WO2008059752A1 - Procédé d'encodage et procédé de décodage de signal d'image, procédé d'encodage et procédé de décodage de source d'information, dispositifs pour ceux-ci, leurs programmes, et support de mémoire avec programme enregistré - Google Patents
Procédé d'encodage et procédé de décodage de signal d'image, procédé d'encodage et procédé de décodage de source d'information, dispositifs pour ceux-ci, leurs programmes, et support de mémoire avec programme enregistré Download PDFInfo
- Publication number
- WO2008059752A1 WO2008059752A1 PCT/JP2007/071719 JP2007071719W WO2008059752A1 WO 2008059752 A1 WO2008059752 A1 WO 2008059752A1 JP 2007071719 W JP2007071719 W JP 2007071719W WO 2008059752 A1 WO2008059752 A1 WO 2008059752A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- information source
- integer value
- integer
- dimensional
- value
- Prior art date
Links
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N1/00—Scanning, transmission or reproduction of documents or the like, e.g. facsimile transmission; Details thereof
- H04N1/41—Bandwidth or redundancy reduction
-
- 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
- H03M7/40—Conversion to or from variable length codes, e.g. Shannon-Fano code, Huffman code, Morse code
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T9/00—Image coding
-
- 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
- H03M7/40—Conversion to or from variable length codes, e.g. Shannon-Fano code, Huffman code, Morse code
- H03M7/42—Conversion to or from variable length codes, e.g. Shannon-Fano code, Huffman code, Morse code using table look-up for the coding or decoding process, e.g. using read-only memory
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/10—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding
- H04N19/102—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the element, parameter or selection affected or controlled by the adaptive coding
- H04N19/13—Adaptive entropy coding, e.g. adaptive variable length coding [AVLC] or context adaptive binary arithmetic coding [CABAC]
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/90—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using coding techniques not provided for in groups H04N19/10-H04N19/85, e.g. fractals
- H04N19/91—Entropy coding, e.g. variable length coding [VLC] or arithmetic coding
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/90—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using coding techniques not provided for in groups H04N19/10-H04N19/85, e.g. fractals
Definitions
- Image signal encoding method and decoding method information source encoding and decoding method, apparatus thereof, program thereof, and storage medium storing the program
- the present invention relates to an image signal encoding method for simply and efficiently encoding an image signal indicating a Gaussian integer signal, and an image signal decoding method for decoding encoded data generated by the image signal encoding method
- An information source encoding method and apparatus for encoding a Gaussian integer signal simply and efficiently, an information source decoding method and apparatus for decoding encoded data generated by the information source encoding method, and Information source encoding program used for realizing the information source encoding method, a computer-readable recording medium storing the program, an information source decoding program used for realizing the information source decoding method, and the program And a computer-readable recording medium on which is recorded.
- a Gaussian signal is a signal whose occurrence probability follows a normal distribution (also called a Gaussian distribution), and this normal distribution is an extremely important distribution that appears in various mathematical / engineering situations.
- the Gaussian signal taken up as an encoding target here is assumed to have an integer signal value. Assuming that generality is not lost, the average signal value is 0 and iid (independent and identically distributed).
- the Golomb code is a signal that follows an exponential distribution (also called a Laplace distribution or a geometric distribution; unless otherwise noted, it refers to a bilateral exponential distribution). It is a code that can be encoded efficiently, does not require a table for encoding and decoding, is extremely simple to process, and is widely used. Furthermore, there is a feature that an arbitrarily large integer input value can be encoded.
- the Golomb code is a Golomb code parameter (g) that takes an integer value of 1 or more. ) Changes the code representation.
- the Golomb code has a relationship in which, for each code parameter g, the code length power increases as the value power 1 ⁇ 2 of the encoded integer z increases.
- the Golomb code parameter g increases, the increase in the code length becomes gradual.For this reason, it is recommended to increase the Golomb code parameter g to encode a gentle distribution. Conversely, as the Golomb code parameter g decreases, the code length increases. Since the increase is steep, the Golomb code parameter g is suitable for encoding a steep distribution that concentrates on 0.
- the Huffman code is a code (compact code) that can be instantaneously decoded and has the shortest average code amount among all the variable length codes using a code table
- the arithmetic code is a variable length code. It is a code that can be compressed to the theoretical limit (sometimes exceeding the compact code) by a different method.
- Patent Document 1 shown below describes a power S that describes an invention in which a feature quantity of an image is vector-quantized according to a binary search tree by a neural network.
- This invention uses dynamic Huffman coding. This technique encodes an image.
- FIG. 21 shows the frequency distribution of normal distribution and exponential distribution with mean 0 and variance 16.
- the vertical axis is the frequency probability, and the logarithm is displayed so that both tails of the distribution are easy to understand.
- the normal distribution is a parabola and the exponential distribution is a triangle.
- each value must be converted to an integer of 0 or more.
- a method of separately expressing positive / negative information and absolute value information, and the following relatively simple conversion are often used.
- the encoding efficiency is significantly lower than that of the Huffman code or the arithmetic code. This is an essential problem because the signal occurrence probability assumed by the Golomb code follows an exponential distribution rather than a normal distribution.
- both encoder and decoder require a frequency table that is not required for Golomb codes.
- Non-Patent Document 2 shown below proposes a hybrid golomb code capable of instantaneous decoding of a generalized Gaussian signal source. This is capable of encoding / decoding arbitrarily large integer input values, but the structure is complicated, and it targets distributions that are steeper than exponential distributions! /, And! / was there. [0019]
- Golomb encoding is performed at the expense of encoding efficiency. The method to do was taken. However, this method has a problem that it is necessary to optimize Golomb code parameters.
- Patent Document 1 Japanese Patent Laid-Open No. 2001-5967
- Non-patent 1 P Boldi, 3 ⁇ 4 Vigna: ompressed perfect embedded skip lists for quick ⁇ nvertea-index lookups, Proceedings of String Processing and Information Retrieval, 12th International Conference, pp.1-15, 2005
- Non-Patent Document 2 S Xue, Y Xu, B Oelmann: "Hybrid Golomb codes for a group of quant ised GG sources, IEE Proceedings, Vision Image and signal Processing, Vol.150, No.4, pp.256-260 , 2003
- the Golomb code is a code that can efficiently encode a signal according to an exponential distribution, does not require a table for encoding and decoding, and can be instantaneously decoded with extremely simple processing.
- the feature is that an arbitrarily large integer input value can be encoded.
- the Golomb code when used for coding a Gaussian signal, assumes the signal generation probability of exponential distribution, and the coding efficiency is higher than that of the Huffinan code or arithmetic code. It will drop greatly.
- Non-Patent Document 2 proposes a hybrid golomb code.
- this method has a problem that the configuration is complicated, and also has a problem that it targets a steeper distribution than a normal distribution.
- Non-Patent Document 1 the Golomb code parameter is optimized and the Golomb code is used by giving priority to the simplicity of the encoding process without considering the encoding efficiency.
- this method has the problem that the coding efficiency is lower than that of Huffman codes and arithmetic codes, and the Golomb code parameters need to be optimized.
- the present invention has been made in view of power and circumstances, and is a new information source encoding / decoding technique that enables simple and efficient encoding / decoding of a Gaussian integer signal.
- the purpose is to provide.
- the information source encoding device of the present invention provides (i) input means for inputting a signal value sequence of a Gaussian integer signal to be encoded in order to realize simple and efficient encoding of a Gaussian integer signal. (Port) Integer value pair conversion means for converting the signal values included in the signal value series input by the input means into two integer value pairs in the input order; and (c) Conversion of the integer value pair conversion means.
- a mapping means that obtains an integer value of 0 or more by applying a 2D ⁇ 1D mapping that maps each integer value pair as a grid point on 2D coordinates and maps the grid point closer to the origin to a smaller value
- the encoding means may encode the integer value obtained by the mapping means using a Golomb code suitable for encoding the information source according to the exponential distribution.
- a Golomb code suitable for encoding the information source according to the exponential distribution.
- variance calculation means for calculating the variance of the signal value input by the input means
- a value proportional to the variance value calculated by the variance calculation means Code parameter determining means for determining a code parameter of a Golomb code having.
- the information source encoding apparatus of the present invention can also encode an image signal indicating a Gaussian integer signal. In this case, the information source encoding apparatus of the present invention It will function as an image signal encoding device.
- the information source encoding method of the present invention realized by the operation of each processing means described above can also be realized by a computer program.
- This computer program is recorded on an appropriate computer-readable recording medium.
- the present invention is realized by operating on a control means such as a CPU, which is provided when the present invention is implemented and installed on the network.
- the information source decoding apparatus provides (i) an integer value generated by the information source encoding apparatus of the present invention in order to realize decoding of the encoded data generated by the information source encoding apparatus of the present invention.
- a one-dimensional ⁇ two-dimensional mapping that is a reverse mapping of the two-dimensional mapping
- a restoring means that restores the integer value pair that is the mapping source of the integer value
- a restored integer value pair is constructed.
- an output means for outputting the integer values in that order.
- the decoding means When adopting this configuration, when the information source encoding device of the present invention generates encoded data of integer values using Golomb codes, the decoding means performs Golo mb decoding of the encoded data of integer values. By decoding the integer value.
- the information source encoding apparatus of the present invention determines the code parameter of the Golomb code having a value proportional to the variance of the signal value to be encoded, it is determined as such as the Golomb code parameter used for decoding.
- a Golomb code parameter input means for inputting the Golomb code parameters thus obtained is provided.
- the information source decoding device of the present invention when the information source encoding device of the present invention generates encoded data using an image signal indicating a Gaussian integer signal as an encoding target, the information source decoding device of the present invention performs image signal decoding. It will function as a device.
- the information source decoding method of the present invention realized by the operation of each processing means described above can also be realized by a computer program.
- the present invention is realized by being provided on a computer-readable recording medium or provided via a network, installed when the present invention is carried out, and operating on a control means such as a CPU. become.
- the encoding efficiency of the variable-length code is improved by expanding the information source.
- the integer-to-one integer value conversion process used in the present invention is nothing but secondary expansion of the information source. From now on, according to the present invention, it is possible to improve the encoding efficiency! RU
- FIG. 1 is a diagram showing an example of a two-dimensional ⁇ one-dimensional mapping executed in the present invention.
- FIG. 2 is an explanatory diagram of a table that stores the correspondence between integer value pairs and integer values.
- FIG. 3 shows an apparatus configuration of an information source encoding device and an information source decoding device according to the present invention.
- FIG. 4 is a flowchart executed by the information source encoding apparatus of the above embodiment.
- FIG. 5 is an explanatory diagram of an algorithm for realizing a two-dimensional to one-dimensional mapping executed in the present invention.
- FIG. 6 is an explanatory diagram of a result of an experiment performed to verify the effectiveness of the present invention.
- FIG. 7] is also an explanatory diagram of the results of experiments conducted to verify the effectiveness of the present invention.
- FIG. 8 is also an explanatory diagram of the results of experiments conducted to verify the effectiveness of the present invention.
- FIG. 9 Similarly, it is an explanatory diagram of the results of experiments conducted to verify the effectiveness of the present invention.
- FIG. 10 Similarly, it is an explanatory diagram of the results of experiments conducted to verify the effectiveness of the present invention.
- FIG. 11 is a flowchart executed by the information source decoding apparatus of the above embodiment.
- FIG. 12 is a detailed flowchart executed by the information source encoding apparatus of the embodiment.
- FIG. 13 is a detailed flowchart executed by the information source encoding apparatus of the embodiment.
- FIG. 14 is a detailed flowchart executed by the information source encoding apparatus of the embodiment.
- FIG. 15 is a detailed flowchart executed by the information source encoding apparatus of the embodiment.
- FIG. 16 is a detailed flowchart executed by the information source encoding apparatus according to the above-described embodiment.
- FIG. 16 is a detailed flowchart executed by the information source decoding apparatus according to the above-described embodiment.
- 18 This is a detailed flowchart executed by the information source decoding apparatus of the above embodiment.
- 19 This is a detailed flowchart executed by the information source decoding apparatus of the above embodiment.
- FIG. 20 is an explanatory diagram of Golomb codes.
- Gaussian signals are input in the order of a, a, a, a,.
- a two-element integer in the order of its input such as (a, a), (a, a), ...
- an integer value z greater than or equal to 0 is obtained by applying a two-dimensional to one-dimensional mapping that maps the integer value pair (X, y) to a smaller value at a lattice point closer to the origin.
- This two-dimensional to one-dimensional mapping is such that a grid point closer to the origin maps to a smaller value.
- the integer value pair (X, y) is mapped to the integer value z in the form shown in Figure 1.
- mapping process for example, as shown in FIG. 2, a table for storing the correspondence between integer value pairs and integer values is prepared in advance, and the integer value pair (X, y) is used as a key. By referring to the table, it is possible to obtain the integer value z that is the mapping result for the integer value pair (X, y).
- the table prepared at this time is a general one that does not depend on the information source to be encoded. Can be prepared.
- the minimum value of the distance between the unprocessed (not listed) grid point and the origin is obtained from the grid point of the origin, and the grid point of the minimum distance is defined. It is possible to obtain an integer value z that is the mapping result for an integer value pair (X, y) by repeating enumeration and assigning integer values according to the order of .
- the z value of a lattice point that is separated from the origin by a distance L is approximately equal to the number of lattice points in the circumference of radius L, which is substantially equal to the area of the circumference. From now on, it becomes "z ⁇ ". Note that this is a global approximation (the actual number of grid points is quite large).
- the probability f of this lattice point can be expressed as “f ex p ( —a L 2 ) (where a is a constant)” by a normal distribution. Therefore, from the relation of “z TT L 2 ”, “f cc exp (— az / ⁇ ) "Exponential distribution for z.
- a two-dimensional to one-dimensional mapping is performed by using a Gaussian signal source as an exponential distribution signal source. Realizes a reversible mapping that converts to.
- the Golomb code is a code that can efficiently encode a signal according to an exponential distribution, does not require a table for encoding and decoding, and can be instantaneously decoded with extremely simple processing.
- the feature is that an arbitrarily large integer input value can be encoded.
- the integer value ⁇ encoded by the information encoding method of the present invention is decoded in the input order of the encoded data.
- the information source encoding method of the present invention when the information source encoding method of the present invention generates encoded data of an integer value using a Golomb code, the integer value is decoded by Golomb decoding the encoded data.
- the decoded integer value z is subjected to a 1D ⁇ 2D mapping, which is a reverse mapping of the 2D ⁇ 1D mapping used by the information source coding method of the present invention,
- the integer value pair (X, y) from which the integer value is mapped is restored.
- This one-dimensional to two-dimensional mapping is performed by executing the mapping process of the two-dimensional to one-dimensional mapping used by the information source encoding method of the present invention, and when the decoded integer value z appears at that time. Then, the integer value pair (X, y) associated with it is specified.
- this mapping process is similar to the mapping process performed by the information source encoding method of the present invention.
- a table that stores beforehand the correspondence between integer value pairs and integer values as shown in FIG.
- you can obtain an integer value pair (X, y) that is the mapping result for the integer value z. is there.
- this mapping process starts from the grid point of the origin, obtains the minimum value of the distance between the grid point not listed and the original point, and lists the grid points of the minimum distance according to a prescribed order. Then, by assigning integer values repeatedly, it is possible to obtain an integer value pair (X, y) that is a mapping result for the integer value z. [0061] Next, the information source decoding method of the present invention outputs the integer values X and y constituting the restored integer value pair (X, y) in that order.
- a Gaussian integer signal is obtained by decoding encoded data encoded with a code used for encoding an information source according to an exponential distribution. Decoding can be performed easily and efficiently.
- the Gaussian signal source is converted into an exponential signal source through quadratic expansion and one-dimensional serialization of the Gaussian signal source, and this is encoded by a Golomb code having a property suitable for the exponential signal source. According to the embodiment, the present invention will be described in detail.
- FIG. 3 shows an embodiment of the device configuration of the information source encoding device 1 and the information source decoding device 2 provided with the present invention.
- the information source encoding device 1 of the present invention has a signal input unit 10 for inputting a signal value sequence of a Gaussian integer signal to be encoded, and an input of the signal input unit 10.
- the integer value pair conversion unit 11 that converts signal values into two integer value pairs in the order of input, and the integer value pair converted by the integer value pair conversion unit 11, the smaller the lattice point closer to the origin. 2D ⁇ 1D mapping to obtain an integer value greater than or equal to 0 2D ⁇ 1D mapping unit 12 and Golomb code, 2D ⁇ 1D mapping unit 12 obtained integer value And a Golomb encoding unit 13.
- the information source decoding device 2 of the present invention includes an encoded data input unit 20 for inputting encoded data of an integer value generated by the information source encoding device 1 of the present invention, and an encoded data input unit 20
- the Golomb decoding unit 21 that decodes the integer value by performing Golomb decoding on the encoded data of the integer value input in step 2 and the mapping of the two-dimensional to one-dimensional mapping unit 12 included in the information source encoding device 1 of the present invention.
- the Golomb decoding unit 21 When the integer value decoded by the Golomb decoding unit 21 appears at that time, by specifying the integer value pair associated therewith, the Golomb decoding unit 21 1 The integer values that make up the restored integer value pairs in the 1D ⁇ 2D inverse mapping unit 22 and the 1D ⁇ 2D inverse mapping unit 22 are output in that order. A sequential output unit 23 is provided.
- FIG. 4 shows an example of a flow chart executed by the information source encoding apparatus 1 of the present invention configured as shown in FIG. Next, processing executed by the information source encoding device 1 of the present invention will be described in detail according to this flowchart.
- step S101 the information source encoding apparatus 1 of the present invention inputs a signal value of a Gaussian integer signal two elements at a time from the top.
- step S102 the Gaussian properties are inputted in the order of a, a ⁇ a, a ⁇ ,.
- a numerical signal is converted into an integer value pair of two elements in the order of input, such as (a, a), (a, a), ...
- step S 103 a two-dimensional to one-dimensional mapping is performed on this integer value pair (X, y), which is mapped to a smaller value at a lattice point closer to the origin. get z.
- step S 104 the integer value z is Golomb-coded using a separately provided Golomb code parameter g, and in the subsequent step S 105, the encoded data is output.
- step S 106 it is determined whether or not the input of the Gaussian integer signal has been exhausted. If it has not been exhausted, the process returns to step S 101 again, and if it has been exhausted, the process ends.
- step S 103 Next, the 2D ⁇ 1D mapping executed in step S 103 will be described.
- l (x, y) shown in Fig. 5 gives the distance between the origin and the point (x, y) and is usually the Euclidean distance.
- the distance function l (x, y) is not necessarily isotropic, and the input Gaussian source is not strictly iid (ind. Endent and identically distributed). If there is, use an appropriate constant ⁇ ,
- L l norm or more generally the L y norm ( ⁇ of each signal element) shown in the following equation.
- FIG. 6 shows a frequency distribution of a signal source that is very close to a Gaussian signal obtained by processing an actual image.
- the vertical axis is the logarithm of probability. It can be seen that the frequency distribution of this signal source is parabolic, that is, follows a normal distribution.
- Fig. 7 shows the signals of this signal source sequentially taken out as two numerical pairs (X, y) and their frequency distributions displayed in three dimensions. Similarly, the vertical axis is the logarithm of probability. The graph is a paraboloid that rotates the parabola.
- FIG. 8 shows the contour lines of the paraboloid of revolution. Although a slight positive correlation is observed, it is almost concentric.
- the exponential distribution shows a triangular force signal as shown in Fig. 21.
- the one-sided distribution by this conversion becomes a straight line that descends to the right.
- FIG. 9 shows the frequency distribution of the two-dimensional to one-dimensional mapping applied to (X, y) according to the present invention and converted to the one-dimensional value z as a vertical logarithmic graph similar to FIG.
- FIG. 10 shows a comparison of the obtained encoding efficiencies.
- datal corresponds to the above data.
- the coding efficiency of the present invention has achieved at least about 90%.
- the present invention has an average of 94.1% and the conventional method has a difference of about 7 points, 87.4%.
- FIG. 11 shows an example of a flow chart executed by the information source decoding apparatus 2 of the present invention configured as shown in FIG.
- the information source decoding device 2 of the present invention firstly, at step S201, the encoded data (Golomb code of integer value z) generated by the information source encoding device 1 of the present invention from the head. Enter one element at a time according to the order.
- step S202 the input Golomb code is decoded using the Golomb code parameter g to obtain an integer value z of 0 or greater.
- step S203 the integer value z is converted to an integer value pair (X
- this integer value pair (X, y) is output in the order of X, y.
- the one-dimensional ⁇ two-dimensional mapping executed in step S203 executes the mapping process of the two-dimensional ⁇ one-dimensional mapping used by the information source coding apparatus 1 of the present invention, and is then decoded.
- an integer value z appears, specify the integer value pair (X, y) associated with it.
- step S205 it is determined whether or not the input of the Golomb code has been exhausted. If V, Na! /, Return to Step S 201, and if it is exhausted! /, End.
- FIG. 12 shows a detailed flowchart of the process executed in step S102 of the flowchart of FIG.
- step S102 when the signal values of the two Gaussian integer signals input in step S101 are received, as shown in the flowchart of Fig. 12, first, in step S301, of them. The first signal value X is input and stored in the memory X. Subsequently, in step S302, the other signal value y is input and stored in the memory y.
- step S303 the signal X is read from the memory X, the signal value y is read from the memory y, and these X and y are output as an integer value pair (X, y).
- step S102 by executing the flowchart of FIG. 12, the process of bundling and outputting two signal values in the signal sequence of the Gaussian integer signal is performed. .
- FIGS. 13 to 16 show detailed flowcharts of the processing executed in step S103 of the flowchart of FIG.
- step S103 the two-dimensional to one-dimensional mapping realized by the algorithm shown in Fig. 5 is performed to map the integer value pair (X, y) to the integer value z. Performs processing to output the numerical value z.
- step S401 when the integer values xO and yO to be mapped are input, the lattice point storage memory Z is emptied in the subsequent step S402. At the same time, the variable z is initialized to 0.
- step S403 procedure A defined by the processing of the flowchart shown in FIG. Run and continue
- step S404 procedure B defined by the processing of the flowchart shown in FIG. 15 is executed.
- procedure A is set by procedure B "dmin ⁇ x ⁇ dmin,- For dmin ⁇ y ⁇ dmin ", the minimum value dmin of the distance to the origin is determined for the unenumerated grid point (x, y) to which no integer value is assigned. At this time, there are usually a plurality of grid points that realize this minimum value dmin.
- procedure B the grid points are selected one by one in the prescribed order, and an integer value one larger than the previous one is assigned to them! /, And so on.
- X and y are both dmin and less than dmin (dmin is an integer, and is incremented by 1 by procedure B).
- the minimum value of the distance from the origin is obtained as a processing target, and it is set as dmin.
- step S501 the integer value dmin is initialized to zero. Subsequently, in step S502, one set is generated for each loop (S502 to S508) in accordance with the specified order without any interval from X to dmin and y from dmin to dmin.
- step S503 the generated (X, y) is already stored in the lattice point storage memory Z. If it is already stored, the process proceeds to step S508. If it is not stored, the process proceeds to step S504, and the variable d is processed by step S505. Substitute the distance l (x, y) between the required grid point (X, y) and the origin.
- the distance Kx, y) used here is, for example,
- step S506 the magnitudes of d and dmin are compared. If d is greater than or equal to dmin, the process immediately proceeds to step S508. If d is less than dmin, the process proceeds to step S507. Proceed to set the value of dmin to d, and then proceed to step S508.
- step S508 it is determined whether all combinations of X and y are within the range from dmin to dmin. If not, the process of step S502 is performed. Return to, and if it is exhausted, the processing as the procedure A ends.
- step S601 the flag found is set to 0. Subsequently, in step S602, X is generated from dmin to dmin and y from dmin to dmin, one set for each one crepe (S602 to S609) in the prescribed order.
- step S603 it is determined whether or not the generated (x, y) has already been stored in the lattice point storage memory Z. Proceeding to S609, if not stored, proceeding to step S604, and for the variable d, the distance l (x, y) between the lattice point (X, y) obtained by the processing of step S605 and the origin. Substitute y).
- the distance Kx, y) used here is, for example,
- step S606 d and dmin are compared. If d and dmin do not match! /, Proceed immediately to step S609, and if d and dmin match, proceed to step S607 and execute procedure X.
- step S608 a new integer value pair (x, y) is stored in the lattice point storage memory Z, the value of the separately defined integer z is incremented by 1, and the flag Set the found value to 1 before proceeding to step S609.
- step S609 it is determined whether all combinations of X and y are within the range from dmin to dmin. If not, the process of step S602 is performed. If it returns and runs out, the process proceeds to step S610 to determine whether the value of the flag found is equal to 0! /.
- step S610 When it is determined that the value of flag found is not equal to 0 according to the determination process of step S610, the process returns to step S601, and when it is determined that the value of flag found is equal to 0, step Proceeding to S611, the value of dmin is incremented by 1, and the processing as procedure B ends.
- step S103 of the flowchart of Fig. 4 the two-dimensional to one-dimensional mapping realized by the algorithm shown in Fig. 5 is performed. By doing so, the process of mapping the integer value pair (X, y) to the integer value z is performed, and the process of outputting the integer value z is performed.
- step S203 the integer value z decoded in the process of step S202 is converted into an integer value pair (X, y) by applying a one-dimensional to two-dimensional mapping realized by the algorithm shown in FIG. The mapping process is performed, and the integer value pair (X, y) is output.
- step S801 when the integer value ⁇ to be mapped is input, in the subsequent step S802, the lattice point storage memory Z is emptied, Initialize variable z to 0.
- step S803 procedure A defined by the process of the flowchart shown in FIG. 14 is executed, and in step S804, procedure B defined by the process of the flowchart shown in FIG. 15 is executed ( However, “Procedure X” is changed to “Procedure X '” described below)
- step S803 if the condition is not satisfied, nothing is output and the process returns to step A by returning to step S803.
- the numerical value (X, y) is output as the mapping result, the process proceeds to step S903, and the one-dimensional ⁇ two-dimensional mapping process is terminated.
- the procedure X ′ is terminated (that is, the procedure returns to step S608 of procedure B).
- step S203 of the flowchart of Fig. 11 the one-dimensional to two-dimensional copy realized by the algorithm shown in Fig. 5 is performed.
- the process of mapping the integer value z to the integer value pair (x, y) is performed, and the process of outputting the integer value pair (X, y) is performed.
- FIG. 19 shows a detailed flowchart of the process executed in step S204 of the flowchart of FIG.
- step S204 when the integer value pair (X, y) obtained in the process of step S203 is received, as shown in the flowchart of FIG. 19, first, in step S 1001, the top of them is displayed. Is stored in the memory X, and the other signal value y is stored in the memory y.
- step S1002 the first signal value X is read out from the memory X and output, and in the subsequent step S1003, the other signal value y is read out from the memory y and output.
- step S204 by executing the flowchart of FIG. 19, the integer value X, y constituting the integer value pair (X, y) of the mapping result is output according to the order. It does.
- the present invention encodes and decodes a Gaussian integer signal, and uses a configuration that realizes a reversible mapping that converts a Gaussian signal source to a signal source having an exponential distribution. Although it appears in various scenes, it is possible to simply and efficiently encode and decode Gaussian integer signals that could not be encoded efficiently by conventional Golomb codes.
Landscapes
- Engineering & Computer Science (AREA)
- Multimedia (AREA)
- Signal Processing (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Compression Or Coding Systems Of Tv Signals (AREA)
- Compression, Expansion, Code Conversion, And Decoders (AREA)
- Compression Of Band Width Or Redundancy In Fax (AREA)
- Image Processing (AREA)
Description
Claims
Priority Applications (7)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
US12/446,818 US8300966B2 (en) | 2006-11-14 | 2007-11-08 | Image signal encoding method and decoding method, information source encoding method and decoding method, apparatuses therefor, programs therefor, and storage media which store the programs to process gaussian integer signal |
BRPI0718239A BRPI0718239B1 (pt) | 2006-11-14 | 2007-11-08 | processo de codificação de fonte de informação, processo de decodificação de fonte de informação, aparelho e codificação de fonte de informação e aparelho de decodificação de fonte de informação |
EP20070831450 EP2083560B1 (en) | 2006-11-14 | 2007-11-08 | Image signal coding method and decoding mehod, information source coding method and decoding mehod, devices for them, their programs, and memory medium with recorded program |
CA2667727A CA2667727C (en) | 2006-11-14 | 2007-11-08 | Information source encoding method and decoding method, apparatuses therefor, programs therefor, and storage media which store the programs |
ES07831450.7T ES2450265T3 (es) | 2006-11-14 | 2007-11-08 | Método de codificación y método de decodificación de señal de imagen, método de codificación y método de decodificación de fuente de información, dispositivos para ellos, sus programas, y medio de memoria con programa registrado |
CN2007800412601A CN101536484B (zh) | 2006-11-14 | 2007-11-08 | 图像信号编码方法及解码方法、信息源编码及解码方法、它们的装置 |
JP2008544115A JP4769305B2 (ja) | 2006-11-14 | 2007-11-08 | 画像信号符号化方法及び復号方法、情報源符号化及び復号方法、それらの装置、及びそれらのプログラム並びにプログラムを記録した記憶媒体 |
Applications Claiming Priority (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
JP2006-307512 | 2006-11-14 | ||
JP2006307512 | 2006-11-14 |
Publications (1)
Publication Number | Publication Date |
---|---|
WO2008059752A1 true WO2008059752A1 (fr) | 2008-05-22 |
Family
ID=39401559
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/JP2007/071719 WO2008059752A1 (fr) | 2006-11-14 | 2007-11-08 | Procédé d'encodage et procédé de décodage de signal d'image, procédé d'encodage et procédé de décodage de source d'information, dispositifs pour ceux-ci, leurs programmes, et support de mémoire avec programme enregistré |
Country Status (11)
Country | Link |
---|---|
US (1) | US8300966B2 (ja) |
EP (1) | EP2083560B1 (ja) |
JP (1) | JP4769305B2 (ja) |
KR (1) | KR101010938B1 (ja) |
CN (1) | CN101536484B (ja) |
BR (1) | BRPI0718239B1 (ja) |
CA (1) | CA2667727C (ja) |
ES (1) | ES2450265T3 (ja) |
RU (1) | RU2406222C1 (ja) |
TW (1) | TWI335147B (ja) |
WO (1) | WO2008059752A1 (ja) |
Cited By (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2010005359A1 (en) * | 2008-07-10 | 2010-01-14 | Ali Majeed | Data compression involving spiral transformation |
WO2018235418A1 (ja) * | 2017-06-22 | 2018-12-27 | 日本電信電話株式会社 | 符号化装置、復号装置、符号化方法、復号方法、およびプログラム |
WO2019021552A1 (ja) * | 2017-07-25 | 2019-01-31 | 日本電信電話株式会社 | 符号化装置、復号装置、符号列のデータ構造、符号化方法、復号方法、符号化プログラム、復号プログラム |
WO2019198383A1 (ja) * | 2018-04-13 | 2019-10-17 | 日本電信電話株式会社 | 符号化装置、復号装置、符号化方法、復号方法、プログラム、および記録媒体 |
WO2020095706A1 (ja) * | 2018-11-07 | 2020-05-14 | 日本電信電話株式会社 | 符号化装置、復号装置、符号列のデータ構造、符号化方法、復号方法、符号化プログラム、復号プログラム |
Families Citing this family (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US8817882B2 (en) | 2010-07-30 | 2014-08-26 | Qualcomm Incorporated | Coding blocks of data using a generalized form of golomb codes |
EP3621206B1 (en) * | 2011-10-24 | 2024-06-05 | Velos Media International Limited | Significant map decoding using partition selection |
US8891630B2 (en) | 2011-10-24 | 2014-11-18 | Blackberry Limited | Significance map encoding and decoding using partition set based context assignment |
CN104038234B (zh) | 2013-03-07 | 2017-09-29 | 华为技术有限公司 | 极性码的译码方法和译码器 |
GB201321851D0 (en) | 2013-12-10 | 2014-01-22 | Canon Kk | Run length parameters coding for palette mode |
CN105009461B (zh) * | 2013-12-24 | 2017-12-22 | 华为技术有限公司 | 极性码的译码方法和译码装置 |
EP3761313B1 (en) * | 2018-03-02 | 2023-01-18 | Nippon Telegraph And Telephone Corporation | Encoding device, encoding method, program, and recording medium |
CN114124413A (zh) * | 2020-08-10 | 2022-03-01 | 中国电信股份有限公司 | 媒体文件加密方法、解密方法、传输方法和装置 |
Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JPH10243395A (ja) * | 1997-02-28 | 1998-09-11 | Graphics Commun Lab:Kk | 画像の符号化および復号方法と装置 |
JP2001005967A (ja) | 1999-06-21 | 2001-01-12 | Matsushita Electric Ind Co Ltd | 画像送信装置及びニューラルネットワーク |
Family Cites Families (14)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US5303265A (en) | 1990-06-29 | 1994-04-12 | Digital Equipment Corporation | Frequency independent encoding technique and apparatus for digital communications |
US5471207A (en) * | 1994-02-23 | 1995-11-28 | Ricoh Company Ltd. | Compression of palettized images and binarization for bitwise coding of M-ary alphabets therefor |
CA2156889C (en) | 1994-09-30 | 1999-11-02 | Edward L. Schwartz | Method and apparatus for encoding and decoding data |
US6137589A (en) * | 1995-09-20 | 2000-10-24 | Hewlett-Packard Company | Efficient storage of dithered raster image data in a printer or the like |
US5764374A (en) | 1996-02-05 | 1998-06-09 | Hewlett-Packard Company | System and method for lossless image compression having improved sequential determination of golomb parameter |
JP2000115782A (ja) | 1998-10-06 | 2000-04-21 | Canon Inc | 符号化装置及び方法及び記憶媒体 |
KR100317279B1 (ko) | 1998-11-04 | 2002-01-15 | 구자홍 | 무손실부호화방법및장치 |
US6668092B1 (en) * | 1999-07-30 | 2003-12-23 | Sun Microsystems, Inc. | Memory efficient variable-length encoding/decoding system |
KR100399932B1 (ko) | 2001-05-07 | 2003-09-29 | 주식회사 하이닉스반도체 | 메모리의 양을 감소시키기 위한 비디오 프레임의압축/역압축 하드웨어 시스템 |
US7342965B2 (en) | 2002-04-26 | 2008-03-11 | Nokia Corporation | Adaptive method and system for mapping parameter values to codeword indexes |
US6987468B1 (en) * | 2004-10-29 | 2006-01-17 | Microsoft Corporation | Lossless adaptive encoding and decoding of integer data |
US7015837B1 (en) | 2004-10-29 | 2006-03-21 | Microsoft Corporation | Lossless adaptive encoding and decoding of integer data |
US7580585B2 (en) | 2004-10-29 | 2009-08-25 | Microsoft Corporation | Lossless adaptive Golomb/Rice encoding and decoding of integer data using backward-adaptive rules |
JP4468858B2 (ja) | 2005-06-01 | 2010-05-26 | オリンパスイメージング株式会社 | データ符号化装置、データ符号化方法、プログラム |
-
2007
- 2007-11-08 CA CA2667727A patent/CA2667727C/en not_active Expired - Fee Related
- 2007-11-08 BR BRPI0718239A patent/BRPI0718239B1/pt not_active IP Right Cessation
- 2007-11-08 RU RU2009116946/09A patent/RU2406222C1/ru not_active IP Right Cessation
- 2007-11-08 JP JP2008544115A patent/JP4769305B2/ja not_active Expired - Fee Related
- 2007-11-08 US US12/446,818 patent/US8300966B2/en active Active
- 2007-11-08 EP EP20070831450 patent/EP2083560B1/en not_active Not-in-force
- 2007-11-08 ES ES07831450.7T patent/ES2450265T3/es active Active
- 2007-11-08 KR KR20097008940A patent/KR101010938B1/ko active IP Right Grant
- 2007-11-08 CN CN2007800412601A patent/CN101536484B/zh not_active Expired - Fee Related
- 2007-11-08 WO PCT/JP2007/071719 patent/WO2008059752A1/ja active Application Filing
- 2007-11-12 TW TW096142635A patent/TWI335147B/zh not_active IP Right Cessation
Patent Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JPH10243395A (ja) * | 1997-02-28 | 1998-09-11 | Graphics Commun Lab:Kk | 画像の符号化および復号方法と装置 |
JP2001005967A (ja) | 1999-06-21 | 2001-01-12 | Matsushita Electric Ind Co Ltd | 画像送信装置及びニューラルネットワーク |
Non-Patent Citations (3)
Title |
---|
P BOLDI; S VIGNA: "Compressed perfect embedded skip lists for quick inverted-index lookups", PROCEEDINGS OF STRING PROCESSING AND INFORMATION RETRIEVAL, 12TH INTERNATIONAL CONFERENCE, 2005, pages 1 - 15 |
S XUE; Y XU; B OELMANN: "Hybrid Golomb codes for a group of quantised GG sources", IEE PROCEEDINGS, VISION IMAGE AND SIGNAL PROCESSING, vol. 150, no. 4, 2003, pages 256 - 260, XP006021077, DOI: doi:10.1049/ip-vis:20030510 |
See also references of EP2083560A4 * |
Cited By (15)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2010005359A1 (en) * | 2008-07-10 | 2010-01-14 | Ali Majeed | Data compression involving spiral transformation |
WO2018235418A1 (ja) * | 2017-06-22 | 2018-12-27 | 日本電信電話株式会社 | 符号化装置、復号装置、符号化方法、復号方法、およびプログラム |
CN110771045A (zh) * | 2017-06-22 | 2020-02-07 | 日本电信电话株式会社 | 编码装置、解码装置、编码方法、解码方法、以及程序 |
US12028093B2 (en) | 2017-06-22 | 2024-07-02 | Nippon Telegraph And Telephone Corporation | Encoder, decoder, encoding method, decoding method and program |
CN110771045B (zh) * | 2017-06-22 | 2024-03-29 | 日本电信电话株式会社 | 编码装置、解码装置、编码方法、解码方法、以及记录介质 |
JPWO2018235418A1 (ja) * | 2017-06-22 | 2020-05-21 | 日本電信電話株式会社 | 符号化装置、復号装置、符号化方法、復号方法、およびプログラム |
US10840944B2 (en) | 2017-07-25 | 2020-11-17 | Nippon Telegraph And Telephone Corporation | Encoding apparatus, decoding apparatus, data structure of code string, encoding method, decoding method, encoding program and decoding program |
WO2019021552A1 (ja) * | 2017-07-25 | 2019-01-31 | 日本電信電話株式会社 | 符号化装置、復号装置、符号列のデータ構造、符号化方法、復号方法、符号化プログラム、復号プログラム |
JPWO2019021552A1 (ja) * | 2017-07-25 | 2020-04-09 | 日本電信電話株式会社 | 符号化装置、復号装置、符号列のデータ構造、符号化方法、復号方法、符号化プログラム、復号プログラム |
WO2019198383A1 (ja) * | 2018-04-13 | 2019-10-17 | 日本電信電話株式会社 | 符号化装置、復号装置、符号化方法、復号方法、プログラム、および記録媒体 |
JPWO2019198383A1 (ja) * | 2018-04-13 | 2021-04-08 | 日本電信電話株式会社 | 符号化装置、復号装置、符号化方法、復号方法、プログラム、および記録媒体 |
JP7173134B2 (ja) | 2018-04-13 | 2022-11-16 | 日本電信電話株式会社 | 符号化装置、復号装置、符号化方法、復号方法、プログラム、および記録媒体 |
JP7139897B2 (ja) | 2018-11-07 | 2022-09-21 | 日本電信電話株式会社 | 符号化装置、復号装置、符号化方法、復号方法、符号化プログラム、復号プログラム |
JP2020077969A (ja) * | 2018-11-07 | 2020-05-21 | 日本電信電話株式会社 | 符号化装置、復号装置、符号列のデータ構造、符号化方法、復号方法、符号化プログラム、復号プログラム |
WO2020095706A1 (ja) * | 2018-11-07 | 2020-05-14 | 日本電信電話株式会社 | 符号化装置、復号装置、符号列のデータ構造、符号化方法、復号方法、符号化プログラム、復号プログラム |
Also Published As
Publication number | Publication date |
---|---|
KR101010938B1 (ko) | 2011-01-25 |
EP2083560B1 (en) | 2013-12-04 |
KR20090084841A (ko) | 2009-08-05 |
US8300966B2 (en) | 2012-10-30 |
ES2450265T3 (es) | 2014-03-24 |
TWI335147B (en) | 2010-12-21 |
JPWO2008059752A1 (ja) | 2010-03-04 |
EP2083560A1 (en) | 2009-07-29 |
CN101536484B (zh) | 2011-10-05 |
JP4769305B2 (ja) | 2011-09-07 |
RU2406222C1 (ru) | 2010-12-10 |
CA2667727C (en) | 2014-07-08 |
CN101536484A (zh) | 2009-09-16 |
BRPI0718239A2 (pt) | 2013-11-19 |
CA2667727A1 (en) | 2008-05-22 |
EP2083560A4 (en) | 2011-12-28 |
BRPI0718239B1 (pt) | 2019-12-24 |
US20100027903A1 (en) | 2010-02-04 |
TW200830732A (en) | 2008-07-16 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
WO2008059752A1 (fr) | Procédé d'encodage et procédé de décodage de signal d'image, procédé d'encodage et procédé de décodage de source d'information, dispositifs pour ceux-ci, leurs programmes, et support de mémoire avec programme enregistré | |
US11632560B2 (en) | Methods and apparatuses for encoding and decoding a bytestream | |
TWI239489B (en) | System and method for progressively transforming and coding digital data | |
US20190228050A1 (en) | Bezier volume representation of point cloud attributes | |
Kavitha | A survey on lossless and lossy data compression methods | |
JP6242074B2 (ja) | 信号データの圧縮および圧縮解除のための方法および装置(信号データの圧縮および圧縮解除) | |
US9543980B2 (en) | Systems and methods for model-free compression and model-based decompression | |
US20130159263A1 (en) | Lossy compression of data points using point-wise error constraints | |
JPH06224777A (ja) | 符号化方法、符号化装置、復号方法、復号器、データ圧縮装置、ビットストリーム生成方法及び遷移マシン生成方法 | |
JP6246233B2 (ja) | 頂点誤差訂正のための方法および装置 | |
US10366698B2 (en) | Variable length coding of indices and bit scheduling in a pyramid vector quantizer | |
CN116156072B (zh) | 隐写图像生成方法、隐写信息提取方法及相关装置 | |
CN116128985A (zh) | 基于双分支神经网络的大规模点云几何压缩方法 | |
Balado et al. | Asymptotically optimum perfect universal steganography of finite memoryless sources | |
JP3603070B2 (ja) | 圧縮コード生成方法、および圧縮コード展開方法 | |
Maleika et al. | Lossless compression method for digital terrain model of seabed shape | |
Al-Azawi et al. | Compression of Audio Using Transform Coding. | |
Negi et al. | A review on fractal image compression | |
JP4694511B2 (ja) | 符号装置、復号装置、符号化復号化システム、符号化復号化方法、符号化プログラム、復号化プログラム、符号化復号化プログラム及び記録媒体 | |
WO2023132329A1 (ja) | 点群復号装置、点群復号方法及びプログラム | |
Nandi et al. | Fractal image compression using Fast Context Independent HV partitioning scheme and MRBH, MRBHM coding | |
WO2002058038A1 (fr) | Appareil, procede et programme de cryptage a compression arithmetique | |
Bakulina | Efficient image compression by coding of low-entropy sources | |
KR20240009371A (ko) | 메모리 장치, 데이터 압축 및 압축 해제 방법 | |
Gupta et al. | Adaptive Byte Compression and Decompression: A New Approach for Fractal Image Compression |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
WWE | Wipo information: entry into national phase |
Ref document number: 200780041260.1 Country of ref document: CN |
|
121 | Ep: the epo has been informed by wipo that ep was designated in this application |
Ref document number: 07831450 Country of ref document: EP Kind code of ref document: A1 |
|
WWE | Wipo information: entry into national phase |
Ref document number: 2008544115 Country of ref document: JP |
|
WWE | Wipo information: entry into national phase |
Ref document number: 2007831450 Country of ref document: EP |
|
WWE | Wipo information: entry into national phase |
Ref document number: 2667727 Country of ref document: CA |
|
WWE | Wipo information: entry into national phase |
Ref document number: 12446818 Country of ref document: US Ref document number: 2253/CHENP/2009 Country of ref document: IN |
|
WWE | Wipo information: entry into national phase |
Ref document number: 1020097008940 Country of ref document: KR |
|
ENP | Entry into the national phase |
Ref document number: 2009116946 Country of ref document: RU Kind code of ref document: A |
|
NENP | Non-entry into the national phase |
Ref country code: DE |
|
ENP | Entry into the national phase |
Ref document number: PI0718239 Country of ref document: BR Kind code of ref document: A2 Effective date: 20090505 |