second correction used by the offset parameter calculation unit 1600. The parameter is written to the header (step S902). Further, the offset parameter writing unit 1602 calculates the number of <5 offset values to be written (step S904). For example, the calculation (the calculation of the number of 5 offset values is performed by finding the last non-zero (the position of the 5 offset values) in the scanning order in the 2-dimensional matrix. However, the 2-dimensional matrix corresponds to the quantization. The matrix of the offset matrix is a matrix including d offset values in the element. Next, the offset parameter writing unit 1602 writes the number of the 5 offset values thus calculated into the header (step S906). The shift parameter writing unit 1602 initializes the count value to 〇 (step S908), and repeats the processing of the loop B as long as there is a condition that the count value is smaller than the number of the offset values written in step S906. In B, the offset parameter writing unit 1602 writes a &lt;5 offset value into the header in accordance with the complex number (5 offset value) calculated by the offset parameter calculating unit 1600 (step S910). The offset parameter writing unit 1602 increments the count value (step S912). That is, the offset parameter writing unit 1602 repeats the processing of steps S910 and S912 until it is written in step S906 (5 offset) The number of values is equal to the count value. In other words, in steps S910 and S912, the partial The shift parameter writing section 1602 successively writes (5 offset values into the header until the position of the 5 offset value written in step S910 is equal to the last non-zero in the scanning order in the above-described 2-dimensional matrix ( 5 The position of the offset value. 32 201138477 Fig. 16 is a flow chart of the processing of the image decoding device _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 6IS1 shows the detailed processing of the step _16. The figure 16 is initialized to the value defined by the pre-calculation unit 15°2 at the beginning of the initial quantization offset value (step S1GGG). For example, the pre-shifted selection The shifting parameter is 15 嶋 quantized biasing will repeat the processing of the loop c. The grading offset value dissimilaring part 1502 is shifted more = the middle 'offset parameter analyzing part 15 ° ° will be from the header The new complex continuous shift value is analyzed by shifting the value of one of the scan handles (step_4). The quantized offset value calculation unit (10) adds the initial offset value of the 5 offset value to the initial stage. The value is used to calculate a new quantized offset value (step S1_). Further, the quantized offset value calculating unit 15〇2 will calculate the initial money (4) Sl_hat The new quantized offset value (step S_). The updated initial value is used as the value added to the next (5 offset value) in the scan order. However, the complex number stored in the offset update parameter ^ The offset value is formed as a two-dimensional matrix by the complex conversion coefficient (frequency component) corresponding to the conversion block, as in the calculation of the complex new quantized offset value, and the respective values are from the low frequency component to the high frequency component. The scanning order is arranged. That is, steps S1 〇〇4, S1006, and S1008 are repeated until the resolution is resolved (the position of the 5 offset value reaches the predetermined scanning position within the 2-dimensional matrix corresponding to the conversion block. In other words, 'Steps S1004, S1006, and S1008 will be repeated until the position of the quantized offset value of the calculated object reaches the quantized offset moment 33 corresponding to the conversion block. The position of the offset value), or until the time point when the quantized offset value calculated in step s" 006 is zero. When the quantized offset value calculated by Litian is zero, the quantized offset value calculation IU502 sets the residual quantized offset value from the position of the quantized offset value of zero to the last scan position 2 to be set to - The calculated non-zeroized offset value (step_2). ❿, the above scanning sequence is like a zigzag scanning sequence. Fig. 17 is a flow chart showing the process of writing the offset value of the 3 offset value by the image encoding device 2 of the present embodiment by the second updating method. However, the seventeenth figure shows the detailed processing of step S716 of Fig. 11. First, the offset parameter calculation unit 16 initializes the first reference offset value to a previously defined value (step s i i(9)), and initializes the count value to 〇 (step S1102). For example, the value defined in advance is 8 or 16. Further, only the first condition 'offset parameter calculation unit 1600 and the offset parameter writing unit i 6〇2 satisfying the number of the iUi small conversion block _(four) remaining (fresh component) are repeatedly executed. On the other hand, if the second condition is satisfied by the processing of step S111G described later, the processing of repeating the loop D is stopped. In the path, first, the offset parameter calculation unit 16 calculates a 5-offset value by subtracting the reference offset value from the new quantized offset value in the scanning order (step SU〇4). Next, the 'offset parameter writing portion 16G2 writes the calculated 5-offset value into the header (step S11G6). Then, the offset parameter calculation unit 16〇〇 updates the reference offset value to the new quantization offset value used in step SUG4 (step S1). The reference offset value thus updated is in a later step si104, 34 201138477 It is subtracted from the next new quantized offset value in the scan order. Next, the offset parameter calculation unit _ 敎 within the new quantization offset 'all remaining quantization offset values (in the scan order, the new quantization offset value used in the previous step S1104 is located further rearward) The scalp shift value is the same value (step UG). Here, if the offset parameter calculation unit _ is determined to be the same (Yes in step S111), each of the five offset values corresponding to all of the remaining new offset values is calculated as a negative reference offset value. At this time, the above second condition is satisfied, and the repetition of the processing of the loop D is stopped. Moreover, the offset parameter "part 16G2 will bias the negative reference offset value by 5 _ ^ header (step 2). On the other hand, if the offset parameter calculation unit _^ step SU1G sneaked the converted offset When the value is different from the reference offset value (No in step S1110), the count value is incremented (step s(1) sentence. Thus, in loop 0, the processes of steps S1104, S1106, S11〇8, Du and Sl114 are processed. Only in the case of the first condition described above and not satisfying l, (4) complex execution is performed. However, the upper zigzag scanning sequence is shown in Fig. 18, which is shown in the inverse of the image decoding split 1000 of the present embodiment. A flowchart of the process of adjusting the inverse quantized value by the quantized offset value by the inverse quantization unit (6) 2 of the 1U5G8 and the image grading device 2 ri ri, the &amp; quantization unit obtains the quantization from the first memory unit Offset value (step 8) Here, the inverse grading unit takes the quantized offset value as a quantized offset score: the processing. Then the 'inverse quantization unit subtracts the quantized offset from the previously meaningful value to calculate the scale. Value (step S14〇2). Moreover, the inverse quantization unit multiplies the scale value by the quantization scale step size coffee ( Step si4, and then inverse quantization 35 201138477 The division of the product of the multiplication result obtained in step S1404 by the same value as the previously defined value used in step si4〇2, thereby calculating the quantization horizontal offset value (step S1406) Finally, the inverse quantization unit subtracts the quantization horizontal offset value from the inverse quantized value (step S1408). The processing shown in Fig. 14 is shown by the following (formula u).

NewAbsVal=AbsVal-(PredefinedVal-〇ffset Numerator) Quant Quantization Step Size/PredefinedVal (Expression 11) Here, 'NewAbsVal' is the absolute value of the adjusted new inverse quantized value, and “PredefinedVal” is a previously defined value. And r 〇ffset Numerator" is a quantized offset molecule. Also, "Quantizati〇nStepSize" is a quantized scale step size, and the previously defined value is such as 128. In this way, in the embodiment, the quantization offset matrix is calculated by using the offset update parameter, so that it is not necessary to put the quantization offset matrix itself into the encoded stream in terms of the image coding device, and only the quantization offset needs to be compressed. The matrix, and the information used to restore the compressed quantized offset matrix to its original state is included as an offset update parameter in the header of the encoded_stream. Therefore, the number of bits required to quantize the offset matrix can be reduced, and the number of codes of the encoded stream can be reduced. Further, in the present embodiment, the quantization offset matrix is calculated by the update method corresponding to the update type identifier. Therefore, in the image coding apparatus, an appropriate update method can be selected in the quantization or inverse quantization, and The offset update parameter and the update type identifier corresponding to the selected update method may be included in the header of the encoded stream. Therefore, the update or calculation of the quantization offset matrix can be made more flexible. (Modification 1) 36 201138477 Here, an i-th modification of the embodiment will be described. The complex offset parameters stored in the header of the above-described embodiment are respectively constituted by the any-term updating method in the third to third updating methods, and can be constituted by mutually different updating methods. However, the complex offset parameters of the present modification are all constituted by the third update method. Fig. I9 is a view showing the constitution of the offset parameter set (1) of the present modification. The offset parameter set Dl7_ is equivalent to the offset parameter set shown in the second and second maps, and the complex offset parameter is deleted by the complex offset parameter. When there are complex conversion blocks of different sizes in the sequence or picture, the offset parameter 〇17〇2 is included in the offset parameter set D1700 according to the size. For example, as shown in Fig. 19, the offset parameter set D1 includes #Μ offset parameters D1702, which correspond to the sizes of the first to third conversion blocks, respectively. The offset parameter D1702 includes an offset update flag D17〇1 and an offset update parameter 1) 1703 for calculating a quantization offset matrix. The offset update flag D17〇1 is the same as the offset update flag 0100 shown in Fig. 7, and is used to indicate whether the flag of the quantization offset value of the conversion block should be updated. That is, the offset update flag D1701 indicates whether or not the quantization offset matrix should be calculated to update the respective quantization offset values of the quantization offset matrix before the update. The offset update parameter D1703 is configured immediately after the offset update flag D1701. Further, the offset update parameter D1703 is different from the configuration of the offset update parameter D102 corresponding to the first update method shown in Fig. 7, and does not include the update type identifier D104. That is, the offset update parameter D1703 includes: first and second correction parameters D1704, and D1706, (5 number of offset values D1708, 37 201138477, and N consecutive 5 offset values Dni〇. The image decoding apparatus 1000 and the image encoding apparatus 2000 are the same as the above-described embodiment, and the processing shown in Figs. 8 and 9 is executed. 2 = Displaying the new array corresponding to one conversion block in the present modification is calculated. Flowchart H of the process of processing H shows the detailed processing of step S4〇6 and s marriage of Fig. 8. 'First,' the offset parameter analysis unit 15 analyzes the offset update flag in the header. (Step S19G()) The offset parameter analysis unit 1 determines whether or not the offset update flag is 1 (step S19G2). Here, if the offset parameter analysis object determines the offset update flag is 1 (step _Yes), that is, the offset update parameter in the header is parsed without parsing the update type identifier (the step is profitable. Moreover, the quantized shift value calculation unit 15 calculates the new quantized offset value by the offset update parameter thereof. (Quantization offset matrix) (Step 6). At this time, the quantization offset value is calculated. [U5G2 or higher The i-th update method of the embodiment calculates a quantization offset matrix. Fig. 21 is a flowchart showing a process of calculating and writing an offset parameter corresponding to one conversion block in the present modification. The detailed processing of steps S508 and S510 of Fig. 9 is displayed. First, the 'offset parameter calculation unit deletes the offset update flag, and the offset parameter write unit 1602 writes the offset update flag determined by the offset parameter write unit 1602. Header (step S2GGG). Then the 'offset parameter calculation is based on the offset update flag determined by the decision, and the judgment is money (four) quantization offset matrix (step s2〇〇2). Here, if the offset parameter is calculated It is determined that the quantization offset matrix is (Yes in step S2002) 'that is, the decision and writing of the update type identifier are not executed, 38 201138477, and the offset update parameter is calculated based on the new quantization offset matrix (step S2004). The shift parameter writing unit 16〇2 writes the calculated offset update parameter into the header (step S2006), and in steps S2004 and S2006, the offset parameter calculating unit 1600 and the offset parameter writing unit 1602 are the same. In the above embodiment, the first one is The new method calculates the offset update parameter and writes it. (Modification 2) Here, a second modification of the embodiment will be described. The complex offset parameters stored in the header of the above embodiment are respectively the first The updating method configuration of any one of the third updating methods can be configured by mutually different updating methods. However, the complex offset parameters of the present modification are the same as the modified example, and all of them are based on the first updating method. Further, in the above-described embodiment and the first modification, in order to calculate a new quantization offset matrix, a previous (for example, the previous) quantization offset matrix which has been set in advance is used, but in the present modification, the previous use is utilized. The calculated arbitrary quantization offset matrix is used to calculate a new quantization offset matrix. Fig. 22 is a view showing the configuration of the offset parameter set D1800 of the present modification. The offset parameter set D1800 is equivalent to the offset parameter sets D306 and D312 shown in Figs. 6A and 6B, and is formed by the complex offset parameter d18〇2b. When there are complex conversion blocks of different sizes in the sequence or picture, the offset parameter D18〇2 is included in the offset parameter set D1800 according to the sizes. For example, as shown in Fig. 22, the offset parameter set D18〇〇 contains one offset parameter 1) 1802, which corresponds to the size of the first to third conversion blocks, respectively. The offset parameter D1802 includes an offset update flag Dl8〇1, and an offset update parameter _(10)3 for calculating the quantization offset. The offset is more the same as the offset _ 39 201138477 is the same as the offset update flag D100 shown in Figure 7, which is used to indicate whether the flag of the quantization offset value of the conversion block should be updated. That is, the offset update flag D1801 indicates whether or not the quantization offset matrix should be calculated, thereby updating the flags of the quantization offset values included in the quantization offset matrix before the update. The offset update parameter D18 03 is configured immediately after the offset update flag D18 01. Further, the offset update parameter D1703 is different from the configuration of the offset update parameter D102 corresponding to the first update method shown in Fig. 7, and does not include the update type identifier D104. Further, the offset update parameter D1803 differs from the configuration of the offset update parameter D1703 shown in Fig. 19 as containing the matrix identifier D1804. That is, the offset update parameter D1803 includes a matrix identifier D1804, first and second correction parameters D1806, and D1808, (5 number of offset values D1810, and N consecutive 5 offset values D1812. Matrix identifier D18 04 is an identifier for identifying a quantization offset matrix or a quantization scaling matrix. The matrix identifier D1804 can be specifically utilized to calculate a new quantization matrix and to use the quantization matrix previously used or The image decoding device 1000 and the image encoding device 2 according to the present modification are the same as the above-described embodiment, and the processes shown in Figs. 8 and 9 are executed. The decoding device 1000 and the image encoding device 2 perform the processes shown in Fig. 20 and Fig. 21 in the same manner as in the first modification. Fig. 23 shows the image decoding device 1 of the present modification. The flowchart of the process of analyzing the offset update parameter is shown in Fig. 23. The figure 23 shows the detailed processing of step S1904 of Fig. 20. First, the offset parameter analysis unit 1500 analyzes the offset stored in the header more 40 201138477 Within the new parameters The matrix identifier Dl8〇4 (step S21〇〇), and the offset parameter analysis unit 1500 analyzes the first correction parameter stored in the offset update parameter of the header (step S2102), and analyzes the second correction separately. Parameter (Step S2104) Then the 'Offset parameter analysis unit 15' analyzes the number of δ offset values stored in the offset update parameter of the header (Step S21〇6). Further, the offset parameter analysis unit 1500 The complex number (5 offset value) stored in the offset update parameter of the header is parsed in the scanning order according to the above-mentioned number 'supposed in step S21〇6. (Step S2108). Here, the offset is stored. The complex number in the update parameter is continuous (5 offset values are arranged in scan order in a 2-dimensional matrix corresponding to the transform block. For example, the δ offset values are in a 2-dimensional matrix corresponding to the transform block, and are subjected to zigzag scanning. The order is arranged from the low frequency component to the high frequency component, and the scanning order differs depending on the size of the conversion block. For example, the scanning order of the block (8x2 block) formed by the frequency component (conversion coefficient) of 8χ2 is different. Block formed by a frequency component of 2x8 (2x8 area) The image decoding apparatus 1000 of the present modification is configured to analyze the offset update parameter by the processing shown in FIG. 23, and then perform the offset update parameter analyzed by the image decoding apparatus 1000 as shown in FIG. In the same manner, the image decoding device 1 of the present modification calculates a quantization offset value (quantization offset matrix). Here, the image decoding device 1 of the present modification is present. When the quantization offset value is obtained in step S1200 shown in Fig. 13, the matrix identifier D1804 parsed in step S2100 of Fig. 23 is used. That is, the quantization offset value calculation unit 1502 specifies the quantization offset matrix or the quantization scaling matrix which is stored in the first memory portion 1506 and which is recognized by the matrix identifier D1804. Further, the quantization offset value calculation unit 1502 reads out, from the first memory unit 1506, the quantization offset at the same position as the quantization offset value of the calculation target in the localization offset matrix or the quantization scaling matrix defined by the 201138477. value. Fig. 24 is a flow chart showing the processing of the matrix before (previously) the image decoding apparatus of the present modification. First, the quantization offset value calculation unit 1502 specifies the value displayed by the matrix identifier analyzed by the offset parameter analysis unit 15 (step S23), and the quantization offset value calculation unit 1502 determines the matrix identifier display. If the value determined in advance is the value set in advance (Yes in step S2302), the value is determined from the first memory unit 15〇6. a quantization scaling matrix corresponding to the value displayed by the matrix identifier (step S23〇6). Further, the quantization offset value calculation unit 1502 sets the quantized scaling matrix retrieved therein to be used to calculate a new quantization offset matrix. The old quantization offset matrix (that is, the quantization offset matrix of the update target) (step S23〇8). On the other hand, the quantization offset value calculation unit 15〇2 determines that the matrix identifier is displayed in step S23〇2. The value is not a value set in advance (No in step 823〇2), that is, the quantization offset matrix directly corresponding to the value displayed by the matrix identifier is retrieved from the first memory unit 1506 (step S23〇4). The value calculation unit 1502 will The searched quantization offset matrix is used as an old quantization offset matrix (i.e., a quantization offset matrix to be updated) used to calculate a new degraded offset matrix. Image coding apparatus 2 of the present modification The system performs the same processing as the processing shown in Fig. 4, and calculates a 5-offset value (a two-dimensional matrix including five offset values). Here, the image encoding apparatus 2 of the present modification 〇〇〇 The offset parameter calculation 42 201138477 deletes the quantized offset matrix or the quantization scale matrix stored in the first memory portion 16G4 when the quantization offset value is obtained in step S13 (K) shown in the figure. Among them, the specific matrix is used as the matrix before the old update. At this time, the image encoding device 2000 performs the same processing as the processing shown in Fig. 24 to specify a matrix. ^ reads from the first memory unit 1604 the quantized offset value at the same position as the δ offset value of the calculated object in the matrix (quantization offset matrix or quantization scaling matrix) that is specified before the update ( Or quantify proportionally increase or decrease step size Fig. 25 is a flow chart showing the process of writing the offset update parameter into the header by the image encoding device 2 of the present modification, and the twenty-fifth figure shows the step S2 of Fig. 21 First, the offset parameter writing unit 1602 writes a matrix identifier for identifying a matrix (quantization offset matrix or quantization scaling matrix) specified by the offset parameter calculation unit 1600 into a header (step S2200) Next, the offset parameter writing unit 1602 writes the third correction parameter used in the offset parameter calculation unit 16 to the header (step S2202), and then the offset parameter calculation unit 16〇〇 The second correction parameter used in the writing is written in the header (step S22〇4). Further, the offset parameter writing unit 1602 calculates the number of the offset values to be written (step S22〇6^, for example, calculation (the calculation of the number of the 5 offset values is performed in the 2-dimensional matrix) The position of the last non-zero 5 offset value in the scan order is performed. However, the matrix corresponding to the quantization offset matrix is a matrix containing 5 offset values in the element. Then the 'offset parameter write The input 1602 will be calculated as such The number of &lt;5 offset values is written to the header (step S22〇8). However, the offset parameter writing unit 16〇2 also 43 201138477 can write a parameter indicating the number of its δ offset value to the header. Further, the offset parameter writing unit 1602 writes the offset value calculated by the offset parameter calculating unit 160 to the header in the scanning order in accordance with the number calculated in step S2206 (step S2210). That is, the offset parameter writing section 16〇2 successively writes the offset value to the header until the position of the 5 offset value written is equal to the last non-zero delta offset in the scanning order in the above 2-dimensional matrix. The location of the value. The image decoding method and the image encoding method of the present invention have been described with reference to the above embodiments and modifications thereof, but the present invention is not limited thereto. For example, in the present embodiment and its modifications, the offset update denominators D304 and D310 are arranged in the update offset parameter flags D302, D308, and the offset parameter set as shown in FIG. 6 and FIG. Between d306 and 〇312, but it may not be in its place. At this time, the offset update denominator may be included in the offset update parameter D102 constituted by the first update method. Further, in the above-described Modification 3, as shown in Fig. 24, the image decoding device 1 specifies or retrieves the pre-update (previous) matrix by the matrix identifier, but the 'image encoding device 2000 may Similarly to the processing shown in Fig. 24, the matrix identifier is used to identify or retrieve the pre-update (previous) moment brake. Further, in the above-described embodiments and their modifications, the flags (the update offset parameter flag or the offset update flag) are respectively displayed as 丨 and the display is 〇 when the processes are different from each other, however, the flag is The relationship between the values shown and the treatment of their values may also be the opposite. For example, as shown in FIG. 8, when the image decoding apparatus 1000 determines in step S402 that the update offset parameter flag is displayed as 1, the processing of steps S404 to S408 is executed, but conversely, the determination is an update offset. When the parameter flag is displayed as 〇, the processing of steps S4〇4~S4〇8 44 201138477 may be performed. Furthermore, an aspect of the present invention may also be a decoding method as described below. The decoding method decodes the quantization offset parameter (offset parameter) used for the image stencil horse, and parses the flag (S4 〇〇) from the header of the picture, and determines whether the flag has a display beforehand. The set value (_), if the flag has a value set in advance, that is, the parameter of the offset update denominator is parsed from the header (S404), and the newly updated quantization offset matrix value is decoded from the header. (Quantization offset value) (S406 and S408). Here, the analysis of the offset update denominator is not necessary, and the above-described offset update denominator may be set in a value defined in advance. Moreover, the process of decoding the new quantization offset matrix (quantization offset matrix value) is as follows: the offset update flag is parsed from the header of the picture, and the offset update flag is determined. There is a value set in advance (S601), and if the offset update flag indicates that the value set in advance is set, that is, the parameter of the updated type identifier is parsed from the reference flag (S6〇2), It is also determined whether or not the parameter analyzed as described above indicates the second predetermined value (S6〇4). (4) The parameter of the analysis _ _ 别 有 有 有 有 有 参数 参数 参数 参数 参数 参数 参数 参数 参数 参数 参数 参数 参数 参数 参数 参数 参数 参数 参数 参数 参数 参数 参数 偏移 偏移 偏移 偏移 偏移 偏移 偏移 偏移 偏移 偏移 偏移 偏移 偏移 偏移 偏移 偏移 偏移 偏移 偏移Quantization offset value _ The parameter of the updated lion identifier is not set to the value set by $2, and whether the parameter of the updated type identifier analyzed above is displayed or not is displayed. S6〇6). If the parameter of the update type of the resolved itj is the value of the third predetermined value, the new quantized offset matrix value is parsed from the header (S616). 45 201138477, the parameter of the updated type identifier that is parsed does not show the third pre-valued 'that is, the single-quantization offset value is parsed from the header (S612), and all the new quantized offsets are converted. The value may be set to a single internalization offset value (S614) that has been resolved before. In addition, 'the process of decoding the new quantized offset matrix (quantization offset matrix value), such as the header of the following system picture parsing out the offset update flag (si9〇〇&quot; and determining whether the flag is 'reported' or not The first set value (8) is moved in advance. If the flag has a value set in advance by 7F, that is, the offset update parameter (Sl9〇4) is taken from the header, and the new quantization offset value (S1906) is calculated according to the analyzed offset update parameter. can. In addition, in the case where the offset update parameter is analyzed, the first correction parameter (S8GG) is parsed from the header, and the second correction parameter (S802) is parsed from the header, and the display is parsed from the header. The parameter (S8〇6) of the number of encoded offset values may be based on the number of previously encoded g-shift values of the parsed ^, and the delta offset value (S8l〇) may be parsed from the header. Further, the processing of the offset update parameter is analyzed. For example, the parameter (10) 00 of the matrix identifier is parsed from the header, and the 钐 correction parameter (s2i 〇 2) is parsed from the header, and then the header is parsed. 2 Correction parameter (8) (4) After the self-described scale, the parameter indicating the number of encoded δ offset values is precipitated (S2106), and the header is self-recited according to the number of the encoded δ offset values analyzed. It is also possible to analyze the delta offset value (S21〇8). Moreover, the processing of the quantized offset matrix value is analyzed, as follows: the initial quantization offset value is initialized to a value defined in advance (8), and the offset value is analyzed from the header in the scan order (4 And adding the initial offset value to the solution 46 201138477, and then calculating _ (8)_), determining that the calculated quantization offset value is the last quantization offset value of the offset matrix is called Household =: The offset value is not the last quantized offset value w. The value of the _ value of the matrix is initialized to: (S1 ° ° 8), if the value calculated above is not ambiguous, it will be Scanning/two (four) offset value hx decoding (measurement 4), if the calculated 1 quantization value is 0, the remaining material; t-offset value is set to (4) Μ _. The process of calculating the new quantized offset value by using the offset update parameter, as shown in the following figure, and determining the coefficient determined by the third correction parameter The old quantized offset value is increased by a factor of 1 to obtain a first value (S1202)' and the difference between the old quantized offset value and the fixed value is derived (S1204), and then the second correction parameter is used by || The coefficient of determination is calculated by increasing or decreasing the difference 2 to calculate the second value (S1206), and then the first! The value is added to the scalar 2 value to obtain the calculated offset value (S1208), and the δ offset value is added to the aforementioned calculated offset (4), thereby calculating the new quantized offset value (S1210). Here, the coefficient obtained by the first correction parameter is obtained by calculating the sum of the value of the first correction parameter and the offset update denominator, and dividing the sum of the sum by the value of the offset update denominator. Calculate the ratio. Further, the coefficient obtained by the second correction parameter is a ratio obtained by dividing the second correction parameter by the value of the offset further. Further, the aforementioned fixed value is a value defined in advance. Alternatively, the aforementioned fixed value is a quantized offset value of the DC position of the conversion block. 47 201138477 or 'The aforementioned fixed value is the average offset value of the conversion block. And one aspect of the present invention may also be an encoding method as described below. The encoding method is a method of encoding the quantization offset parameter (offset parameter) used in image encoding to have the following processing: writing a flag to the header of the picture (S 5 00), and determining the quantization offset matrix value Whether (quantization offset matrix) should be updated (S502) 'If the quantization offset matrix value should be updated, the parameter of the offset update denominator is written to the aforementioned header (S504), and the newly updated quantization offset matrix value is encoded. Go to the aforementioned header (S508 and S510). Here, the writing of the offset update denominator is not necessary, and the offset update denominator may be set to a value defined in advance. In addition, the process of encoding a new quantized offset matrix, as described below, writes an offset update flag to the header of the picture (S700), and determines whether to use a new quantized offset matrix (S701)' when using a new offset matrix. At this time, the parameter of the update type identifier is written in the aforementioned header (S702), and it is determined whether or not the aforementioned quantization offset matrix value should be encoded by the update method (S704). If the foregoing quantization offset matrix value should be encoded using the foregoing update method, that is, the offset update parameter is calculated using the foregoing update method according to the new quantization offset matrix value (S7〇8), and the offset update parameter is written into the foregoing label. Head (S710). If the foregoing quantization offset matrix value is not to be encoded using the foregoing update method (s 7 〇 6 ), it is determined whether the new quantized offset matrix value is the same value, and if the new quantized offset matrix value is not the same value, then The new quantized offset matrix value is written to the aforementioned header (S716). If the value of the new quantization offset matrix is the same value, a single quantization offset value (quantization offset matrix value) may be written in the header (S712). In addition, the process of encoding the new quantized offset matrix is as follows: the new flag is written to the header of the picture (S2000), and the new quantized pseudo-shift matrix is used (S2002). The offset matrix, that is, the offset update parameter is calculated according to the new quantized shift value (S2004), and the offset update parameter is also the same as the foregoing header (S2006). Further, the process of writing the offset update parameter is as follows: the first correction parameter is written in the header (S900), and the second correction parameter is written in the front frame (S902) to calculate the 5 offset value. The number (S904), and the parameter 'displaying the number of the encoded offset value' is written to the foregoing header (S906), according to the number of 5 offset values written and encoded by the previous fan, 5 offset value is written to the previous block to find the header (S910). 'And' the process of writing the offset update parameter, as shown below, the parameter of the de-interval identifier is written into the header (S2200), and When the first correction parameter is written into the header (S2202) and the second correction parameter is written into the header, the number of the 5 offset values is calculated (S2206), and the number of the encoded offset values is displayed. After the parameter is written into the foregoing header (S2208), the δ offset value may be written into the former fascinating head (S2210) according to the number of δ offset values written and encoded by the former hacker. The process of quantizing the offset matrix value to be written, as the following initializes the initial reference offset value to a previously defined value (sll〇〇), and scans from The next new quantized offset value in the order, minus the aforementioned reference offset value, thereby calculating the 5 offset value (S1104), encoding the 5 offset value to the aforementioned header (S1106)' and the conversion block The next position corresponding to the reference offset value is set to the aforementioned new quantized offset value (S1108), and it is determined whether the remaining quantized offset value in the scan order is the same value as the aforementioned reference offset value (SU1〇). 49 201138477 If the remaining quantization offset value in the scanning order is the same value as the aforementioned reference offset value, the aforementioned δ offset value is encoded to the negative value of the aforementioned reference offset value (S1112) 'If remaining in the scanning order If the quantized offset value and the reference offset value are different values, the next offset value is calculated in the scan order until the last position of the conversion block is calculated (sll 〇 4). The process of calculating the offset update parameter is as follows: the old quantized offset value is determined as follows (S1300) and the old offset offset value is scaled up and down by a coefficient determined by the second correction parameter, thereby calculating The third value (§ 13〇2), and derive the aforementioned old quantization The difference between the shift value and the fixed value (sl3〇4), and the difference is proportionally increased or decreased by the coefficient determined by the second correction parameter, thereby calculating the second value (S1306), and then the 丨 value And adding the second value, obtaining the calculated offset value (S1308), and subtracting the calculated offset value from the new quantized offset value to calculate the S offset value (S1310). The coefficient obtained by the first correction parameter is calculated by dividing the sum of the first correction parameter and the value of the offset update denominator and dividing the sum by the value of the offset update denominator. The coefficient obtained by the second correction parameter is a ratio calculated by dividing the second correction parameter of the month by the value of the offset update denominator. Further, the aforementioned fixed value is a value defined in advance. Or the aforementioned fixed value is an offset value of the DC position of the conversion block. Here, in the method of encoding and decoding the above-described quantization offset parameter, the processing for determining the value of the old quantization offset matrix may be as follows. The parameter value of the matrix identifier is determined (S2300), and it is determined whether or not the parameter of the aforementioned matrix identifier indicates a value set in advance (S2302). If the parameter of the aforementioned matrix 50 201138477 identifier shows a value set in advance, the matrix value is quantized from the memory (S2306), and the quantized offset matrix value is set as the retrieved quantization scale matrix value (S2308). If the parameter of the matrix identifier does not show a value set in advance, the quantization offset matrix value is retrieved from the memory (S2304). Further, one aspect of the present invention may be the following adjustment method. The adjustment method adjusts the inverse quantization value by the quantization offset value, and has a process of acquiring the quantization offset value (S1400), and subtracting the quantization offset value from the previously defined value, thereby calculating the scale value. (sl4〇2), and multiplying the aforementioned scale value by the quantization scale step size (S1404), and dividing the value of the product by the aforementioned predefined value, thereby calculating the adjustment value (sl4〇6) Thereafter, the aforementioned adjustment value (sl4〇8) is subtracted from the aforementioned inversely quantized value. Here, the aforementioned value defined in advance is 128. However, the present invention can be implemented not only by the above decoding method, encoding method, and fading method, but also as a device for performing decoding, encoding, and adjustment in this way, an integrated circuit, and a computer for performing processing according to the method. The program, and the storage medium for storing the program. (Embodiment 2) By recording a program for realizing the configuration of the image encoding method or the image decoding method described in the above embodiment to a memory medium, the processing shown in the above-described target form can be easily performed. The ground is delivered in a separate computer system. § I have recalled that the media can be a disk, a CD, a CD, a 1 (^ card, a semiconductor memory, etc.). In addition, the image coding method shown in the above embodiment will be described here. And an application example of the image decoding method and a system using the same. Fig. 26 is a diagram showing the overall configuration of the content supply system ex100 that realizes the content delivery service. The communication service providing area is divided into a desired size' Each of the service areas is provided with a base station exl06~exl1 of a fixed wireless station, the content supply system ex100 is transmitted through the Internet service provider ex102 and the telephone network exl〇4, and the base station exl〇6~exll〇 The computer exlll, the PDA (Personal Digital Assistant) exll2, the camera exll3, the mobile phone exll4, and the game machine are connected to the Internet exlOl with each machine such as 115. However, the content supply system exlOO is not limited to the figure 26 The configuration may be such that the optional elements are combined and connected. Alternatively, the base stations exl06 to exll〇 may not be passed through the fixed wireless station. It is connected to the telephone network exl04. In addition, each device can be directly connected to each other through short-range wireless, etc. The camera e X113 is a device that can perform motion picture recording such as a digital video camera, and the camera exll6 can perform still image photography such as a digital camera. And the machine for motion picture photography. In addition, the mobile phone exl 14 can be GSM (Global System for Mobile Communications), CDMA (Code Division Multiple Access), W-CDMA (Wideband- Code Division Multiple Access: LTE (Long Term Evolution), HSPA (High Speed Packet Access), or PHS (Personal Handyphone System: Personal Hand) 52 201138477 Machine system), etc. In the content supply system exlOO, the camera exll3 or the like is connected to the streaming server exl03 through the base station exl09 or the telephone network ex104, thereby achieving on-site instant transmission, etc. The user can use the camera exl as described in the above embodiment. The content of the l3 shot (for example, an image of a live music concert, etc.) is encoded and transmitted to the streaming server ex103. On the other hand, the streaming server ex103 can stream the content data transmitted by the customer in need. Among the customers, there are a computer exlU, a PDAexll2, a camera exll3, a mobile phone exll4, or a game machine exll5 which can decode the above-mentioned encoded data. In each machine that receives the transmitted data, the received data can be decoded and reproduced. The encoding processing of the photographed data may be performed by the camera exll3, or by the streaming server ex103 transmitting the processing data, or may be shared with each other. Similarly, the decoding process of the transmitted data can be performed by the client, or by the streaming server ex103, or can be shared with each other. Further, it is not limited to the camera exll3, and the still image and/or moving image data captured by the camera exll6 may be transmitted to the streaming server ex103 via the computer exl 11. The encoding processing at this time may be performed by any of the camera exll6, the computer exl11, or the streaming server ex103, or may be shared with each other. Further, these encoding processing and decoding processing are generally performed in a computer exlll and an LSI (Large Scale Integration) ex500 owned by each device. The LSI ex500 can be a single wafer or a composite wafer 53 201138477. Further, the software for image encoding and image decoding is mounted to a recording medium (CD_R〇M, floppy disk, hard disk, etc.) that can be read in a computer exlll or the like and the encoding process is performed by the software. Decoding processing is also possible. In addition, when the mobile phone exll4 is attached with a camera, it can be transmitted with the dynamic image data obtained by the camera. The motion picture data at this time is encoded by the LSI ex500 equipped with the mobile phone ex 114. Moreover, the serial server exl03 can be a plurality of servers or a plurality of computers, and can also process or record the distributed data and pass the wheel. As described above, in the content supply system ex100, the client can receive the encoded material and reproduce it. In this way, in the content supply system exl〇〇, the customer can immediately receive the information transmitted by the user and decode and reproduce it, without the user having special rights or equipment, and can also realize personal play. The image encoding method or image demapping method shown in the above-described embodiment can be applied to the encoding or decoding of each device constituting the content supply system. As an example, the mobile phone exll4 will be described. Fig. 27 is a view showing a mobile phone ex14 using the image coding method and image decoding method described in the above embodiment. The mobile phone exi μ has an antenna ex601 for transmitting and receiving radio waves between the base station and the base station, a camera unit ex603 for capturing images and static images such as a CCD camera, and a display unit ex602 for decoding. The image captured by the camera unit ex6〇3 or the image received by the antenna ex601 is displayed on the liquid crystal display or the like; the main body is composed of the operation keys of 6〇4 groups; the voice output unit Ex608, used for voice output, σ 八 八; voice input 54 201138477 into 4ex605, used for voice input microphone, etc.; recording media ^ 7, is used to save the recorded dynamic image or static image data, Received mail data, dynamic data or static image data, and encoded or decoded data; and slots (81 〇〇 以 6 〇 6, can be used to install. Recorded media ex6G7 installation The mobile phone is 114. The recording medium ex6〇7 is stored in the flash memory component, and (4) the flash memory component is EEPROM (Electncally Erasable and Programmable Read Only Memory) One type of read-only memory is a non-electrical memory that can be electronically exchanged and erased in a plastic case such as an SD card. In addition, the mobile phone exll4 is illustrated in Figure 28. The mobile phone is looking for the 114 series through the synchronous bus ex713 The power supply circuit unit ex71, the operation input control unit ex704, the image coding unit ex712, the camera interface ex703, the LCD (Liquid Crystal Display) control unit ex7〇2, the image decoding unit ex709, and the demultiplexing unit Ex708, recording/reproduction unit ex7〇7, modulation/demodulation circuit unit ex706, and sf·sound processing unit ex7〇5 are connected to main control unit ex7丨1 (for display control unit ex602 and operation key ex604) When the user ends up the call and the power button is turned on by the user's operation, the power supply circuit unit ex710 supplies power from the battery pack to each unit, thereby enabling the camera digital mobile phone exl 14 to be activated. The state of the mobile phone exll4 is based on the control of the main control unit ex711 formed by the CPU, the ROM and the RAM, and the voice input unit 6 in the voice call mode; &lt;6()5 The voice signal collected by the voice processing unit ex705 is converted into a digital voice resource 55 201138477 ^ To adjust the spread spectrum processing unit ex706 to perform the spread spectrum processing and to receive the transmission circuit The portion is transmitted through the antenna ex601 after the digital analog conversion processing and the frequency conversion processing are applied at 7.1. Moreover, the mobile phone eXU4 can expand the received data received by the antenna ex601 in the voice call mode and apply frequency conversion processing and analog digital conversion processing, and then despread the frequency by the modulation and demodulation circuit unit ex7〇6 ( The dispreading spectrum is processed and outputted by the voice output unit ex6〇8 after being converted into analog voice data by the voice processing unit. Further, when the e-mail is transmitted in the data communication mode, the file information of the e-mail input by the operation of the operation key ex6〇4 of the body unit is transmitted to the main control unit ex7u via the operation input control unit ex7〇4. The main control unit ex711 spreads the file data by the modulation/demodulation circuit unit ex7〇6, applies digital analog conversion processing and frequency conversion processing to the reception transmission circuit unit at 7〇1, and then transmits the antenna through the ex6〇1 The base station is transmitted at 11 miles. When the image data is transmitted in the data communication mode, the image data captured by the camera unit ex6〇3 is supplied to the image encoding unit ex712 via the camera interface ex703. Further, when the image data is not transmitted, the image data captured by the camera unit ex6〇3 can be directly displayed on the display unit ex6〇2 via the camera face ex703 and the LCD control unit ex702. The image encoding unit ex712 is provided with the image encoding device described in the present invention, and the image encoding device shown in the above-mentioned application is provided with a (four) encoding method 'supplied from the camera unit ex6Q3. It is compressed into a coded image data, and is sent to the multiplexed department ex7G8. In addition, at the same time, the mobile phone, such as the 14th series, uses the speech processing unit eX705 to transmit the voice collected by the camera unit ex6G3 with the voice input unit ex6〇5 as digital voice data to the multiplex unit π· . The multiplexed section 8 is a modified solution from the image data supplied from the image encoding unit ex712 and the voice supplied from the voice processing unit ex7〇5. The modulation circuit unit π· purely spreads the multi-X data obtained as a result, and applies the digital analog conversion processing and the frequency conversion processing to the transfer transmission circuit unit ex7〇1, and then transmits the data through the antenna ex601. When receiving the dynamic image slot data that has been connected to the first page in the data communication mode, the data is processed by the modulation demodulation circuit unit ex7 through the antenna sub-interface and received from the base station exllO. The multiplexed data obtained as a result is sent to the multiplexing department ex7〇8. Moreover, in order to decode the multiplexed data received through the antenna ex6〇i, the demultiplexing unit ex708 multiplexes the multiplexed data to separate the bit stream of the image data and the bit of the voice data. The stream is supplied to the image decoding unit ex7〇9 via the sync bus ex113, and the voice data is supplied to the voice processing unit ex7〇5. Next, the image decoding unit ex709 includes the image decoding device described in the present application, and decodes the bit stream of the image data by the decoding method corresponding to the encoding method shown in the above embodiment. Thereby, the reproduced moving image data is generated and supplied to the display unit 602 via the LCD control unit ex702. Thereby, for example, dynamic image data contained in the sorrow image file connected to the home page can be displayed. At the same time, the voice processing unit ex7〇5 57 201138477 converts the voice data into analog voice data, and supplies the voice data to the voice wheeling unit ex608. By this, for example, the voice material contained in the motion picture file connected to the home page can be reproduced. And 'not limited to the above system example, recently, digital playback of satellites and ground waves (gr〇Und wave) has become a topic'. As shown in FIG. 29, at least the image coding apparatus or the diagram of the above embodiment may be used. Like a decoding device, it is installed in a digital playback system. Specifically, in the broadcast station ex2〇i, the voice data, the image data, or the bit stream of the multiplexed data may be transmitted to the communication or broadcast satellite ex202 by radio waves. Receiving the broadcast satellites at 2〇2, the broadcast wave will be transmitted, and the home antenna ex204 having the satellite broadcast receiving device receives the wave, and the television (receiver) ex3〇〇 or the transponder (STB) ex217 The device will decode the bit stream and regenerate it. Further, the image decoding device shown in the above embodiment may be mounted in the reader/writer ex218. The reader/recorder ex2丨8 reads the recording medium such as CD and DVD recorded on the recording medium with the image data of 215 and 216, and the bit stream of the multiplexed voice data or Decoder. At this time, the image signal reproduced by the household can be displayed on the monitor ex219. Further, the image decoding device may be attached to the transponder ex2_ connected to the antenna ex2G4 of the L2〇h satellite/ground wave for cable television, and the image is reproduced by the monitor 19 of the electric device. Composition. At this point, the image decoding body can also be mounted to the TV - not to the transponder 1. Further, in the car ex210 having the antenna eX2〇5, it is also possible to receive the speech from the satellite ex2〇2 or the base station or the like and reproduce the moving image in a display device such as the ex: 211 provided in the car ex21G. 58 201138477 Further, the image decoding device or the image encoding device shown in the above embodiment may be attached to the reader/writer ex218. The reader/recorder ex218 reads and decodes the voice data, the image data, or the encoded bit stream of the recording medium ex215 that has been recorded to the DVD, BD, etc. (7) recording medium ex215, or The voice data, the video data, or the data is encoded as multiplexed data and recorded to the recording medium ex215. At this time, the reproduced image signal can be displayed on the monitor ex219. Further, by recording the recording medium ex215 of the encoded bit stream, other devices and systems can reproduce the video signal. For example, the other reproduction device ex212 can reproduce the video signal to the monitor ex213 using the recording medium ex214 in which the encoded bit stream is copied. Alternatively, the image decoding apparatus may be mounted in a transponder ex217 connected to the cable ex203 for cable television or the antenna ex204 for satellite/ground broadcasting, and displayed on the monitor ex219 of the television. At this time, the image decoding device can be mounted to the television instead of the transponder one. Fig. 30 is a view showing a television (receiver) ex300 using the image decoding method and the image coding method described in the above embodiment. The television ex300 includes a tuner ex301 that acquires or outputs a bit stream of video information by receiving the antenna ex204 or the cable ex203 that is being played, and the modulation/demodulation unit ex302 decodes the received encoded data. Tuning, or, for transmitting the generated encoded data to the outside, and multiplexing/demultiplexing, ex303, to demodulate the demodulated image data and voice data, or The multiplexed image data and voice data are multiplexed. Further, the television ex300 includes a signal processing unit ex306, which has a voice signal processing unit ex3〇4 and a scene signal processing unit that decodes the voice data and the video data, respectively, or encodes each of the funds 59 201138477. Ex305; and the output unit is 3〇9, which is a display unit ex such as a display having a decoded voice signal output, and a display unit ex such that a decoded video signal is displayed. Further, the television set has an interface ex317' which has an operation input unit 6X312 for receiving a user operation. Further, the television set 300 has a control unit eX31 for controlling the respective units and a power supply circuit unit ex311 that supplies power to each unit. The interface portion 317 may have a bridge ex313 connected to an external device such as the reader/recorder 6X218, an SD card, etc., in addition to the operation of the wheel portion. The slot ex3i4 of the recorded media ex2i6, the driver ex315 connected to an external recording medium such as a hard disk, and the data processor 6?&quot;316 connected to the telephone network. However, the recording media technology 216 can perform electronic information recorders by storing non-electrical/electrical semiconductor recording elements. Each part of the television ex300 is connected to each other through a synchronous bus. First, the configuration in which the television ex300 reproduces the poems obtained from the outside through the antenna ex2〇4 or the like is described. The television ex3GG receives the user operation from the remote controller ex220, etc., according to the control of the control unit J〇Pex310 having cpu, etc., and the multiplexer/demultiplexing unit will use the modulation/demodulation unit ex302 at 3〇3. The demodulated image data or voice data is demultiplexed. Further, the television ex300 decodes the demultiplexed voice data by the voice signal processing unit by 3〇4, and uses the decoding method described in the above embodiment to use the decoding method described in the above embodiment. The image data of Xie Duowei is decoded. The decoded voice signal and video signal are respectively rotated outward from the output unit ex3〇9. In order to synchronize the voice signal and the video signal during the output, 60 201138477 temporarily accumulates the signal to the buffer to 318, 6) (319 etc. In addition, the TV ex300 can also be from the magnetic gas / CD, SD card The recording medium is read by a coded bit stream of 215, ex216 instead of playing, etc. Next, it is explained that the television ex300 encodes the voice signal and the video signal, and transmits it to the outside or writes it. The composition of the recording medium 4. The television receives the user operation from the remote controller ex220 and the like by 3, and encodes the voice signal by the voice signal processing unit ex3〇4 according to the control of the control unit ex31, and processes the image signal by the video signal. The part ex305' encodes the video signal using the encoding method described in the above embodiment. The encoded voice signal and the video signal are multiplexed by the multiplex/demultiplexing unit ex303 and output to the outside. The signal is synchronized with the image sfl number, and the signals may be temporarily accumulated in the buffers ex320, ex321, etc., and as shown, the buffers ex318 to ex321 may be provided in plurality, or may be shared In addition to the display, for example, between the modulation/demodulation unit ex3〇2 and the multiplex/demultiplexing unit ex303, data may be accumulated in the buffer as In addition to the acquisition of voice data and video data from playback and recording media, the ex300 of the TV can also be configured to receive the input of the microphone and the camera. The data obtained is subjected to encoding processing, etc. Here, although the television ex300 is described as a configuration that can achieve the above-described encoding processing, multiplexing, and external output, it may be impossible to perform all of the processing. Only the configuration of the above-described reception, decoding processing, and external output can be achieved. Further, when the reader/recorder ex218 reads or writes the encoded 61 201138477 bit stream from the recording medium, the above decoding processing or editing The drinking process can be optionally performed between the television ex. and the reader/recorder ex218. Alternatively, the television T can be shared by the television ex300 and the reader/recorder ex218. When the resource information, the first 31δ) Display read or write data from the CD reproducing constituting / recording unit of eX_. The information reproduction/recording department has the following. The element of the moon ex4Gl ex4G7. The optical head ex4Qi irradiates the laser spot to the recording surface of the recording medium 5 of the optical disc to write information, and (4) the reflected light from the recording surface of the recording medium ex215, and reads the information. The modulation recording unit ex402 electrically drives the semiconductor laser built in the optical head ex4〇1, and performs modulation of the laser light in accordance with the recorded data. The reproduction demodulation unit ex4()3 expands the reproduction signal from which the reflected light from the recording surface is electrically detected by the optical detector of the optical magnetic (4), and the signal component recorded in the recording medium ex215 The information needed to solve the problem of multi-guard, demodulation, and regeneration. The buffer ex4〇4 temporarily holds the information for recording to the recording medium ex215, and the information reproduced from the recording medium technology 215. The disk motor ex405 can rotate the recording medium at 215. The servo control unit ex406 controls the rotational driving of the disk motor ex405, moves the optical head ex401 to a predetermined information track, and performs tracking processing of the laser spot. The system control unit ex407 performs overall control of the information reproduction/recording unit ex4. The processing of the above-mentioned s purchase and writing is realized by the system control unit ex407 generating and adding new information as necessary, and using the various information held in the buffer ex4〇4, and making the modulation recording unit necessary. The ex4〇2, the reproduction demodulation unit ex403, and the servo control unit ex4〇6 coordinate the operation to perform recording and reproduction of information through the optical head ex401. For example, the system control unit ex4〇7 is configured by a microprocessor of 62 201138477, and executes such processing so as to execute a program for reading and writing. Although the optical head ex401 has been described as an irradiation laser spot as described above, it is also possible to use a near-field light to perform high-density recording. Fig. 32 is a view showing the recording medium ex215 of the optical disc. On the recording surface of the recording medium ex215, a race (groove) is formed in a spiral manner, and the information track ex230 is changed in advance according to the shape of the groove, and an address indicating the absolute position on the disk is recorded. News. The address information includes information for specifying the location of the block (the unit of recorded data) ex231, and the device for performing recording and reproduction can record the block by reproducing the information track ex230 and reading the address information. Further, the recording medium ex215 includes a data recording area ex233, an inner area ex232, and a peripheral area ex234. The area used to record the user data is the data recording area ex233, and the inner area ex232 and the outer area ex234 disposed in the inner or outer periphery of the data recording area ex233 are used in addition to the recording of the user data. Use. The information reproducing/recording unit ex400 reads and writes the encoded voice data and video data or the encoded data of the multiplexed data in the data recording area ex233 of the recording medium ex215. Although a single optical disc such as a DVD or a BD has been exemplified above, it is not limited to these, and may be a multi-layered structure and a disc which can be recorded outside the surface. Further, it is also possible to perform a multi-dimensional recording/reproduction structure of the optical disc, for example, at the same place of the magnetic disc, recording information with various kinds of different wavelengths of color light, or recording different information layers from various angles. Further, in the digital broadcasting system ex200, data can be received from the satellite ex202 or the like in the car ex210 having the antenna ex205 63 201138477, and the moving image can be reproduced to the car navigation provided in the car ex210 on the display device such as 211. . The configuration of the car navigation ex211 is a configuration in which the GPS receiving unit is added to the configuration shown in the third drawing, and the computer 6) (111 and the mobile phone may have the same configuration with 114, etc.) 'The above mobile phone exll4# The terminal is the same as the television ex300. In addition to the transmission and reception type terminals of both the encoder and the decoder, there are three types of installations, such as a transmission terminal having only an encoder and a receiving terminal having only a decoder. The image encoding method or the image decoding method shown in the above embodiment can be used in the above-described optional device and system, whereby the effects described in the above embodiments can be obtained. It is not limited to the above-described embodiments, and various modifications and corrections can be made without departing from the scope of the invention. (Embodiment 3) The image encoding method and apparatus, image decoding method and apparatus shown in each of the above embodiments, Typically, it can be implemented in an LSI of an integrated circuit. For example, Figure 33 shows the configuration of a single-wafered LSI ex5. The LSI ex500 has the following elements. Ex5〇1~ex5〇9, each element is connected through the bus bar ex51〇. The power supply circuit unit ex5〇5 supplies power to each unit when the power is turned on, thereby starting the operation state. For example, 'execution coding At the time of processing, the LSIex5 is based on the control unit ex5G1 with CPUex502's memory controller with 5〇3, and the stream controller to find 5〇4, etc., with AVI/〇ex5()9j^ microphone Receive the Av signal by 117 and camera exll3, etc. The signal of the input will temporarily accumulate in the memory of external memory such as SDRAM at 511. According to the control of the control unit ex5〇l, the accumulated data will be processed according to the processing amount. And the processing speed is appropriately divided into a plurality of times and then transmitted to the signal processing unit ex5〇7. The signal processing unit ex5〇7 performs encoding of the voice signal and/or encoding of the video signal. Here, the encoding process of the video signal is The encoding process described in the above embodiment. Further, in the signal processing unit ex5〇7, the multiplexed processing of the encoded voice data and the encoded video data is performed as appropriate, and the stream I/ is streamed. Oex506 Output to the outside. The output bit stream will be transmitted to the base station exl07 or write to the recording medium ex215. However, in order to achieve synchronization, the data can be temporarily accumulated in the buffer ex5〇8. When the decoding process is executed, the LSI ex500 temporarily stores the coded data obtained by the stream I/0ex5〇6 through the base station ex丨〇7 or the coded data read from the recording medium ex215 under the control of the control unit ex501, and temporarily stores the coded data in the memory. In the control of the control unit ex5〇i, the accumulated data can be appropriately divided into a plurality of times in accordance with the processing amount and the processing speed, and then transmitted to the signal processing unit ex507. The signal processing unit ex5〇7 performs decoding of the voice data and/or decoding of the video data. Here, the decoding process of the video signal is the decoding process described in the above embodiment. In addition, depending on the situation, in order to synchronize the decoded voice signal and the decoded video signal, the respective signals may be temporarily accumulated in the buffer ex5〇8 and the like. The decoded round-trip signal can be output from each of the output units such as the mobile phone exll4, the game machine exll5, and the television ex3〇0 through the memory ex511 or the like as appropriate. In the above, the memory ex511 is described as an external configuration of the LSI ex500. However, the configuration may be included in the LSI ex500. Buffer 65 201138477 ex508 is also not limited to a single one, but may also have a complex buffer. Further, the LSI ex5 〇 0 can be single-wafered or can be multi-wafered. The term "LSI" is used herein, but may be referred to as 1C, system LSI, Super LSI, or Ultra LSI depending on the degree of integration. Further, the method of integrating the circuit is not limited to the LSI, and can be implemented in a dedicated circuit or a general-purpose processor. After the LSI is manufactured, a programmable FPGA (Field Programmable Gate Array) or a reconfigurable processor that can be connected and set by a circuit cell in a reconfigurable LSI can be used. In addition, if there is a technology that replaces the LSI's integrated circuitization due to advances in semiconductor technology or other technologies derived from it, it is a matter of course, and it is also possible to use the technology to perform the integration of functional blocks. It may also be applicable to applications in biotechnology. Although the coding method, the coding apparatus, the decoding method, and the decoding apparatus of the present invention have been described above based on the embodiments, the present invention is not limited to the embodiments. Other aspects that can be applied to the embodiments of the present invention and the constituent elements and steps of the different embodiments are combined without departing from the scope of the present invention. It is intended to be included within the scope of the invention. INDUSTRIAL APPLICABILITY The image decoding method and the image encoding method of the present invention have an effect of reducing the number of codes of a coded stream, and can be applied to, for example, an image decoding device, an image encoding device, a video camera, and a reproducing device. A mobile phone with a dynamic image reproduction function, and a personal computer. 66 201138477 [Simple description of the diagram] Figure 1 shows a graph of the inverse quantized values adjusted by the quantized horizontal offset value. Fig. 2A is a block diagram showing the configuration of an image decoding apparatus in the first embodiment of the present invention. Fig. 2B is a flow chart showing an image decoding method in the first embodiment of the present invention. Fig. 3A is a block diagram showing the configuration of an image coding apparatus in the first embodiment of the present invention. Fig. 3B is a flow chart showing an image encoding method in the first embodiment of the present invention. Fig. 4 is a block diagram showing an example of a detailed configuration of an image decoding apparatus in the first embodiment of the present invention. Fig. 5 is a block diagram showing an example of a detailed configuration of an image coding apparatus in the first embodiment of the present invention. Fig. 6A is a view showing the construction of a sequence header in the first embodiment of the present invention. Fig. 6B is a view showing the construction of a picture header in the first embodiment of the present invention. Fig. 7 is a view showing the configuration of an offset parameter set in the first embodiment of the present invention. Fig. 8 is a flow chart showing the process of decoding the quantization offset matrix in the first embodiment of the present invention. Fig. 9 is a flow chart showing the processing of encoding the quantization offset 67 201138477 matrix in the first embodiment of the present invention. Fig. 10 is a flow chart showing the process of calculating a new quantization offset matrix corresponding to one conversion block in the first embodiment of the present invention. Fig. 11 is a flow chart showing the process of calculating and writing an offset parameter corresponding to one conversion block in the first embodiment of the present invention. Fig. 12 is a flowchart showing a process in which the image decoding apparatus analyzes the written offset update parameter by the first update method in the first embodiment of the present invention. Figure 13 is a flowchart showing a process in which the image decoding apparatus calculates the quantization offset value by the first updating method in the first embodiment of the present invention. Fig. 14 is a flowchart showing a process of calculating the δ offset value of the offset update parameter by the first updating method in the first embodiment of the present invention. Fig. 15 is a flowchart showing the process of the image encoding apparatus writing the offset update parameter into the header by the first updating method in the first embodiment of the present invention. Figure 16 is a flowchart showing a process in which the image decoding device calculates the quantization offset value by the second updating method in the first embodiment of the present invention. Figure 17 is a diagram showing the image encoding apparatus according to the first embodiment of the present invention. &lt;5 Flowchart of the process of writing the offset value to the header. Fig. 18 is a flowchart showing a process of adjusting the inverse quantized value by the quantization offset value by the inverse quantization unit of the image decoding device and the inverse quantization unit of the image coding device according to the first embodiment of the present invention. Fig. 19 is a view showing the configuration of the offset parameter set in the first modification of the first embodiment of the present invention. 68 201138477 Fig. 20 is a flowchart showing a process of calculating a new quantization offset matrix corresponding to one conversion block in the first modification of the first embodiment of the present invention. Fig. 21 is a flowchart showing a process of calculating and writing an offset parameter corresponding to one conversion block in the first modification of the first embodiment of the present invention. Fig. 22 is a view showing the configuration of the offset parameter set in the second modification of the first embodiment of the present invention. Figure 23 is a flowchart showing the process of analyzing the offset update parameter by the image decoding device in the second modification of the first embodiment of the present invention. Fig. 24 is a flowchart showing the processing of the image decoding device specific matrix in the second modification of the first embodiment of the present invention. Figure 25 is a flowchart showing a process in which the image encoding device writes the offset update parameter into the header in the second modification of the first embodiment of the present invention. Fig. 26 is a view showing an example of the overall configuration of a content supply system that realizes a content delivery service. Figure 27 shows a diagram of the appearance of the phone. Figure 28 is a block diagram showing a configuration example of a mobile phone. Fig. 29 is a schematic view showing an example of the overall configuration of the digital broadcasting system. Figure 30 is a block diagram showing a configuration example of a television set. Fig. 3 is a block diagram showing an example of the configuration of an information reproduction recording unit that performs reading and writing of information on a disc recording medium. Fig. 32 is a view showing a configuration example of a disc recording medium. Fig. 33 is a block diagram showing an example of the configuration of an integrated circuit for realizing the image encoding method and the image decoding method of the respective embodiments. 69 201138477 [Description of main component symbols] Bu··Image decoding device 2··Image encoding device 10...flag analysis unit 11...determination unit 12...quantization offset matrix calculation unit 13...inverse quantization decoding unit 20...determination Part 21...flag writing unit 22...offset update parameter calculation unit 23...parameter writing unit 24··quantization coding unit 1000...image decoding device 1500...offset parameter analysis unit 1502...quantization offset value calculation unit 1504...quantization offset value storage unit 1506...first memory unit 1508...inverse quantization unit 1510...inverse conversion unit 1512...sample recombination unit 1514...sample prediction unit 1516...second memory unit 1600...offset parameter calculation unit 1602 ...offset parameter writing unit 70 201138477 1604...first memory unit 1606...subtracting unit 1608...converting unit 1610...quantizing unit 1612...inverse quantization unit 1614...inverse conversion unit 1616...sample prediction unit 1618...second memory unit 1620.·.Addition Department 2000...Image Encoding Device

Qss...quantization scale order size D100...offset update flag D102···offset update parameter D104...update type identifier D106...first correction parameter D108..·second correction parameter D110...5 number of offset values D112 , D114...δ offset value D116···quantization offset value D302, D308...update offset parameter flag D304, D310···offset update denominator D306, D312.··offset parameter set D1501...header D1501a ···Sequence header 71 201138477 D1501b· •• Picture header D1503.. • Offset parameter D1505·· • New quantized offset value D1509.. • Quantization offset value D1511·· • New quantized offset value D1513· · • Encoding block D1515&quot; • Inverse quantized value D1517.. • Decode residual D1519&quot; • Reassembly block D1521·· • Prediction sample D1523·· • Reference image D1607·· • New quantized offset value D1609.· • Uncompressed Block D1611·· • Residual block D1613&quot; • Coefficient block D1615&quot; • Prediction sample D1619&quot; • Inverse quantized value D1621·· • Code block D1623.. • Reference image D1625·· • Reassembly block D1627·. • Recombination residual Block D1700.· • Offset parameters D1701·· • Offset update flag D1702. • Offset parameter 201138477 D1703... Offset update parameter D1704... First correction parameter D1706... Second correction parameter D1708··· Number of δ offset values D1710... (5 partial Shift value D1800...offset parameter set D1801...offset update flag D1802...offset parameter D1803...offset update parameter D1804...matrix identifier D1806...first correction parameter D1808...second correction parameter D1810...5 offset value Number D1812...δ offset value S10~S14...Steps S20~S26...Steps S400, S402, S404, S406, S408...Steps S500, S502, S504, S508, S510...Steps S600, S601, S602, S604, S606, S608 , S610, S612, S614, S616 &quot; ♦ steps S700, S7 (H, S702, S704, S706, S708, S710, S712, S716 ... steps S800, S802, S806, S808, S810, S812... steps S900, S902, S904 Steps S1100, S1102, S1104, S1106, S1108, S1110, S1112, S1114, steps S1200, S1202, S1204, S1206, S1208, S1210.··Step S1300 S1302, S1304, S1306, S1308, S1310. Step S1400, S1402, S1404, S1406, S1408.. Step S1900, S1902, S1904, S1906... Steps S2000, S2002, S2004, S2006... Steps S2100, S2102, S2104 S2106, S2108...Steps S2200, S2202, S2204, S2206, S2208, S2210...Steps S2300, S2302, S2304, S2306, S2308..·Step exlOO···Content supply system exlOl···Internet ex 102...Internet Internet service provider exl03···Streaming server exl04._·Telephone network exl06, exl07, exl08, exl09, exllO... base station exm···computer

Exl 12---PDA exll3, exll6···camera exll4···with camera digital mobile phone (mobile phone) exll5...game machine exl 17···microphone ex200...digital playback system 74 201138477 ex201···playing bureau ex202... Play satellite (satellite) ex203···cables ex204, ex205, ex601·. antenna ex210...car ex211...car navigation (Car Navi) ex212...reproduction device ex213, ex219···monitor ex214, ex215, ex216, ex607... Recording medium ex217...Turner (STB) ex218...Reader/recorder ex220...Remote controller ex230...Information magnetic execution ex23 Bu·· Recording block ex232... Inner area ex233... Data recording area ex234... Peripheral area ex300 ··············································································· Processing unit 75 201138477 6乂307...°Hedgehog eight ex308, ex602.&quot; Display unit ex309···output unit ex310, ex501·.. control unit ex311, ex505, ex710...power supply circuit unit ex312...operation input unit X313...bridge ex314, ex 606···Slot part ex315...driver ex316...data machine ex317···face ex318, ex319, ex320, ex321, ex404, ex508...buffer ex400...information reproduction/recording unit ex401...optical head ex402...modulation recording Part ex403...reproduction demodulation unit ex405...disk motor ex406...service control unit ex407...system control unit

Ex500---LSI

Ex502 ...CPU ex503...memory controller ex504...stream controller

Ex506...Streaming I/O 76 201138477

Ex509.&quot;AVI/O ex510··· busbar ex603···camera unit ex604···operation key ex605···voice input unit ex608···voice output unit ex701···received transmission circuit unit ex702.&quot LCD control unit ex703...camera interface (camera I/F unit) ex704...operation input control unit ex705···voice processing unit ex706···modulation demodulation circuit unit ex707...recording/reproduction unit ex708...demultiplexing unit Ex709···Image decoding unit ex711··· main control unit ex712...image encoding unit ex713···synchronous bus bar 77

Claims (1)

201138477 VII. Patent application scope: 1. An image decoding method for decoding a coded image contained in a coded stream, which has the following processing: parsing out the header included in the code stream Deviating the update flag; determining whether the parsed update flag is displayed by displaying the previously set value; and determining that the parsed offset update flag is displaying the preset value, Calculating a quantization offset matrix applicable to the encoded image by using an offset update parameter included in the foregoing header; performing inverse quantization on the encoded image by using the calculated quantization offset matrix, thereby decoding the foregoing Encode the image. 2. The image decoding method according to claim 1, wherein the header includes an update type identifier indicating an update method of the quantization offset matrix, and the header is parsed when the quantization offset matrix is calculated. The update type identifier is included, and the quantization offset matrix is calculated in accordance with an update method indicated by the updated update type identifier. 3. The image decoding method according to claim 1, wherein when calculating the quantization offset matrix, the offset update parameter is applied to another quantization offset matrix stored in the memory, thereby calculating the object The quantization offset matrix is calculated. 78 201138477 4. The image decoding method according to claim 3, wherein when calculating the quantization offset matrix, the last calculated quantization offset matrix is used as the other quantization offset matrix, and the offset is applied. Update parameters. 5. The image decoding method of claim 3, wherein the header includes a matrix identifier for specifying the other quantization offset matrix, and when the quantization offset matrix is calculated, the header is parsed The matrix identifier is included, and the other quantization offset matrix identified by the parsed matrix identifier is retrieved from the memory. 6. The image decoding method according to claim 3, wherein the header includes first and second correction parameters and a 5 offset (Delta offset) value, wherein the (5 offset value corresponds to the calculation target When calculating each of the elements included in the offset matrix, when calculating the quantization offset matrix, the first and second coefficients are calculated based on the first and second correction parameters included in the header, and Each of the known quantized offset values, which is an element of the other quantization offset matrix, multiplies the first coefficient by the known quantized offset value, thereby calculating a first value, and multiplying the second coefficient by the known quantized offset value The difference between the fixed value and the fixed value is used to calculate the second value, and the first value and the second value are added to calculate a predicted quantized offset value, and the predicted quantized offset value is added before the calculation target. 201138477 describes the aforementioned δ offset value corresponding to the element of the quantization offset matrix, thereby calculating the quantization offset value as the aforementioned element. 7. The image decoding method according to claim 6, wherein the foregoing header is further Contains offset update denominator When calculating the first coefficient, the first coefficient is calculated by adding the offset update denominator to the third correction parameter and dividing the value obtained by the addition by the offset update denominator. The image decoding method according to claim 7, wherein when the second coefficient is calculated, the second coefficient is calculated by dividing the second correction parameter by (4) the offset update denominator. The method of decoding the sixth item of the range, wherein the fixed value is a value set in advance for the quantization offset matrix of the calculation target. [1] The image decoding method according to claim 6, wherein the fixed value is an element. And a quantization offset value included in the other quantization offset matrix and located at a position corresponding to the DC component. η. The image decoding method according to claim 6, wherein the fixed value is included as an element. The average value of the complex quantized offset values of the other quantized offset matrices. The image decoding method of claim [0002], wherein the offset update parameter includes a complex a value_value indicating a _ difference of the complex quantized offset value, wherein the complex quantized offset value is an element of the quantization offset matrix of the object, and the quantization offset moment (4) is calculated, and the following processing is repeated : Calculating the new quantized offset value by adding the offset value to the offset value of the aforementioned complex (four) offset value, and calculating the new quantized offset value. The image decoding method of claim 1, wherein the offset update parameter includes only one quantization offset value as a single quantization offset value, wherein the quantization offset value is an element of the quantization offset matrix of the object; When calculating the quantization offset matrix, the single quantization offset value included in the offset update parameter is set for all elements included in the quantization offset matrix to be calculated. 14. An image encoding method for encoding image data, thereby generating a coded streamer having the following processing: determining whether a quantization offset matrix should be updated; an offset update flag showing a result of the foregoing determination Writing a header of the foregoing encoded stream; when it is determined that the update is to be performed in the foregoing determination, calculating an offset update parameter according to the updated new quantization offset matrix; and writing the offset update parameter to the header; The aforementioned image data is subjected to quantization using the aforementioned new quantization offset matrix, whereby the aforementioned image data is encoded. 15. An image decoding apparatus for displaying a coded image included in a coded stream, comprising: a flag analysis unit that analyzes a header included in the code stream; The offset update flag is used by the determination unit to determine whether the offset update flag is 81 or not, and the predetermined offset value is determined. The quantized offset matrix calculation unit determines that the offset is calculated. When the offset update flag is a value set in advance, the offset update parameter included in the header is used to calculate a quantization offset matrix applicable to the coded image; and an inverse quantization decoding unit is provided. And performing inverse quantization on the encoded image using the calculated quantization offset matrix, thereby decoding the encoded image. 16. An image encoding apparatus that encodes image data to generate a coded streamer, comprising: a determining unit that determines whether a quantization offset matrix should be updated; and a flag writing unit An offset update flag indicating the result of the determination is written in the header of the encoded stream; and an offset update parameter calculation unit is determined based on the updated new quantization offset matrix when it is determined that the update is to be performed in the determination. Calculating the offset update parameter; the parameter writing unit writes the offset update parameter to the header; and the quantization coding unit performs quantization on the image data by using the new quantization offset matrix The aforementioned image data is encoded. 17. A program for causing a computer to perform a process of decoding a coded image for decoding a coded image contained in a coded stream, the process comprising: parsing the encoded stream The offset 82 201138477 update flag in the header; determining whether the parsed update flag is displayed in advance or not; and determining that the offset update flag is displayed When the value is set in advance, the quantization offset parameter applicable to the encoded image is calculated using the offset update parameter included in the header; and the calculated quantization offset matrix is used to perform the encoded image on the encoded image. Inverse quantization, thereby decoding the aforementioned encoded image. 18. A program for causing a computer to perform image processing to encode image data to generate a code stream, the process comprising: determining whether a quantization offset matrix should be updated; An offset update flag indicating the result of the foregoing determination is written into the header of the encoded stream; when it is determined that the update is to be performed in the foregoing determination, the offset update parameter is calculated according to the updated new quantization offset matrix; The offset update parameter is written to the aforementioned header; the aforementioned image data is quantized using the aforementioned new quantized offset matrix, whereby the aforementioned image data is encoded. 19. An integrated circuit for displaying a coded image included in a coded stream, comprising: a flag analysis unit that parses an offset update in a header included in the coded stream; The flag indicating unit determines whether the offset update flag is 83 or not, and determines the value to be set in advance. The quantization offset matrix calculating unit determines that the offset is determined. When the update flag is used to display the value set in advance, the offset update parameter included in the header is used to calculate the quantization offset matrix applicable to the coded image; and the inverse quantization decoding unit is used. The quantization offset matrix is calculated, and inverse quantization is performed on the coded image, thereby decoding the coded image. 20. An integrated circuit for encoding an image data to generate a coded streamer, comprising: a determination unit that determines whether the quantization offset matrix should be updated; and a flag writing unit that displays The offset update flag of the result of the determination is written in the header of the encoded stream; and the offset update parameter calculation unit calculates the updated quantization offset matrix based on the updated new quantization offset matrix when it is determined that the update is to be performed in the determination. a parameter of the offset update parameter; the parameter writing unit writes the offset update parameter to the header; and the quantization coding unit performs quantization on the image data by using the new quantization offset matrix, thereby The aforementioned image data is encoded. 84