WO2010032934A2 - B-변환을 위한 부호화 방법 및 장치와 그를 위한 부호화 데이터 - Google Patents
B-변환을 위한 부호화 방법 및 장치와 그를 위한 부호화 데이터 Download PDFInfo
- Publication number
- WO2010032934A2 WO2010032934A2 PCT/KR2009/005095 KR2009005095W WO2010032934A2 WO 2010032934 A2 WO2010032934 A2 WO 2010032934A2 KR 2009005095 W KR2009005095 W KR 2009005095W WO 2010032934 A2 WO2010032934 A2 WO 2010032934A2
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- binary
- encoding
- context
- symbol
- data
- Prior art date
Links
Images
Classifications
-
- H—ELECTRICITY
- H03—ELECTRONIC CIRCUITRY
- H03M—CODING; DECODING; CODE CONVERSION IN GENERAL
- H03M7/00—Conversion of a code where information is represented by a given sequence or number of digits to a code where the same, similar or subset of information is represented by a different sequence or number of digits
- H03M7/30—Compression; Expansion; Suppression of unnecessary data, e.g. redundancy reduction
Definitions
- the present invention relates to an encoding method and apparatus for B-transformation and encoded data therefor. More particularly, the present invention relates to a method and apparatus for encoding and compressing data in order to effectively use a storage medium and a communication medium in the compression technology field.
- Techniques for compressing random data include Huffman Coding, Arithmetic Coding, and Run-Length Coding based data compression techniques.
- entropy was defined by CE Shannon, and the lower limit of the efficient symbol coding based on the stochastic model was proposed. Were proposed. These coding techniques have a problem that higher coding efficiency can be obtained when the theoretical probability of entropy is skewed to a few symbols, so that coding efficiency of a uniform distribution cannot be obtained.
- a nonlinear B-Transform technique has been proposed in order to increase the coding efficiency for data having a probability model of a uniform distribution with probability not biased in a few symbols.
- the proposed nonlinear B-transformation technique only mathematically derives the upper limit of compression of uniform distribution data, and the method of efficiently encoding binary symbols ('0' and '1') generated by nonlinear B-transformation
- a coding scheme for this has not been proposed yet, which causes a problem in encoding. Therefore, there is a need for development of a technique for efficiently encoding binary symbols generated due to nonlinear B-transformation.
- the present invention in compressing data, efficiently compresses data by uniformly distributing random data of a uniform distribution into binary symbols, analyzing the context, and encoding based on the analyzed context.
- the main purpose of the binary representation generated by B-converting data is to increase the coding efficiency by reducing the number of bits by using the characteristics of the B-transform and thereby to improve the compression performance.
- the present invention provides an apparatus for encoding input data, comprising: a B-transformer for generating a binary representation by B-transforming the input data; And an encoder which removes one or more binary symbols from the binary representation and scans the binary representation from which the one or more binary symbols have been removed to generate a bit string.
- a method of encoding input data comprising: a B-transforming step of B-transforming the input data to generate a binary representation; And an encoding step of removing one or more binary symbols from the binary representation and scanning a binary representation from which one or more binary symbols have been removed to generate a bit string.
- a data generated by an encoding apparatus by encoding input data comprising: a step number field including a number of B-transform steps; And a data field comprising a bit string in which one or more binary symbols are removed and scanned from the binary representation generated by the B-transformation of the input data by the number of B-transform steps.
- the context analysis for determining the context for the samples by analyzing the correlation between the samples of the data is converted to a binary symbol represented by the input data part;
- an encoder configured to generate a bit string by encoding samples according to the determined context.
- the context analysis for determining the context for the samples by analyzing the correlation between the samples of the data converted to the binary symbol represented by the input data step; And an encoding step of generating a bit string by encoding samples according to the determined context.
- the apparatus for analyzing a context for encoding selecting a sample to determine the context in the table representing the samples represented by the binary symbol uniform distribution data, the right side of the selected sample in the table
- a method of analyzing a context for encoding comprising: selecting a sample for determining a context from a table representing samples represented by binary symbols with uniform distribution data; Detecting a binary symbol of a sample located at the upper right of the selected sample in the table; And a context determination step of determining the detected binary symbol as the context of the selected sample.
- the context is determined more accurately by determining the context using the characteristic that the diagonal samples among the samples of the binary representation generated by non-linear B-conversion of uniform distribution data are correlated. It is possible to accurately update the probability table, thereby compressing uniform distribution data more efficiently, and encoding the binary representation generated after the B-transformation by using the characteristics of the B-transformation. Since not only can improve the compression performance, but also a separate operation process is not required, the implementation of the encoding apparatus can be simplified.
- FIG. 1 is a block diagram schematically illustrating a configuration of an encoding apparatus according to a first embodiment of the present invention
- FIG. 2 is an exemplary diagram showing properties used for B-transforming and encoding input data according to the first embodiment of the present invention
- FIG. 3 is an exemplary diagram for explaining a process of generating a binary representation by B-converting input data according to the first embodiment of the present invention
- FIG. 4 is an exemplary view showing a table in the form of a property by applying property 1 to a binary representation occurring after B-conversion according to the first embodiment of the present invention
- FIG 5 is an exemplary view showing a table in the form of a property by applying property 2 to a binary representation generated after B-conversion according to the first embodiment of the present invention.
- FIG. 6 is an exemplary view illustrating a property of applying property 2 to a binary representation generated after B-transformation and scanning in a diagonal direction according to the first embodiment of the present invention
- FIG. 7 is an exemplary diagram showing the number of bits required for encoding when scanning a binary representation after B-transformation in a diagonal direction according to the first embodiment of the present invention
- FIG. 8 is an exemplary diagram illustrating a binary representation for encoding the number of '1' reads when scanning in a diagonal direction according to the first embodiment of the present invention
- 9 is an exemplary view showing the number of bits that can be reduced by using the number of steps according to the first embodiment of the present invention.
- FIG. 10 is an exemplary view showing a condition that can reduce a bit when encoding a position of an alignment identifier requiring a maximum number of bits according to a first embodiment of the present invention
- FIG. 11 is an exemplary diagram showing a table in the form of a table by applying property 3 to a binary representation generated after B-conversion according to the first embodiment of the present invention
- FIG. 12 is an exemplary view illustrating a property of applying property 3 to a binary representation generated after B-transformation and scanning in a diagonal direction according to the first embodiment of the present invention
- FIG. 13 is an exemplary diagram showing a structure of encoded data obtained by encoding a binary representation occurring after B-transform according to the first embodiment of the present invention
- 15 is a block diagram schematically illustrating a configuration of a context-based encoding device according to a second embodiment of the present invention.
- 16 is an exemplary diagram illustrating a table showing a binary representation generated through a nonlinear B-transformation.
- 17 is a flowchart illustrating a context-based encoding method according to a second embodiment of the present invention.
- FIG. 18 is a flowchart illustrating a process of determining a context according to a second embodiment of the present invention.
- FIG. 1 is a block diagram schematically illustrating a configuration of an encoding apparatus according to a first embodiment of the present invention.
- the encoding apparatus 100 may include a B-transformer 110 and an encoder 120.
- the encoding apparatus 100 may be a personal computer (PC), a notebook computer, a personal digital assistant (PDA), a portable multimedia player (PMP), or a PlayStation Portable (PSP).
- PC personal computer
- PDA personal digital assistant
- PMP portable multimedia player
- PGP PlayStation Portable
- a device for encoding various data such as a mobile communication terminal, a communication device such as a communication modem for performing communication with various devices or a wired / wireless communication network, and storing various programs and data for encoding data.
- Means a variety of devices including a memory for, a microprocessor for executing and operating a program.
- the B-transformer 110 generates a binary representation by B-converting the input data.
- the input data may be uniform distribution data
- the B-transformation is a technique of converting the uniform distribution data into a binary representation in order to efficiently encode the uniform distribution data.
- the encoder 120 removes one or more binary symbols from the binary representation converted by the B-transformer 110, and generates a bit string by scanning the binary representation from which one or more binary symbols have been removed.
- the encoder 120 may remove the one or more binary symbols from the binary representation by analyzing correlations between the binary symbols of the binary representation, and the raster scan direction or the diagonal of the binary representation from which one or more binary symbols have been removed.
- the bit stream can be generated by scanning in the direction.
- the one or more binary symbols removed from the binary representation may be binary symbols from the last binary symbol of the binary representation generated by performing the i-th B-transformation on the symbol string of the input data to the i-th binary symbol.
- the one or more binary symbols removed from the binary representation are i + 1 th B-converted when the value of the j th binary symbol of the binary representation generated by the i th B-conversion of the symbol string of the input data is '0'. It may be a j-1 th binary symbol of the generated binary representation.
- the encoder 120 may scan the binary representation in the 135 ° direction as the diagonal direction.
- the one or more binary symbols removed from the binary representation may be i-1 th B-converted when the value of the j th binary symbol of the binary representation generated by i-th B-converting the symbol string of the input data is '1'. It may be a j + 1 th binary symbol of the generated binary representation.
- the encoder 120 may scan the binary representation in the ⁇ 45 ° direction as the diagonal direction.
- the encoder 120 may generate a bit string by encoding the number of '1's to be read when scanning in the diagonal direction.
- the number of bits required to encode the number of '1's is determined, and the number of' 1's can be encoded by representing the number of '1's in binary numbers according to the determined number of bits.
- the most significant bit of the binary number representing the number is '1', only the most significant bit may be included in the bit string.
- the encoder 120 determines the number of bits required to encode the number of '1' when generating a bit string by encoding the number of '1's to be read when scanning in the diagonal direction, and determines the threshold value.
- a sort identifier having a number greater than or equal to '1' is encoded with the determined number of bits, and a sort identifier having a maximum number of bits from an sort identifier having a number greater than or equal to the number of '1's is determined as the determined number of bits.
- the number of bits obtained by subtracting one bit from the determined number of bits may be encoded, but the position of the alignment identifier requiring the maximum number of bits may be further encoded.
- the encoder 120 determines the number of bits required to encode the number of '1's, encodes up to an alignment identifier having a number of' 1's greater than or equal to the threshold value with the determined number of bits, and the threshold value. From the alignment identifier next to the alignment identifier having the number of greater than or equal to '1', it may be encoded by the number of bits obtained by subtracting one bit from the determined number of bits. In this case, when the sorting identifier having the number of '1' greater than or equal to the threshold value is located near or at the end of the order of all sort identifiers, there is a problem that the number of bits that can be reduced is small. And the number of '1's may be encoded by adaptively determining one of the reverse directions.
- the encoder may determine the smaller of the binary symbols of the binary representation from which one or more binary symbols have been removed as the most probable symbol (MPS).
- MPS most probable symbol
- FIG. 2 is an exemplary diagram showing properties used for B-transforming and encoding input data according to the first embodiment of the present invention.
- Bubble-Transform is a set of binary symbols that occur after consecutive M X B-transformations for the occurrence symbol sequence X This is defined as a binary representation of X. set Through these results, it can be seen that there are properties of properties 1 to 3 as shown in FIG. 1, and these properties can effectively encode a binary representation of X occurring after B-transformation.
- Equation 1 n, i and j are each an arbitrary integer.
- property 1 indicates that whenever a B-transformation is performed on a symbol string X of binary data, the number of binary symbols of a binary representation can be reduced.
- property 1 means that in the binary representation generated by performing the i-th B-transform on the symbol string X of the input data, the binary symbol from the last binary symbol of the binary representation to the binary symbol located in the i th is always' Since it has a value of 0 ', which means that encoding can be omitted, using property 1 can reduce the bits of data to be encoded.
- the two properties is the symbol sequence of the binary data X i-th B- is the j-th binary symbol of the converted binary representation of a '0', i 1 + j-1-th binary symbol of the second B- converted binary representation is a "0"
- property 2 can reduce the bit of data to be encoded.
- Property 3 is that if the jth order binary symbol of the i-th B-converted binary representation of the input data symbol X of the input data is '1', the j + 1th binary symbol of the i-1th B-transformed binary representation is '1'. 1 '. That is, when the binary symbols of the binary representation of B-transformed symbol strings of the input data are represented in a table according to the number of steps of the B-transformation, the binary symbol at an arbitrary position in the table is '1'. The other binary symbols located diagonally upwards from the binary symbol position are all '1'.
- property 3 can reduce the bits of data to be encoded.
- FIG. 3 is an exemplary diagram for explaining a process of generating a binary representation by B-converting input data according to the first embodiment of the present invention.
- Population S ⁇ 1,2,... , 16 ⁇ is a symbol column of the input data is generated to be uniformly distributed within the X ⁇ 2,5,3,1,11,12,15,8,7,6,16,4,9,10,14,13 ⁇ , It is possible to B-convert the symbol string X , as shown in FIG.
- X is a symbol string of input data
- X (M1) is a symbol string of the rearranged input data after the first B-transformation
- X (M2) is a symbol string of the rearranged input data after the second B-transformation.
- Heat, Is a binary representation generated after the first B-transformation Represents the binary representation produced after the second transform.
- the number of binary symbols of the binary representation generated by each B-transformation is one smaller than the number of symbol strings of the input data.
- the B-conversion process is performed by comparing each symbol of the symbol string X of the input data with the next symbol in order, and if the two symbols are arranged in ascending order, the binary symbol '0' If the two symbols are not sorted in ascending order, the binary symbol '1' is output. If two symbols are not sorted in ascending order and the binary symbol '1' is outputted, the two symbols are sorted in ascending order by changing the sorting order of the two symbols. And the next symbol.
- FIG. 4 is an exemplary diagram in the form of a table by applying property 1 to a binary representation generated after B-conversion according to the first embodiment of the present invention.
- the binary representation resulting from the B-transformation in each step is represented in the form of a table by applying the property 1, as shown in FIG. 4.
- Can be represented. 4 the total number of binary symbols of the binary representation generated after the first B-transformation is 15, and the total number of binary symbols of the binary representation generated after the second B-transformation is 14, the last 8 The total number of binary symbols of the binary representation generated after the first B-transformation is eight.
- the total number of binary symbols in the binary representation is 15 regardless of the number of steps of the B-transform, but as the number of steps of the B-transform increases, the binary symbols of the binary representation generated are reversed from the last binary symbol in the reverse order.
- the last binary symbols of the binary representation can be predicted according to the number of steps of the B-transformation. Therefore, some of the binary symbols that can occur according to the B-transformation can be predicted by the property 1, so that the predictable binary symbols can be restored without encoding them, thereby reducing the number of binary symbols to be encoded. The number of bits can be reduced.
- d 1 to d 15 represent respective binary symbols of a binary representation generated by B-converting a symbol string of input data, and may be generalized to d n .
- FIG. 5 is an exemplary diagram in the form of a table applying property 2 to a binary representation generated after B-conversion according to the first embodiment of the present invention.
- property 2 When property 2 is applied to the binary data shown in FIG. 4, it can be expressed as shown in FIG. 5. According to property 2, when a binary symbol at an arbitrary position in the table is '0', it can be predicted that all other binary symbols located diagonally downward and left at the position of the corresponding binary symbol are all '0'. That is, among the binary symbols located on the same line in the diagonal direction (for example, 135 °) from the upper right to the lower left in the table of FIG. 4, the binary symbols after the binary symbol having a value of '0' for the first time are all ' It can be seen that 0 '.
- FIG. 6 is an exemplary view illustrating a property of applying property 2 to a binary representation generated after B-transformation and scanning in a diagonal direction according to the first embodiment of the present invention.
- the scanned bits can be scanned in a diagonal direction of 135 [deg.] And transmitted as bit strings, and reverse scanning the transmitted bit strings in the same manner can restore the original binary representation.
- the bit string of the binary representation shown in FIG. 5 when scanning in the raster scan direction, the bit string of the binary representation shown in FIG. 5 may be ⁇ 01100011101111101111111111111111010111110011000111 ⁇ . 6, when scanning diagonally, the bit string of the binary representation shown in FIG. 5 may be ⁇ 01011100011101111011111001111111111110111101101110 ⁇ .
- the decoding apparatus may inversely scan the received bit string according to the scan direction to restore all symbols of the binary representation.
- n is, if they have, or is less than M X, the binary symbols that is, the maximum number of bits which may have in the position of d n (max_ bit (d n)) is n Games It can be seen.
- M n is greater than X, the maximum number (max_ bit (d n)) of bits which may have in the position of d n can know the M X Games.
- max_ bit ( k ) represents the maximum bit that can be generated at the k position.
- Equation 3 when the binary representation shown in FIG. 5 is scanned in the diagonal direction shown in FIG. 6, a rule such as Equation 3 can be further found. That is, when a binary symbol having a value of '0' occurs in the process of scanning and reading in a diagonal direction according to property 2, it is not necessary to read and encode the next binary symbol located at the next position in the diagonal direction. The reason that the encoding apparatus does not need to read and encode the next binary symbol is that if the decoding apparatus knows the number of binary symbols having a value of '1' among the binary symbols read when scanning diagonally from the position of d n . This means that all binary representations shown in FIG. 4 can be restored.
- the encoding apparatus encodes only the number of binary symbols having a value of '1' among binary symbols read when scanning in a diagonal direction, and at this time, the number of binary symbols having a value of '1' for each position of d n ( bit ( d n )) may be calculated by Equation 3.
- Ceil ( x ) is the largest integer among numbers less than or equal to x .
- FIG. 7 is an exemplary diagram illustrating the number of bits required for encoding a binary representation after B-transformation in the diagonal direction according to the first embodiment of the present invention.
- Equation 3 Applying Equation 3 to, for example, the symbol string of the input data presented (i.e., when N is 16), encoding the number of binary symbols having a value of '1' at the position of d n (1 ⁇ n ⁇ 16).
- the number of bits required to do this can be represented as shown in FIG. Referring to FIG. 7, the total bits required for encoding are 49 bits.
- Equation 4 the left part of the right term has a range of n 1 ⁇ n ⁇
- Equation 5 shows an equation for calculating the total number of bits when encoding using entropy encoding, various equations may be derived according to the encoding technique.
- a gain Gain may be calculated through Equation 6.
- FIG. 8 is an exemplary diagram illustrating a binary representation for encoding the number of '1's to be read when scanning in a diagonal direction according to the first embodiment of the present invention.
- the encoding may be performed in the same order as in FIG. 8. That is, after calculating the number of bits necessary for encoding the binary symbol value '1' of each d n of the binary data shown in FIG. 2, and calculating the number of '1' when scanning diagonally for each d n When the number of '1's is converted to binary numbers and represented, a new binary representation is generated as shown in FIG. 8.
- Equation 3 when the number of bits required to encode the number of '1' at the position of d 1 is calculated using Equation 3, the number of bits is one, and when scanning diagonally at the position of d 1, the number of '1' Since the number is 0, a binary representation for encoding the same is represented by '1'.
- the number of bits required to encode the number of '1' at the position of d 2 is calculated by using Equation 3, and the number of '1' is 1 when scanning diagonally at the position of d 2 . Therefore, the binary representation for encoding this is represented by '01'.
- the number of bits required to encode the number of '1' at the position of d 7 is calculated by using Equation 3, and the number of '1' is 3 when scanning diagonally at the position of d 7 . Therefore, the binary representation for encoding this is represented by '011'.
- Encoding d 1 to d 15 in this manner produces a binary representation as shown in FIG. 8. Therefore, when encoded as described above, the symbol string X generated as uniform distribution data may be encoded and represented as a bit string ⁇ 1011100000000001101000101000010000100010000100011 ⁇ .
- the number of '1's is calculated after scanning diagonally for each d n , the number of'1's are converted to binary numbers, encoded, and then described later using information of M X. Likewise, you can remove the bit further. That is, M X is the number of steps of the B-transformation, and since the number of 1s in each d n cannot be greater than M X , Bits of some of the bits from d to N may be removed.
- M X is 8 in the binary representation shown in FIG. 4, the binary representation is '1000'.
- d 8 to d 15 may have a value ranging from '0000' to '1000' since they should be less than or equal to M X.
- d 11 since the number of '1' is 8, it may be represented as '1000'. Since the maximum value of M X is '1000', when the most significant bit (MSB) becomes '1', the remaining 3 bits except for the most significant bit of the total 4 bits must be '000'.
- the decoding apparatus when the decoding apparatus receives the most significant bit, it is possible to know the number of original '1's without receiving the remaining three bits'000' and to restore the binary symbols accordingly. Accordingly, the encoding apparatus does not need to transmit the remaining 3 bits '000', and thus can further compress 3 bits.
- 9 is an exemplary view showing the number of bits that can be reduced by using the number of steps according to the first embodiment of the present invention.
- the above-described example can be generalized and summed up when N is 16 and M X is equal to or larger than 8, as shown in FIG. 9. That is, in the case of d n where the number of '1' is M X , as shown in FIG. 9, the bit can be reduced.
- d n having a number of '1' can be represented as '11' by omitting '00' from the original binary representation '1001', thereby reducing two bits. have.
- the number of 1s in each d n may not be greater than M X , so that a bit may be reduced from a maximum of 3 bits to 0 bits per one alignment identifier d n .
- Such a method is not limited to the example shown in FIG. It can be applied to all symbols in the positions of the N d.
- the symbols of the binary representation representing the number of '1's in binary may be omitted as described below. That is, when N is small (e.g., N ⁇ 64), Of d from to N Greater than or equal to d N is less. In the example shown in FIG. 7, when the number of '1's between d 8 and d 15 is greater than or equal to 8, there is one case of d 11 . Therefore, the number of '1's can be sufficiently represented by 3 bits except d 11 among d 8 to d 15 . However, the decoding apparatus must encode information that can be represented by 3 bits except d 11 among the actual d 8 to d 15. There are two methods.
- the first method is to encode the position of d 11 .
- d 8 to d 15 are a total of eight, all the positions can be represented by three bits. Therefore, when d 8 is represented as '000' as a reference point, d 11 may indicate its position as '100'.
- the decoding apparatus can decode d 8 , d 9 , d 10 , d 12 , d 13 , d 14 , and d 15 using only 3 bits and decode only d 11 using 4 bits. In this case, an additional three bits are needed to indicate the position, but the total number of four bits can be reduced by reducing seven bits.
- d n is greater than or equal to 8 but not more than one, the number of bits required for encoding the actual position may be greater than the number of bits that can be reduced thereby.
- the number of '1's is encoded by performing a B-transformation on the length N of the generated symbol string X and scanning the binary data, i.e., from d 1 to d N-1 diagonally, the same number of bits.
- There are intervals using i.e., alignment identifiers ( d n ) encoded using the same number of bits
- T can be calculated using Equation 7.
- Equation 8 Equation 8
- FIG. 10 is an exemplary view showing a condition in which a bit can be reduced when encoding a position of an alignment identifier requiring a maximum number of bits according to the first embodiment of the present invention.
- Fig. 10 shows Equation 8 summarized for various Ks and n obtained by satisfying Equation 8 is shown. 10, the larger the size n of which is greater the more K allows, n is the greater number of bits can be reduced The more smaller. However, the additional information is required to be transmitted to the number of n is also encoded.
- the second way is to have a number of '1's Greater than or equal d n Send only the number of and the number of '1' Greater than or equal d n All bits are transmitted until the data is transmitted, and the next bit is transmitted one bit less.
- d 8 To d 15 Suppose we encode d 8 To d 15 medium Greater than or equal d n silver d 11 Since only one exists, the number is one. After encoding '1', real data is transmitted in order. In other words, d 8 Is '0100', d 9 Is '0101', d 10 Is '0000', d 11 Transmits to '1000'.
- d 11 Is a number greater than or equal to '8' ( N Is 16, Is '8'), d 12 It can be seen that the number is smaller than '7', and only 3 bits can be encoded. therefore, d 12 Is '100', d 13 Is '100', d 14 Is '010', d 15 Is encoded as '011', and can reduce a total of 3 bits.
- the problem with the second method is that the number of '1' The difference that the bit can be reduced according to the position of d n which is greater than or equal is very large. If the number of '1' If a position greater than or equal to d n is located in the coding order (transmission order), many bits can be reduced, whereas if it is located later, fewer bits can be reduced.
- backward search may be used.
- d 15 , d 14 , d 13 ,. which is encoded (or transmitted) in the order of d 8 .
- forward search means that the number of '1' For d n greater than or equal to d 8 , d 9 , d 10 ,. , Refers to encoding a sequence of d 15, backward seek is d 15, d 14, d 13 , ... , which is coded in the order of d 8 .
- Equation 9 it is assumed that the distribution of n has a uniform distribution.
- the number of bits that can reduce (Adaptive Bit) can be calculated using the equation (10).
- FIG. 11 is an exemplary diagram in the form of a table applying property 3 to a binary representation generated after B-conversion according to the first embodiment of the present invention.
- property 3 is applied to the table of FIG. 4 that illustrates the binary representation that occurs after B-conversion of the input data, it may be represented as shown in FIG. 11.
- the signals are sequentially scanned in the raster scan direction from the left of the bottom row, encoded, and transmitted, or as shown in FIG. 12, from -45 ° from the binary symbol at the upper left. In the direction of, i.e., scan diagonally, encode and transmit.
- FIG. 12 is an exemplary view illustrating a property of applying property 3 to a binary representation generated after a B-transform and scanning in a diagonal direction according to the first embodiment of the present invention.
- n is M, if X has less than or equal to the maximum number of bits which may have in the position of d n (max_ bit (d n)) is can be seen that n Games. And M n is greater than X, the maximum number (max_ bit (d n)) of bits which may have in the position of d n can know the M X Games.
- the number of '1's is equal to the number of' 0's when the binary symbols are scanned and encoded in the diagonal direction according to the property 3, compared with the case where the binary symbols are scanned and encoded in the diagonal direction according to the property 2. All properties are the same except for the changes. Therefore, the description of a method of scanning and encoding a binary symbol in a diagonal direction according to the aforementioned property 2 may be similarly applied to property 3.
- the maximum possible symbol reverse flag (RF) is determined.
- the conventional method analyzes the distribution of binary symbols having '0' and '1' in the binary representation after the B-transformation to determine a highly distributed symbol as the maximum possible symbol. That is, when there are many '0's, the maximum possible symbol is determined as' 0', and when there are many '1', the maximum possible symbol is determined as' 1 '.
- the maximum possible symbol is not determined according to the distributions of '0' and '1', and after the bit is reduced by encoding using the properties 2 and 3, the number of less appearing between '0' and '1' appears. Determine the bit as the maximum possible symbol.
- the reverse flag also determines the direction in which the output bit becomes the minimum after performing both the forward search and the reverse search. For example, if the bit output after the forward search is the minimum, it may be determined as '0', and if the bit output after the backward search is the minimum, it may be determined as '1'.
- the encoder 120 of the encoding apparatus 100 encodes the input data to generate data.
- the generated encoded data includes a step number field including the number of B-transform steps and the input data equal to the number of B-transform steps.
- One or more binary symbols may be removed from the binary representation generated by the B-conversion to include a data field including the bit string generated by scanning.
- the encoded data may further include a maximum possible symbol field including a flag indicating a maximum possible symbol, which may be the smaller of binary symbols of a binary representation from which one or more binary symbols have been removed.
- the encoded data may further include an inverse flag field including a flag indicating whether to search backward.
- FIG. 13 is an exemplary diagram showing a structure of encoded data obtained by encoding a binary representation occurring after B-transform according to the first embodiment of the present invention.
- M X denotes the number of steps of the B-transformation
- the maximum possible symbol represents a binary symbol 0 and 1 frequently occurring binary symbols occurring after the B-transformation, thereby representing an encoding order of binary data.
- the maximum possible symbol is not determined according to the distribution of the binary symbol, but after the bit is reduced by applying the properties 2 and 3 according to the first embodiment of the present invention, a symbol that rarely occurs is determined as the maximum possible symbol. .
- the reverse flag indicates a direction in which the output bit becomes the minimum after performing both the forward search and the reverse search.
- a field that can be compressed by encoding using the properties 1 to 3 is a data field.
- FIG. 14 is a flowchart for explaining an encoding method according to a first embodiment of the present invention.
- the encoding apparatus 100 generates a binary representation by B-converting the symbol string of the input data (S1410), removes one or more binary symbols from the binary representation by applying the properties 1 to 3 to the binary representation (S1420), Each binary symbol of the binary representation from which one or more binary symbols have been removed is scanned in a raster scan direction or a diagonal direction to generate a bit string (S1430).
- the bit string generated as described above may be generated as encoded data by combining with the number of steps or with one or more of the maximum number of possible symbols and the inverse flag and transmitted to the decoding apparatus.
- the encoding apparatus 100 when the encoding apparatus 100 generates a bit string by scanning each binary symbol of a binary representation from which one or more binary symbols have been removed in a diagonal direction, the number of bits using the various techniques described above with reference to FIGS. 4 to 11. Can be further reduced.
- 15 is a block diagram schematically illustrating a configuration of a context based encoding apparatus according to a second embodiment of the present invention.
- the context-based encoding apparatus 1500 may include a B-transformer 1510, a context analyzer 1520, and an encoder 1530.
- the context-based encoding device 1500 may be a personal computer (PC), a notebook computer, a personal digital assistant (PDA), a portable multimedia player (PMP), or a PlayStation Portable (PSP).
- Portable a sub-device or sub-software module for compressing data in a mobile communication terminal, etc., communication interface, image for performing communication with various devices, components, software modules or wired and wireless communication network It may be a variety of devices including a variety of programs for encoding the data and the memory for storing the data, a microprocessor for executing the operation and control the program.
- the B-transformer 1510 converts the input uniform distribution data into a binary symbol to generate data represented by a binary symbol and transmits the data to the context analyzer 1520.
- the binary symbol refers to a symbol having a value of '0' or '1'.
- the context analyzer 1520 analyzes the correlation between the samples of the data represented by the binary symbol by the B-transformer 1510 to determine the context for the samples and transmits the context to the encoder 1530.
- the context analyzer 1520 may analyze the correlation using a table representing the samples represented by the binary symbol.
- the nonlinear B-transform expresses the presence or absence of uncertainty between the uniform distribution data as binary symbols ('0' or '1').
- the non-linear B-transformation may express the presence or absence of uncertainty between the uniform distribution data according to the depth of the transformation.
- 16 exemplarily shows a table showing binary representations generated through nonlinear B-transformation.
- Nonlinear B-transformation of the 16 uniform distribution data produces a binary representation with a binary symbol of '0' or '1'.
- Expressing the transformed binary representation according to the depth of the B-transformation generates the table shown in FIG.
- l 1 to l 8 represent the depth of transformation (Level)
- d 1 to d 16 represent the presence or absence of uncertainty between uniform distribution data.
- the binary representation obtained through the nonlinear B-transformation is represented differently according to the presence or absence of uncertainty between the uniform distribution data and the depth of the nonlinear B-transformation. Also, it can be seen from FIG. 16 that, due to the characteristics of the nonlinear B-transform, the presence or absence of uncertainty at any position in the table correlates in the diagonal direction. That is, in FIG. 16, it can be seen that samples having a value of '1' indicating that there is uncertainty and samples diagonally adjacent to each other (particularly, samples located on the upper right side) have the same value of '1'. Accordingly, the context analyzer 1520 determines the context of the sample using this correlation.
- the context analyzer 1520 selects an arbitrary sample to determine the context from the table representing the samples represented by the binary symbol, and selects a binary symbol of a sample diagonally adjacent to the selected sample.
- the detected binary symbols can be determined as the context of the selected sample.
- the diagonally adjacent samples may be samples on the upper right of the selected sample in the table, and the selected samples may be selected in the order of the raster scan direction. Determining the context by detecting the binary symbol of the sample on the upper right side of the selected sample without detecting the binary symbols of all the diagonally adjacent samples, since the sample on the lower left of the selected sample is not yet encoded, This is because a decoder that receives and decodes a bit string cannot decode the currently selected sample based on the sample on the lower left. Accordingly, when determining the context of the selected sample, the context analyzer 1520 considers only the binary symbols of the upper right sample among the samples correlated diagonally, and does not consider the binary symbols of the lower left sample.
- the selected sample is located at coordinates ( l 2 , d 6 ), it is a binary symbol of the sample located at coordinates ( l 1 , d 7 ) on the upper right of coordinates ( l 2 , d 6 ).
- '1' becomes the context of the selected sample.
- the selected sample is located at coordinates ( l 3 , d 10 )
- '1' This is the context of the selected sample.
- the encoder 1530 generates a bit string by encoding the samples according to the context determined by the context analyzer 1520. That is, when the context of the sample is transmitted from the context analyzer 1520, the encoder 1530 updates the probability table for encoding according to the transferred context, and encodes the sample using the updated probability table.
- the B-transformer 1510 is necessarily included in the context-based encoding apparatus 1500 according to the second embodiment of the present invention, thereby generating data represented by binary symbols by B-converting the input data.
- the context-based encoding apparatus 1500 may not necessarily include the B-transformer 1510.
- the context analyzer 1520 may convert data represented by a binary symbol by B-converting the input data. Can be entered directly.
- 17 is a flowchart illustrating a context-based encoding method according to a second embodiment of the present invention.
- the context-based encoding apparatus 1500 converts the input data into data represented by a binary symbol by B-converting the data (S1710), and analyzes the correlation of the samples of the data represented by the binary symbol to each sample. Determine the context (S1720).
- the context-based encoding apparatus 1500 generates a bit string by encoding each sample according to the determined context (S1730). In this case, whenever the context of each sample is determined, the context-based encoding apparatus 1500 updates the probability table for encoding and encodes the sample using the updated probability table.
- step S1710 may not necessarily be performed. That is, the context-based encoding apparatus 1500 may omit step S1710 if the uniform distribution data is not received as input data and the uniform distribution data is B-transformed to receive data represented by a binary symbol.
- the context-based encoding apparatus 1500 may determine the context as described below with reference to FIG. 18.
- FIG. 18 is a flowchart illustrating a process of determining a context according to a second embodiment of the present invention.
- the context-based encoding apparatus 1500 may generate an arbitrary raster from the sample in the upper left of the table. One by one in order of scanning direction is selected (S1810).
- the context-based encoding apparatus 1500 detects a binary symbol of a sample on the upper right side of the selected sample (S1820). If there is no sample on the upper right, the context is not determined. Accordingly, even when the sample is encoded in step S1730 described above with reference to FIG. 17, the sample may be encoded without updating the probability table. In addition, when determining the context of the selected sample, the context-based encoding apparatus 1500 may consider only the binary symbol of the upper right sample among the samples correlated diagonally and may not consider the binary symbol of the lower left sample. .
- the context-based encoding apparatus 1500 that detects the binary symbol determines the detected binary symbol as the context of the selected sample (S1830).
- the context analyzer 1520 of the context-based encoding apparatus 1500 according to the second embodiment of the present invention may be independently implemented as an apparatus for analyzing a context for encoding according to the third embodiment of the present invention.
- the context analysis apparatus according to the third embodiment of the present invention selects a sample to determine the context from a table representing samples in which uniform distribution data is represented by a binary symbol, and selects a binary symbol of a sample located at the upper right of the selected sample in the table. The detected binary symbol can be determined as the context of the selected sample.
- the present invention can determine the context more accurately by determining the context using the characteristic that the diagonal samples among the samples of the binary representation generated by non-linear B-conversion of the uniform distribution data are correlated.
- the probability table can be updated accurately, so that the uniform distribution data can be compressed more efficiently, and the number of bits can be reduced by encoding the binary representation generated after the B-transformation using the characteristics of the B-transformation. It is possible to improve the compression performance as well as to eliminate the need for a separate operation, which is a very useful invention for generating an effect that can simplify the implementation of the encoding apparatus.
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Compression, Expansion, Code Conversion, And Decoders (AREA)
- Compression Or Coding Systems Of Tv Signals (AREA)
Abstract
Description
Claims (32)
- 입력 데이터를 부호화하는 장치에 있어서,상기 입력 데이터를 B-변환하여 이진 표현을 생성하는 B-변환부; 및상기 이진 표현에서 하나 이상의 이진 심볼을 제거하고, 상기 하나 이상의 이진 심볼이 제거된 이진 표현을 스캔하여 비트열을 생성하는 부호화부를 포함하는 것을 특징으로 하는 부호화 장치.
- 제 1 항에 있어서, 상기 부호화부는,상기 이진 표현의 이진 심볼들 간의 상관 관계를 분석하여 상기 하나 이상의 이진 심볼을 제거하는 것을 특징으로 하는 부호화 장치.
- 제 1 항에 있어서, 상기 하나 이상의 이진 심볼은,상기 입력 데이터의 심볼열을 i 번째 B-변환을 수행하여 생성되는 이진 표현의 마지막 이진 심볼로부터 상기 i 번째에 위치한 이진 심볼까지의 이진 심볼인 것을 특징으로 하는 부호화 장치.
- 제 1 항에 있어서, 상기 부호화부는,상기 하나 이상의 이진 심볼이 제거된 이진 표현을 대각선 방향의 순서로 스캔하는 것을 특징으로 하는 부호화 장치.
- 제 4 항에 있어서, 상기 하나 이상의 이진 심볼은,상기 입력 데이터의 심볼열을 i 번째 B-변환하여 생성되는 이진 표현의 j 번째 이진 심볼의 값이 '0'인 경우, i+1 번째 B-변환하여 생성되는 이진 표현의 j-1 번째 이진 심볼인 것을 특징으로 하는 부호화 장치.
- 제 5 항에 있어서, 상기 대각선 방향은,135°방향인 것을 특징으로 하는 부호화 장치.
- 제 4 항에 있어서, 상기 하나 이상의 이진 심볼은,상기 입력 데이터의 심볼열을 i 번째 B-변환하여 생성되는 이진 표현의 j 번째 이진 심볼의 값이 '1'인 경우, i-1 번째 B-변환하여 생성되는 이진 표현의 j+1 번째 이진 심볼인 것을 특징으로 하는 부호화 장치.
- 제 8 항에 있어서, 상기 대각선 방향은,-45°방향인 것을 특징으로 하는 부호화 장치.
- 제 1 항에 있어서, 상기 부호화부는,상기 대각선 방향으로 스캔할 때 읽혀지는 '1'의 개수를 부호화함으로써 상기 비트열을 생성하는 것을 특징으로 하는 부호화 장치.
- 제 9 항에 있어서, 상기 부호화부는,상기 '1'의 개수를 부호화하는 데 소요되는 비트수를 결정하고, 상기 결정된 비트수에 따라 상기 '1'의 개수를 이진수로 표현함으로써 상기 '1'의 개수를 부호화하는 것을 특징으로 하는 부호화 장치.
- 제 10 항에 있어서, 상기 부호화부는,상기 '1'의 개수를 표현한 이진수의 최상위 비트가 '1'인 경우, 상기 최상위 비트만을 상기 비트열에 포함시키는 것을 특징으로 하는 부호화 장치.
- 제 9 항에 있어서, 상기 부호화부는,상기 '1'의 개수를 부호화하는 데 소요되는 비트수를 결정하고, 임계값보다 크거나 갖은 '1'의 개수를 갖는 정렬 식별자까지는 상기 결정된 비트수로 부호화하며, 임계값보다 크거나 갖은 '1'의 개수를 갖는 정렬 식별자부터는 최대 비트수가 소요되는 정렬 식별자의 경우 상기 결정된 비트수로 부호화하고 나머지 정렬 식별자의 경우 상기 결정된 비트수에서 한 개의 비트를 뺀 비트수로 부호화하되, 상기 최대 비트수가 소요되는 정렬 식별자의 위치를 추가로 부호화하는 것을 특징으로 하는 부호화 장치.
- 제 9 항에 있어서, 상기 부호화부는,상기 '1'의 개수를 부호화하는 데 소요되는 비트수를 결정하고, 임계값보다 크거나 갖은 '1'의 개수를 갖는 정렬 식별자까지는 상기 결정된 비트수로 부호화하며, 상기 임계값보다 크거나 갖은 '1'의 개수를 갖는 정렬 식별자 다음의 정렬 식별자부터는 상기 결정된 비트수에서 한 개의 비트를 뺀 비트수로 부호화하는 것을 특징으로 하는 부호화 장치.
- 제 13 항에 있어서, 상기 부호화부는,정방향 및 역방향 중 하나의 방향을 적응적으로 결정하여 상기 '1'의 개수를 부호화하는 것을 특징으로 하는 부호화 장치.
- 제 1 항에 있어서, 상기 부호화부는,상기 하나 이상의 이진 심볼이 제거된 이진 표현의 이진 심볼들 중 더 적은 심볼을 최대 가능 심볼로서 결정하는 것을 특징으로 하는 부호화 장치.
- 제 1 항에 있어서, 상기 입력 데이터는,균일 분포 데이터인 것을 특징으로 하는 부호화 장치.
- 입력 데이터를 부호화하는 방법에 있어서,상기 입력 데이터를 B-변환하여 이진 표현을 생성하는 B-변환 단계; 및상기 이진 표현에서 하나 이상의 이진 심볼을 제거하고, 상기 하나 이상의 이진 심볼이 제거된 이진 표현을 스캔하여 비트열을 생성하는 부호화 단계를 포함하는 것을 특징으로 하는 부호화 방법.
- 부호화 장치가 입력 데이터를 부호화하여 생성하는 데이터에 있어서,B-변환 단계수를 포함하는 단계수 필드; 및상기 입력 데이터가 상기 B-변환 단계수만큼 B-변환되어 생성된 이진 표현에서 하나 이상의 이진 심볼이 제거되고 스캔되어 생성되는 비트열을 포함하는 데이터 필드를 포함하는 것을 특징으로 하는 부호화 데이터.
- 제 18 항에 있어서, 상기 부호화 데이터는,최대 가능 심볼을 나타내는 플래그를 포함하는 최대 가능 심볼 필드를 추가로 포함하는 것을 특징으로 하는 부호화 데이터
- 제 19 항에 있어서, 상기 최대 가능 심볼은,상기 하나 이상의 이진 심볼이 제거된 이진 표현의 이진 심볼들 중 더 적은 심볼인 것을 특징으로 하는 부호화 데이터.
- 제 18 항에 있어서, 상기 부호화 데이터는,역방향 탐색 여부를 나타내는 플래그를 포함하는 역 플래그 필드를 추가로 포함하는 것을 특징으로 하는 부호화 데이터.
- 입력 데이터를 부호화하는 장치에 있어서,상기 입력 데이터가 변환되어 이진 심볼로 표현된 데이터의 샘플들 간의 상관 관계를 분석하여 상기 샘플들에 대한 문맥을 결정하는 문맥 분석부; 및상기 결정된 문맥에 따라 상기 샘플들을 부호화함으로써 비트열을 생성하는 부호화부를 포함하는 것을 특징으로 하는 문맥 기반 부호화 장치.
- 제 22 항에 있어서, 상기 문맥 분석부는,상기 이진 심볼로 표현된 샘플들을 나타내는 테이블을 이용하여 상기 상관 관계를 분석하는 것을 특징으로 하는 문맥 기반 부호화 장치.
- 제 23 항에 있어서, 상기 문맥 분석부는,상기 테이블에서 상기 문맥을 결정할 샘플을 선택하고, 상기 테이블에서 상기 선택된 샘플과 대각선 방향으로 인접한 샘플의 이진 심볼을 검출하며, 상기 검출된 이진 심볼을 상기 선택된 샘플의 문맥으로서 결정하는 것을 특징으로 하는 문맥 기반 부호화 장치.
- 제 24 항에 있어서, 상기 문맥 분석부는,상기 선택된 샘플의 오른쪽 위의 샘플의 이진 심볼을 상기 선택된 샘플의 문맥으로서 결정하는 것을 특징으로 하는 문맥 기반 부호화 장치.
- 제 24 항에 있어서, 상기 문맥 분석부는,래스터 스캔 방향의 순서대로 상기 샘플을 선택하는 것을 특징으로 하는 문맥 기반 부호화 장치.
- 제 22 항에 있어서, 상기 문맥 기반 부호화 장치는,상기 입력 데이터로서 균일 분포 데이터가 입력되면, 이진 심볼로 변환하여 상기 이진 심볼로 표현된 데이터를 생성하는 B-변환부를 추가로 포함하는 것을 특징으로 하는 문맥 기반 부호화 장치.
- 입력 데이터를 부호화하는 방법에 있어서,상기 입력 데이터가 변환되어 이진 심볼로 표현된 데이터의 샘플들 간의 상관 관계를 분석하여 상기 샘플들에 대한 문맥을 결정하는 문맥 분석 단계; 및상기 결정된 문맥에 따라 상기 샘플들을 부호화함으로써 비트열을 생성하는 부호화 단계를 포함하는 것을 특징으로 하는 문맥 기반 부호화 방법.
- 제 28 항에 있어서, 상기 문맥 기반 부호화 방법은,상기 입력 데이터로서 균일 분포 데이터가 입력되면, 이진 심볼로 변환하여 상기 이진 심볼로 표현된 데이터를 생성하는 B-변환 단계를 추가로 포함하는 것을 특징으로 하는 문맥 기반 부호화 방법.
- 제 28 항에 있어서, 문맥 분석 단계는,상기 이진 심볼로 표현된 샘플들을 나타내는 테이블에서 상기 문맥을 결정할 샘플을 선택하는 샘플 선택 단계;상기 테이블에서 상기 선택된 샘플의 오른쪽 위에 위치한 샘플의 이진 심볼을 검출하는 검출 단계; 및상기 검출된 이진 심볼을 상기 선택된 샘플의 문맥으로서 결정하는 문맥 결정 단계를 포함하는 것을 특징으로 하는 문맥 기반 부호화 방법.
- 부호화를 위한 문맥을 분석하는 장치에 있어서,균일 분포 데이터가 이진 심볼로 표현된 샘플들을 나타내는 테이블에서 상기 문맥을 결정할 샘플을 선택하고, 상기 테이블에서 상기 선택된 샘플의 오른쪽 위에 위치한 샘플의 이진 심볼을 검출하며, 상기 검출된 이진 심볼을 상기 선택된 샘플의 문맥으로서 결정하는 것을 특징으로 하는 문맥 분석 장치.
- 부호화를 위한 문맥을 분석하는 방법에 있어서,균일 분포 데이터가 이진 심볼로 표현된 샘플들을 나타내는 테이블에서 상기 문맥을 결정할 샘플을 선택하는 샘플 선택 단계;상기 테이블에서 상기 선택된 샘플의 오른쪽 위에 위치한 샘플의 이진 심볼을 검출하는 검출 단계; 및상기 검출된 이진 심볼을 상기 선택된 샘플의 문맥으로서 결정하는 문맥 결정 단계를 포함하는 것을 특징으로 하는 문맥 분석 방법.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
US13/063,784 US8462023B2 (en) | 2008-09-16 | 2009-09-09 | Encoding method and encoding apparatus for B-transform, and encoded data for same |
Applications Claiming Priority (4)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
KR1020080090799A KR101516373B1 (ko) | 2008-09-16 | 2008-09-16 | 문맥 분석 방법과 장치 및 그를 이용한 문맥 기반 부호화 방법과 장치 |
KR10-2008-0090799 | 2008-09-16 | ||
KR1020080091578A KR101559824B1 (ko) | 2008-09-18 | 2008-09-18 | B-변환을 위한 부호화 방법 및 장치와 그를 위한 구조를 가진 데이터를 기록한 컴퓨터로 읽을 수 있는 기록매체 |
KR10-2008-0091578 | 2008-09-18 |
Publications (2)
Publication Number | Publication Date |
---|---|
WO2010032934A2 true WO2010032934A2 (ko) | 2010-03-25 |
WO2010032934A3 WO2010032934A3 (ko) | 2010-06-24 |
Family
ID=42039986
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/KR2009/005095 WO2010032934A2 (ko) | 2008-09-16 | 2009-09-09 | B-변환을 위한 부호화 방법 및 장치와 그를 위한 부호화 데이터 |
Country Status (2)
Country | Link |
---|---|
US (1) | US8462023B2 (ko) |
WO (1) | WO2010032934A2 (ko) |
Families Citing this family (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US9306597B1 (en) * | 2015-03-30 | 2016-04-05 | Microsoft Technology Licensing, Llc | Data compression |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US5903676A (en) * | 1994-11-10 | 1999-05-11 | The Chinese University Of Hong Kong | Context-based, adaptive, lossless image codec |
EP1400954A2 (en) * | 2002-09-04 | 2004-03-24 | Microsoft Corporation | Entropy coding by adapting coding between level and run-length/level modes |
KR20040047643A (ko) * | 2002-11-28 | 2004-06-05 | 가부시키가이샤 리코 | 정보 압축 장치, 정보 압축 방법 및 그 프로그램과 기록매체 |
US6927710B2 (en) * | 2002-10-30 | 2005-08-09 | Lsi Logic Corporation | Context based adaptive binary arithmetic CODEC architecture for high quality video compression and decompression |
KR20050090941A (ko) * | 2004-03-10 | 2005-09-14 | 삼성전자주식회사 | 무손실 오디오 부호화/복호화 방법 및 장치 |
Family Cites Families (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US5550540A (en) * | 1992-11-12 | 1996-08-27 | Internatioal Business Machines Corporation | Distributed coding and prediction by use of contexts |
US5432870A (en) * | 1993-06-30 | 1995-07-11 | Ricoh Corporation | Method and apparatus for compressing and decompressing images of documents |
GB2333412B (en) * | 1998-01-20 | 2002-05-15 | Daewoo Electronics Co Ltd | Context-based arithmetic encoding/decoding apparatus |
JP3541930B2 (ja) * | 1998-08-13 | 2004-07-14 | 富士通株式会社 | 符号化装置及び復号化装置 |
-
2009
- 2009-09-09 WO PCT/KR2009/005095 patent/WO2010032934A2/ko active Application Filing
- 2009-09-09 US US13/063,784 patent/US8462023B2/en not_active Expired - Fee Related
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US5903676A (en) * | 1994-11-10 | 1999-05-11 | The Chinese University Of Hong Kong | Context-based, adaptive, lossless image codec |
EP1400954A2 (en) * | 2002-09-04 | 2004-03-24 | Microsoft Corporation | Entropy coding by adapting coding between level and run-length/level modes |
US6927710B2 (en) * | 2002-10-30 | 2005-08-09 | Lsi Logic Corporation | Context based adaptive binary arithmetic CODEC architecture for high quality video compression and decompression |
KR20040047643A (ko) * | 2002-11-28 | 2004-06-05 | 가부시키가이샤 리코 | 정보 압축 장치, 정보 압축 방법 및 그 프로그램과 기록매체 |
KR20050090941A (ko) * | 2004-03-10 | 2005-09-14 | 삼성전자주식회사 | 무손실 오디오 부호화/복호화 방법 및 장치 |
Also Published As
Publication number | Publication date |
---|---|
US8462023B2 (en) | 2013-06-11 |
US20110210875A1 (en) | 2011-09-01 |
WO2010032934A3 (ko) | 2010-06-24 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
WO2010050706A2 (ko) | 움직임 벡터 부호화 방법 및 장치와 그를 이용한 영상 부호화/복호화 방법 및 장치 | |
WO2017043769A1 (ko) | 부호화 장치, 복호화 장치 및 그 부호화 방법 및 복호화 방법 | |
WO2012005551A2 (ko) | 변환 계수의 엔트로피 부호화/복호화 방법 및 장치 | |
WO2012092821A1 (zh) | 一种dna序列数据压缩系统 | |
WO2017043763A1 (ko) | 부호화 장치, 복호화 장치, 그 부호화 및 복호화 방법 | |
WO2011021910A2 (ko) | 인트라 예측 부호화/복호화 방법 및 장치 | |
JP4893957B2 (ja) | 符号化装置、復号化装置、符号化方法及びプログラム | |
WO2010077071A2 (ko) | 블록 모드 부호화/복호화 방법 및 장치와 그를 이용한 영상 부호화/복호화 방법 및 장치 | |
JPS62261230A (ja) | 多重レベル信号の圧縮方法 | |
JPH0358574A (ja) | 網点画像データ圧縮装置 | |
US20130272373A1 (en) | Video encoder with 2-bin per clock cabac encoding | |
WO2014189236A1 (ko) | 무손실 이미지 압축 및 복원 방법과 이를 수행하는 장치 | |
WO2010032934A2 (ko) | B-변환을 위한 부호화 방법 및 장치와 그를 위한 부호화 데이터 | |
US20100246983A1 (en) | Data compression method | |
WO2011053054A9 (ko) | 움직임 벡터 해상도 제한을 이용한 움직임 벡터 부호화/복호화 방법 및 장치와 그를 이용한 영상 부호화/복호화 방법 및 장치 | |
WO2011025301A2 (ko) | 동영상 부호화를 위한 움직임 벡터 부호화/복호화 방법 및 장치와 그를 이용한 영상 부호화/복호화 방법 및 장치 | |
US6771273B2 (en) | Image display apparatus | |
JP4000266B2 (ja) | データ符号化装置、データ符号化方法、及びそのプログラム | |
EP3635953A1 (en) | Electronic device for high-speed compression processing of feature map of cnn utilizing system and controlling method thereof | |
WO2022075754A1 (ko) | 머신비전을 위한 영상의 처리 방법 및 장치 | |
WO2011052821A1 (ko) | 대수우도비를 산출하는 방법 및 복조 장치 | |
WO2020231020A1 (ko) | 연판정 기반으로 선형 부호를 고속 복호화하는 방법 및 장치 | |
US20080025620A1 (en) | Data compression apparatus and data compressing program storage medium | |
KR20050010918A (ko) | 가변길이 복호화를 위한 방법 및 시스템, 및 코드워드들의지역화를 위한 장치 | |
WO2020179966A1 (ko) | 연판정 기반으로 선형 부호를 고속 복호화하는 방법 및 장치 |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
121 | Ep: the epo has been informed by wipo that ep was designated in this application |
Ref document number: 09814750 Country of ref document: EP Kind code of ref document: A2 |
|
NENP | Non-entry into the national phase |
Ref country code: DE |
|
WWE | Wipo information: entry into national phase |
Ref document number: 13063784 Country of ref document: US |
|
32PN | Ep: public notification in the ep bulletin as address of the adressee cannot be established |
Free format text: NOTING OF LOSS OF RIGHTS PERSUANT TO RULE 112(1) EPC, EPO-FORM 1205A SENT 29-06-2011. |
|
122 | Ep: pct application non-entry in european phase |
Ref document number: 09814750 Country of ref document: EP Kind code of ref document: A2 |