JP2005252337A - Image coding apparatus, image coding method, and storage medium therefor - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an image coding apparatus capable of highly efficiently applying reversible compression to a pseudo gradation image at high speed, and to provide an image coding method and a storage medium therefor. <P>SOLUTION: The image coding apparatus includes: a reference pixel acquisition means 303 for acquiring a reference pixel in the vicinity of a pixel to be coded; a storage means 310 for storing a prediction state and a superiority symbol for each context being an incident pattern of the acquired reference pixel; a superiority symbol acquisition means 304 for acquiring the superiority symbol stored in the storage means 310 by using the context; a prediction conversion means 305 for predicting the coded pixel by using the superiority symbol and providing an output of hit/miss data sequence; a prediction transition table 308 for shifting the prediction state on the basis of the prediction transition table 308; and a run length coding means 306 for applying run length coding to the prediction hit/miss data. <P>COPYRIGHT: (C)2005,JPO&NCIPI

Description

  The present invention relates to an image encoding technique for compressing image data with high efficiency.

  Data compression to reduce storage capacity and transmission time of data input / output devices such as printers, scanners, data storage devices, network communication devices, etc., as printers, scanners, copiers, facsimiles, etc. increase in resolution and color. Technology is becoming important.

  General data used in data output devices such as printers and copiers is binary or multi-value pseudo gradation image data composed of 1 bit per pixel or N bits. As the pseudo gradation processing, there are a dither method, an error diffusion method, and the like. In the pseudo gradation image subjected to gradation conversion by such a method, ON / OFF change of dots frequently occurs. Therefore, in the case of lossless compression, simple run-length encoding, MH, MR, MMR, or the like cannot be said to have sufficient encoding efficiency, and efficient expansion of information sources is essential. In order to expand the information source and apply Huffman coding in units of blocks, it is necessary to use memory for reporting order control, overflow countermeasures, memory management, or stripe unit coding processing to separate and encode for each context. (For example, see Non-Patent Document 1).

  On the other hand, since arithmetic coding can be decoded by knowing the symbol allocation area for each context, the above-described notification order control and memory management processing are not required. For this reason, the conventional lossless compression for binary or multi-value pseudo gradation images is based on arithmetic coding as represented by JBIG, which is an international standardization method (see, for example, Non-Patent Document 1).

  A conventional technique in the lossless compression method of a pseudo gradation image using arithmetic coding will be described below.

  FIG. 1 is a block diagram showing the configuration of a conventional image encoding device, and FIG. 2 is an explanatory diagram showing the flow of processing in the conventional image encoding device.

  As shown in FIG. 1, the pseudo gradation image encoding apparatus includes a control unit 101 that controls operation of the apparatus, an input unit 102 to which image data is input, a reference pixel acquisition unit 103, a dominant symbol acquisition unit 104, an appearance probability. Acquisition unit 105, MPS / LPS exchange unit 106, register update unit 107, prediction state transition unit 108, normalization unit 109, prediction transition table 110, encoded data output unit 111, prediction state and register value for each context, or image It has a storage means 112 in which data and control programs are stored.

  With respect to the image encoding apparatus configured as described above, the flow of processing will be described with reference to FIG.

  First, the control means 101 initializes an A register that represents the interval length of arithmetic coding and a C register that represents the start point. For example, when arithmetic encoding is performed with 16-bit precision, the A register is composed of 16 bits, the C register is composed of 32 bits, and the A register is initialized to 0xFFFF and the C register is initialized to 0x00000000. Also, the prediction state is initialized. Here, 0x indicates that the following numerical value is in hexadecimal notation. Since the prediction state needs to be stored for each context, when there are 1024 contexts, the prediction state for each context is initialized to zero. Thereafter, the input image data is received from the image data input means 102 and stored in the storage means 112.

  Next, the reference pixel acquisition unit 103 acquires reference pixels in the vicinity of the encoding target pixel from the image data stored in the storage unit 112 (S201). Then, the dominant symbol acquisition unit 104 refers to the prediction state stored in the storage unit 112 based on the acquired appearance pattern of the reference pixel (hereinafter referred to as “context”), and predicts the encoded pixel. A superior symbol (MPS) is acquired (S202). Further, the appearance probability acquisition unit 105 refers to the register update unit 107 based on the context, and acquires the appearance probability (Amps) when the encoded pixel is the dominant symbol and the appearance probability (Alps) when the encoded pixel is the inferior symbol ( S203).

  Next, when the appearance probability of the dominant symbol acquired by the appearance probability acquisition unit 105 becomes smaller than the appearance probability of the inferior symbol, the MPS / LPS exchanging unit 106 switches the appearance probability of the dominant symbol and the inferior symbol (S204). ). Then, the register updating unit 107 updates the register value based on the appearance probability of the dominant symbol and the inferior symbol and whether or not the encoded pixel is the dominant symbol (S205). When the encoded pixel is a dominant symbol (MPS), the appearance probability (Amps) of the dominant symbol is substituted into the A register, and when the encoded pixel is the inferior symbol (LPS), the appearance probability (Amps) of the dominant symbol is stored in the C register. And the appearance probability (Alps) of the inferior symbol is substituted into the A register.

  When the updated value of the A register is smaller than half of the initial value (A <0x8000 when the initial value is 0xFFFF), the prediction state transition unit 108 determines whether the prediction transition table 110 and the encoded pixel are the dominant symbols. Based on this, a prediction state transition is performed (S206). Further, the normalizing means 109 normalizes the A register (S207). That is, while the value of the A register is smaller than half of the initial value (A <0x8000 when the initial value is 0xFFFF), the left 1-bit shift is repeated with respect to the values of the A register and the C register. Then, the bits overflowing from the lower 16 bits of the C register are output from the output unit 111 as encoded bits.

Note that these operations may be controlled by a control program stored in the storage unit 112.
Hiroshi Watanabe, Fumitaka Ono, “Basic Technology of International Standard Image Coding”, Corona, March 20, 1998, p58-65.

  However, the above-described conventional technique has a problem that a high-speed clock is required to realize hardware with a high processing speed because it is necessary to perform bit operations on a pixel basis.

  Therefore, an object of the present invention is to provide an image encoding technique capable of performing reversible compression of a pseudo gradation image at high speed and high efficiency.

In order to solve this problem, an image encoding device according to the present invention includes a reference pixel acquisition unit that acquires reference pixels in the vicinity of an encoded pixel, and a prediction state and a superiority for each context that is an appearance pattern of the acquired reference pixel. Prediction state storage means for storing symbols, dominant symbol acquisition means for acquiring a dominant symbol stored in the prediction state storage means using a context, prediction of an encoded pixel using the dominant symbol, and a success / failure data series thereof Predictive conversion means for outputting, prediction transition table for transitioning the prediction state based on the prediction success / failure, prediction state transition means for transitioning the prediction state based on the prediction transition table, and the run length code for the prediction success / failure data And a run-length encoding means for converting the data into the run-length encoding means. As a result, adaptive prediction based on the state of surrounding pixels is performed, and the prediction success / failure data is run-length encoded, so that highly accurate prediction conversion can be realized, and more than one pixel can be encoded at a time. And high-speed and high-efficiency lossless compression can be performed on the pseudo gradation image.

  In a preferred embodiment of the present invention, a periodic pixel acquisition unit that acquires a pixel having a high correlation with an encoded pixel at a constant period is provided, and the reference pixel of the reference pixel acquisition unit is a floating pixel that can be replaced with another pixel. In addition, the values of the pixels obtained by the periodic pixel acquisition unit and the floating pixels are exchanged. As a result, by determining the context using periodic pixels, it is possible to realize high-precision predictive conversion even for image patterns having periodicity such as screens and hatch patterns, and with high speed and high efficiency. Reversible compression is possible.

  In a further preferred aspect of the present invention, the encoding method comprises: run length storage means for storing a run length that is the length of a run that has just appeared; and encoding method switching means for switching a method for encoding the run length. The switching means switches the encoding method based on the run length stored in the run length storage means. As a result, encoding suitable for the target run length can be realized by switching the encoding method (for example, encoding table) in the run length encoding based on the run length that appears in the vicinity. Highly efficient lossless compression is possible.

  In a further preferred aspect of the present invention, the method further comprises run-length appearance frequency counting means for counting the appearance frequency of run-lengths appearing within a certain section, and encoding method switching means for switching a method for encoding run-length. The method switching means switches the encoding method to be applied to the next section based on the appearance frequency counted by the run length appearance frequency counting means. As a result, encoding suitable for the target run length can be realized by switching the encoding method (for example, encoding table) in the run length encoding on the basis of the run length appearing in a certain fixed neighborhood. High-speed and high-efficiency lossless compression.

  In a further preferred embodiment of the present invention, when the floating pixel is changed in the periodic pixel acquisition means, the reference position change code indicating the change position of the floating reference pixel and whether or not the change has occurred during the continuation of the run Reference pixel change code generating means for generating Thereby, the reference pixel can be changed during the continuation of the run by having the data indicating whether the run is continuation or not in the reference pixel change code.

  In a further preferred aspect of the present invention, when the run length reaches a preset run length, the reference pixel change code is generated by the reference pixel change code generating means. This makes it possible to generate a run length that does not exceed the maximum value of a run length code table prepared in advance.

  In a further preferred form of the present invention, the prediction state storage means has n × m sets of prediction states and dominant symbols for each context (n and m are integers), and the reference pixel acquisition means encodes n codes. Reference pixels other than floating pixels generate the same context for the pixel unit, and predictive conversion of n × m pixels is performed at the same time. As a result, high-speed predictive conversion can be realized by simultaneously performing predictive conversion of n × m pixels.

In a further preferred embodiment of the present invention, a decoding speed prediction means for performing a decoding speed prediction of a code based on the number of prediction errors within a certain interval with respect to the prediction success / failure data output from the prediction conversion means, and a predetermined image line unit And encoding method switching means for determining whether the raster code output data is run-length encoded data or input data itself from the integrated decoding speed value output from the decoding speed prediction means. As a result, it is possible to perform encoding that provides a desired decoding speed during encoding.

  In order to solve this problem, an image encoding method according to the present invention stores a reference pixel acquisition step for acquiring a reference pixel in the vicinity of an encoded pixel, and a prediction state and a dominant symbol for each context of the acquired reference pixel. A prediction state storage step, a dominant symbol acquisition step of acquiring the dominant symbol stored in the prediction state storage step using the context, and a prediction for performing prediction of the encoded pixel using the dominant symbol and outputting the success / failure data sequence A conversion step, a prediction transition table for transitioning the prediction state based on prediction success / failure, a state transition step for performing a transition of the prediction state based on the prediction transition table, and a run length for performing run-length encoding on the prediction success / failure data And an encoding step. As a result, adaptive prediction based on the state of surrounding pixels is performed, and the prediction success / failure data is run-length encoded, so that highly accurate prediction conversion can be realized, and more than one pixel can be encoded at a time. Therefore, high-speed and high-efficiency lossless compression can be performed on a pseudo gradation image.

  In a preferred embodiment of the present invention, the pixel acquisition step includes a periodic pixel acquisition step of acquiring pixels having a high correlation with the encoded pixel at a constant cycle, and the reference pixel of the reference pixel acquisition step includes a floating pixel and is obtained by the periodic pixel acquisition step. The value of the selected pixel and the value of the floating pixel are exchanged. As a result, by determining the context using periodic pixels, it is possible to realize high-precision predictive conversion even for image patterns having periodicity such as screens and hatch patterns, and with high speed and high efficiency. Reversible compression is possible.

  In a further preferred aspect of the present invention, the method includes a run length storing step for storing a run length that is the length of a run that has just appeared, and an encoding method switching step for switching a method for encoding the run length. In the method switching step, the encoding method is switched based on the run length stored in the run length storing step. As a result, encoding suitable for the target run length can be realized by switching the encoding method (for example, encoding table) in the run length encoding based on the run length that appears in the vicinity. Highly efficient lossless compression is possible.

  In a further preferred aspect of the present invention, there is provided a run length appearance frequency counting step for counting the appearance frequency of run lengths that appear in a certain section, and an encoding method switching step for switching a method for encoding the run length, The encoding method switching step switches the encoding method applied to the next section based on the appearance frequency counted by the run length appearance frequency counting step. As a result, encoding suitable for the target run length can be realized by switching the encoding method (for example, encoding table) in the run length encoding on the basis of the run length appearing in a certain fixed neighborhood. High-speed and high-efficiency lossless compression.

  In order to solve this problem, a storage medium of the present invention stores a control program describing the above-described image encoding method. Thereby, the reversible compression of the pseudo gradation image can be performed at high speed and high efficiency.

  According to the present invention, the predicted value is acquired by the state transition based on the context of the surrounding pixel, and the run-length encoding is performed on the prediction data whether or not the prediction value is correct. Prediction can be performed, multiple pixels can be encoded at once, and parallel processing can be realized without a significant increase in gate amount, resulting in high speed and high efficiency. An effective effect that it is possible to perform reversible compression of the pseudo gradation image is obtained.

  According to the first aspect of the present invention, reference pixel acquisition means for acquiring a reference pixel in the vicinity of an encoded pixel, and prediction for storing a prediction state and a dominant symbol for each context that is an appearance pattern of the acquired reference pixel A state storage means, a dominant symbol acquisition means for acquiring a dominant symbol stored in the prediction state storage means using the context, and a predictive transformation for performing prediction of the encoded pixel using the dominant symbol and outputting the correct / incorrect data series Means, a prediction transition table for transitioning the prediction state based on the prediction success / failure, a prediction state transition means for performing a transition of the prediction state based on the prediction transition table, and a run length code for performing run length encoding on the prediction success / failure data An image encoding device characterized by comprising: an encoding means for performing adaptive prediction based on the state of surrounding pixels, Since the data is run-length encoded, high-precision predictive conversion can be realized, and more than one pixel can be encoded at a time. Has the effect of being able to

  The invention according to claim 2 of the present invention is the invention according to claim 1, further comprising periodic pixel acquisition means for acquiring pixels having a high correlation with the encoded pixel at a constant period, and the reference pixel of the reference pixel acquisition means is An image encoding device including a floating pixel that is a pixel that can be replaced with another pixel and replacing the value of the pixel obtained by the periodic pixel obtaining unit and the floating pixel, using the periodic pixel By determining the context, it is possible to realize highly accurate predictive conversion even for an image pattern having periodicity such as a screen or a hatch pattern, and to perform high-speed and high-efficiency reversible compression.

  According to a third aspect of the present invention, in the first or second aspect of the present invention, run-length storage means for storing a run-length that is the length of a run that appears immediately before, and a method for encoding the run-length Encoding method switching means for switching between, and the encoding method switching means is an image coding device characterized in that the coding method is switched based on the run length stored in the run length storage means. By switching the encoding method (for example, encoding table) in run-length encoding based on the run length that appears, encoding suitable for the target run-length can be realized, and high-speed, high-efficiency reversible It has the effect that it can be compressed.

  According to a fourth aspect of the present invention, in the first or second aspect of the present invention, a run-length appearance frequency counting means for counting the appearance frequency of a run length that appears in a predetermined section, and the run length are encoded. And a coding method switching means for switching a method, wherein the coding method switching means switches a coding method to be applied to the next section based on the appearance frequency counted by the run length appearance frequency counting means. An image encoding device that performs encoding suitable for a target run length by switching an encoding method (for example, an encoding table) in the run length encoding based on a run length that appears in a certain fixed neighborhood. It can be realized and has an effect that high-speed and high-efficiency lossless compression can be performed.

  According to a fifth aspect of the present invention, in the second aspect of the present invention, when the floating pixel is changed in the periodic pixel acquisition means, the change position of the floating reference pixel and the change are continuously performed. An image encoding apparatus comprising a reference pixel change code generating means for generating a reference pixel change code indicating whether or not a run has occurred, wherein data indicating whether or not a run is continuing is referred to as a reference pixel change code By having it inside, it has the effect | action that a reference pixel can be changed during the continuation of a run.

The invention described in claim 6 of the present invention is characterized in that, in the invention described in claim 5, when the run length reaches a preset run length, the reference pixel change code generating means generates the reference pixel change code. This is an image encoding device that can generate a run length that does not exceed the maximum value of a run-length code table prepared in advance.

  In the invention according to claim 7 of the present invention, in the invention according to claim 2, the prediction state storage means has n × m sets (n and m are integers) storing the prediction state and the dominant symbol for each context. In the reference pixel acquisition unit, the reference pixels other than the floating pixels generate the same context for n encoded pixel units, and simultaneously perform predictive conversion of n × m pixels. It has the effect that high-speed predictive conversion can be realized by simultaneously performing predictive conversion of n × m pixels.

  According to an eighth aspect of the present invention, in the first or second aspect of the present invention, the decoding speed of the code based on the number of prediction errors within a certain interval with respect to the prediction success / failure data output from the prediction conversion means. Decoding speed prediction means for performing prediction, and code output data of the raster as run-length encoded data or input data itself from the integrated value of decoding speeds output from the decoding speed prediction means in predetermined image line units And an encoding method switching means for performing the encoding. The image encoding device has an operation of performing encoding to obtain a desired decoding speed at the time of encoding.

  The invention according to claim 9 of the present invention includes a reference pixel acquisition step of acquiring a reference pixel in the vicinity of an encoded pixel, and a prediction state storage step of storing a prediction state and a dominant symbol for each context of the acquired reference pixel. A dominant symbol acquisition step for acquiring the dominant symbol stored in the prediction state storage step using the context; a prediction conversion step for performing prediction of the encoded pixel using the dominant symbol and outputting the correct / incorrect data sequence; A prediction transition table for transitioning the prediction state based on the success / failure, a state transition step for performing the transition of the prediction state based on the prediction transition table, and a run-length encoding step for performing run-length encoding on the prediction success / failure data An image encoding method characterized by having adaptive prediction based on the state of surrounding pixels. In addition, by performing run-length encoding on the prediction success / failure data, high-precision prediction conversion can be realized, and more than one pixel can be encoded at a time. It has the effect of performing highly efficient lossless compression.

  The invention described in claim 10 of the present invention is the invention described in claim 9, further comprising a periodic pixel acquisition step of acquiring pixels having a high correlation with the encoded pixel at a constant cycle, and the reference pixel of the reference pixel acquisition step Is an image encoding method that includes floating pixels and replaces the values of the pixels obtained by the periodic pixel acquisition step and the floating pixels. By determining the context using the periodic pixels, Highly accurate predictive conversion can be realized even for an image pattern having a periodicity such as a hatch pattern, and high-speed and high-efficiency lossless compression can be performed.

  According to an eleventh aspect of the present invention, in the invention according to the ninth or tenth aspect, a run length storing step for storing a run length that is the length of a run that appears immediately before, and a method for encoding the run length A coding method switching step for switching between the coding methods, wherein the coding method switching step switches the coding method based on the run length stored in the run length storage step. By switching the encoding method (for example, encoding table) in run-length encoding based on the run length that appears, encoding suitable for the target run-length can be realized, and high-speed, high-efficiency reversible It has the effect that it can be compressed.

The invention according to claim 12 of the present invention is the invention according to claim 9 or 10, wherein the run-length appearance frequency counting step for counting the appearance frequency of run-lengths appearing within a certain section and the run-length are encoded. And a coding method switching step for switching a method, wherein the coding method switching step switches a coding method to be applied to the next section based on the appearance frequency counted by the run length appearance frequency counting step. This is an image encoding method, and by switching the encoding method (for example, encoding table) in the run length encoding based on the run length appearing in a certain fixed neighborhood, encoding suitable for the target run length is performed. It can be realized and has the effect of being able to perform high-speed, high-efficiency lossless compression.

  A thirteenth aspect of the present invention is a storage medium storing a control program that describes the image encoding method according to the ninth, tenth, eleventh, or twelfth aspect, and is high-speed and high-efficiency. This has the effect that the pseudo gradation image can be reversibly compressed.

  Hereinafter, the best mode for carrying out the present invention will be described more specifically with reference to the drawings. In addition, since description here is the best form by which this invention is implemented, this invention is not limited to the said form.

(Embodiment 1)
FIG. 3 is a block diagram showing the configuration of the image coding apparatus according to Embodiment 1 of the present invention, and FIG. 4 is an explanatory diagram showing the flow of processing in the image coding apparatus according to Embodiment 1 of the present invention.

  As shown in FIG. 3, the image coding apparatus according to the present embodiment includes a control unit 301 that controls processing for processing image data, an input unit 302 to which image data is input, a reference pixel acquisition unit 303, and a dominant symbol. Acquisition unit 304, prediction conversion unit 305, run-length encoding unit 306, prediction state transition unit 307, prediction state transition table 308, output unit 309, storage unit storing a prediction state and image data for each context, or a control program (Predicted state storage means) 310 is provided.

  With respect to the image coding apparatus configured as described above, the flow of processing will be described with reference to FIG.

  First, the control means 301 initializes the run length to zero. Also, the prediction state is initialized. Since the prediction state needs to be stored for each context, when there are 1024 contexts, the prediction state for each context is initialized to zero. The input image data is received from the input unit 302 and stored in the storage unit 310.

  Next, the reference pixel acquisition unit 303 acquires reference pixels in the vicinity of the encoding target pixel from the image data stored in the storage unit 310 (S401). Then, the dominant symbol acquisition unit 304 refers to the prediction state stored in the storage unit 310 based on the appearance pattern (context) of the acquired reference pixel, and acquires the dominant symbol when predicting the encoded pixel ( S402). Further, the predictive conversion unit 305 predicts the encoded pixel using the acquired dominant symbol and outputs a prediction success / failure data series (S403).

  The run-length encoding means 306 increments the run length in the case of per prediction from the prediction success / failure data, and in the case of a prediction failure, the run-length encodes the run length and resets the run length to 0 (S404). When run-length encoding is performed, the encoded data is output from the output unit 309. Finally, the prediction state transition unit 307 performs prediction state transition based on the prediction transition table 308 and whether the encoded pixel is a dominant symbol (S405).

  Note that these operations may be controlled by a control program stored in the storage unit 310.

  The reference pixel acquisition unit 303 acquires pixels near the encoding target pixel in order to perform prediction conversion. In that case, only pixels on the same line may be referred to, or pixels on a plurality of lines including the previous line may be referred to. For example, when 10 pixels are referred to, there are 1024 (2 to the 10th power) appearance patterns (contexts) of reference pixels. In addition, the reference pixel may include a pixel (floating pixel) that can be replaced with one or more other pixels.

  FIG. 5 is an explanatory diagram illustrating an example of the reference pixel and the floating pixel. FIG. 6 is an explanatory diagram showing an example of the prediction transition table. Normally, 10 pixels in addition to the floating pixels are used as the reference pixels in the 9 pixels shown in FIG. 5. However, in the case of an image having a high correlation with the encoded pixels in an 8-pixel cycle, the floating pixels are shown in FIG. It replaces and uses a pixel with high correlation with an encoding pixel. By doing so, for example, highly efficient encoding can be realized even for repetitive patterns with high periodicity such as dither patterns and hatch patterns.

  The dominant symbol acquisition unit 304 acquires the dominant symbol stored in the storage unit 310 using the context. When there are 1024 contexts, the prediction state and the dominant symbol are stored for each context, and the stored dominant symbol is acquired.

  The run-length encoding unit 306 performs run-length encoding on the success / failure data series by the predictive conversion unit 305. At this time, several run lengths appearing in the vicinity are memorized, and when run lengths appearing in the vicinity are long on average, a code table suitable for long runs is used, and the run lengths appearing in the periphery are averaged. If the length is short, it is possible to apply run-length coding using a code table suitable for short runs. Further, as run length coding, Huffman coding may be considered, and systematic codes such as Golomb coding and Weil coding may be considered.

  The prediction state transition unit 307 refers to the prediction transition table 308 and performs state transition of the prediction state for each context. The prediction transition table is configured as shown in FIG. 6 and defines which state is to be changed next based on whether the encoded pixel is the dominant symbol or the inferior symbol. The number of the predicted state for each context, that is, the current state is stored in the storage unit 310. For example, if the current prediction state is state number 13 and a dominant symbol appears, the prediction state for that context transitions to number 29. The storage unit 310 also stores a dominant symbol for each context. Then, the dominant symbol is updated together with the transition of the prediction state. The state in which the value of the dominant symbol is inverted is defined in the prediction transition table as shown in FIG. When the inversion item is 1, the dominant symbol is inverted. In addition, although an example of a table having 47 states as the prediction transition table is shown here, it goes without saying that the number of states is not limited to 47.

  FIG. 7 is a block diagram showing a hardware configuration for realizing the image coding apparatus according to Embodiment 1 of the present invention. FIG. 8 is a block diagram showing a modification of the hardware configuration for realizing the image coding apparatus according to Embodiment 1 of the present invention.

  As shown in FIG. 7, the image encoding apparatus includes an external storage device 701, a central processing unit (hereinafter referred to as “CPU”) 702, a read-only memory (hereinafter referred to as “ROM”) 703, and a random access memory. (Hereinafter referred to as “RAM”) 704, which are coupled to each other via a bus. The ROM 703 has a program storage area, and the RAM 704 has an image storage area.

FIG. 7 is a hardware configuration diagram of an embedded device. When this embodiment is realized by a general-purpose device, there may be a program storage area in the external storage device 801 as shown in FIG. Here, examples of the external storage device include a hard disk, a floppy (registered trademark) disk, a CD-ROM, and an MO. The external storage device may be connected by a network or the like.

  FIG. 9 is a flowchart for realizing the image coding technique according to Embodiment 1 of the present invention. Hereinafter, the flow of processing will be described with reference to FIG.

  In FIG. 9, after image data stored in the external storage device 701 or the like or a prediction transition table stored in the ROM 703 is input to the RAM 704, the pixel number is set to 0, the run length is set to 0, and the context is set for each context. The prediction state (Cx) is initialized to 0, respectively. Next, in S902, reference pixels near the encoding target pixel are acquired. The on / off appearance pattern of the reference pixel is the context Cx. In step S903, the dominant symbol MPS is acquired based on the context Cx and the prediction state (Cx) stored in the RAM 704. In S904, the encoded pixel and the dominant symbol are compared, and if the encoded pixel matches the dominant symbol, the process proceeds to S905. If not, the process proceeds to S906. In S905, the run length is incremented. On the other hand, in S906, the run length is run length encoded, and in S907, the run length is reset to zero. In step S908, the state transition of the prediction state (Cx) is performed using the prediction transition table stored in the RAM 704. Thereafter, in S909, the number of pixels is compared with the pixel number. If the pixel number is smaller, the process proceeds to S910, and if not, the process proceeds to S911. In S910, the pixel number is incremented, the process returns to S902, and the processing from S902 to S910 is repeated until the processing is completed for all the pixels. On the other hand, in S911, the run length of the last run is run-length encoded. Also, the run-length encoded data is stored in the external storage device 701 or the like or transmitted to the outside via a network. The above processing is performed using the CPU 702.

  FIG. 10 is an explanatory diagram showing a memory map of a storage medium that stores various data processing programs that can be read by the print processing apparatus according to the first embodiment of the present invention.

  Although not particularly illustrated, information for managing a program group stored in the storage medium, for example, version information, information such as a creator, and information depending on the OS on the program reading side, for example, an icon for identifying and displaying the program May also be stored.

  Further, data depending on various programs is also managed in the directory. In addition, a program for installing various programs in the computer, and a program for decompressing when the program to be installed is compressed may be stored.

  The functions shown in FIG. 9 in the first embodiment may be performed by a host computer by a program installed from the outside. In this case, the present invention is applied even when an information group including a program is supplied to the output device from a storage medium such as a CD-ROM, a flash memory, or an FD, or from an external storage medium via a network. Is.

  As described above, a storage medium storing software program codes for realizing the functions of the above-described embodiments is supplied to a system or apparatus, and the computer (or CPU or MPU) of the system or apparatus is stored in the storage medium. It goes without saying that the object of the present invention can also be achieved by reading and executing the program code.

In this case, the program code itself read from the storage medium realizes the novel function of the present invention, and the storage medium storing the program code constitutes the present invention.

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

  Further, according to the present invention, not only when the functions of the above-described embodiments are realized by executing the program code read by the computer, but the computer is operating on the basis of the instruction of the program code. It goes without saying that the case where the OS (operating system) or the like performs part or all of the actual processing and the functions of the above-described embodiments are implemented by the processing.

  Further, according to the present invention, a program code read from a storage medium is written in a memory provided in a function expansion board inserted into the computer or a function expansion unit connected to the computer, and then the program code is instructed. Based on this, it goes without saying that the CPU of the function expansion board or function expansion unit performs part or all of the actual processing and the functions of the above-described embodiments are realized by the processing.

(Embodiment 2)
FIG. 11 is a block diagram showing the overall configuration of the image coding apparatus according to Embodiment 2 of the present invention.

  As shown in FIG. 11, the image coding apparatus 1101 of the present embodiment includes a reference pixel acquisition unit 1102, a periodic pixel acquisition unit 1103, a dominant symbol acquisition unit 1104, a prediction state storage unit 1105, a prediction state transition unit 1106, A prediction transition table 1107, a prediction conversion unit 1108, a run length encoding unit 1109, a reference pixel change code generation unit 1110, a decoding speed prediction unit 1111, and an encoding scheme switching unit 1112 are provided.

  Encoding is started by first obtaining a reference pixel from an image input. Here, FIG. 12 is an explanatory diagram showing reference pixels acquired by the reference pixel acquisition unit according to Embodiment 2 of the present invention. In FIG. 12, 6 reference pixels including floating reference pixels f1 and f2 for encoding target pixels T1 and T2 and 6 reference pixels including floating reference pixels f3 and f4 for encoding target pixels T3 and T4 are simultaneously acquired. It shows how they are doing. In FIG. 12, for the encoding target pixels T1 and T2, the four pixels immediately before the encoding target pixel T1 are acquired for the encoding target pixels T1 and T2, and the floating reference pixel f1 is further converted to the encoding target pixel T1. In addition, the floating reference pixel f2 is acquired for the encoding target pixel T2. For the encoding target pixels T3 and T4, the four pixels immediately before the encoding target pixel T3 are acquired for the encoding target pixels T3 and T4, and the floating reference pixel f3 is further changed to the encoding target pixel T3. The floating reference pixel f4 is acquired for the encoding target pixel T4.

  In Embodiment 2 of the present invention, reference pixels are simultaneously acquired for the four pixels to be encoded by the reference pixel acquisition unit 1102. Two sets of two encoding target pixels are acquired at the same time, and prediction conversion is performed simultaneously for four pixels, thereby realizing high-speed encoding. At the time of decoding the code, as can be seen from FIG. 12, since the encoding target pixels T1 and T2 are included in the reference pixels of the encoding target pixels T3 and T4, two pixels are used instead of the four-pixel simultaneous processing. Is the decryption process.

Further, in the reference pixel acquisition unit 1102, floating reference pixels f1, f2, f3,
f4 is acquired, and the floating pixel position is determined by performing periodic pixel position acquisition by the periodic pixel acquisition unit 1103. Here, the flow of acquisition of the periodic pixel position of the periodic pixel acquisition unit 1103 will be described with reference to FIG. FIG. 13 is an explanatory diagram showing the periodic pixel position of the periodic pixel acquisition means in the second embodiment of the present invention.

  In the image coding apparatus according to the second embodiment of the present invention, each block and coding pixel for a block of 8 pixels in which the relative distance of the pixels ahead of the coding pixel is 8 pixels is 1 to 32. A determination is made as to whether or not the eight pixels match, and a counter that adds 1 if there is a match for each relative distance is provided. When the counter value reaches a certain set value, the relative distance is determined as a periodic pixel. The distance is reset, and the counters of all relative distance blocks are reset. The periodic pixel distance is then held until the counter value of any relative distance block reaches a certain set value. A pixel having a periodic pixel distance obtained in this way is acquired as a periodic pixel. In FIG. 13, R8 indicates a pixel with a relative distance of 8, R9 indicates a pixel with a relative distance of 9, and R10 indicates a pixel with a relative distance of 10. The pixel position at the periodic pixel distance obtained in this way is the floating pixel position.

  FIG. 14 is a block diagram showing the predicted state storage means in the second embodiment of the present invention. The prediction state storage unit 1105 has a prediction state storage unit T1, a prediction state storage unit T2, and a prediction state storage unit T3 for each of the four encoding target pixels that have been acquired by the reference pixel acquisition unit 1102. And four prediction state storage units of the prediction state storage unit T4. FIG. 15 shows an internal configuration of each prediction state storage unit T1 to T4. The prediction state storage units T1 to T4 are configured by 32 words by a word composed of 1-bit dominant symbol value and 6-bit transition state, and the word address is a context value composed of an ON / OFF appearance pattern of 5 reference pixels. Correspondingly, the initial value of the transition state of each word is 0.

  The dominant symbol acquisition unit 1104 uses the on / off appearance pattern of the five reference pixels for each encoding target pixel obtained by the reference pixel acquisition unit 1102 as a context value in the prediction state storage units T1 to T4 of the prediction state storage unit 1105. Give the word address and get the dominant symbol value. Using the dominant symbol value obtained here, the predictive conversion means 1108 outputs “0” as the predictive conversion value when the dominant symbol value matches the encoding target pixel value, and encodes the dominant symbol value and the encoding value. If the target pixel values do not match, “1” is output as the predicted conversion value.

  The prediction state transition unit 1106 generates the prediction transition table 1107 storing the values shown in FIG. 6 and the transition state value and the dominant symbol value obtained from the prediction state storage unit 1105 and the reference pixel obtained by the reference pixel acquisition unit 1102. The transition state value and the dominant symbol value in the prediction state storage units T1 to T4 are updated from the context value. The prediction transition table 1107 has a configuration as shown in FIG. 6 and defines which state is to be changed next based on whether the encoded pixel is a dominant symbol or an inferior symbol. The predicted state for each context, that is, the number at which the current state is located is stored in the predicted state storage unit 1105.

  FIG. 15 is an explanatory diagram showing the prediction state storage unit in the second embodiment of the invention, FIG. 16 is an explanatory diagram showing the run-length code table in the second embodiment of the present invention, and FIG. FIG. 18 is an explanatory diagram showing a reference pixel change code in Embodiment 2 of the invention, and FIG. 18 is an explanatory diagram showing a decoding speed prediction table in Embodiment 2 of the present invention.

As shown in FIG. 15, the prediction state storage means 1105 also stores a dominant symbol value for each context, and the dominant symbol value is updated with the transition of the prediction state. The state in which the dominant symbol value is inverted is also defined in the prediction transition table as shown in FIG. 6. When the item of inversion is 1, the dominant symbol is inverted. For example, the current prediction state is state number 14
When the dominant symbol value matches the encoding target pixel value, the prediction state for the context transitions to No. 15. When the dominant symbol value and the encoding target pixel value do not match, the prediction state remains No. 14, but since 1 is indicated in the item of inversion, the prediction state storage unit 1106 corresponds to the prediction state storage. Updates such as inverting the dominant symbol value of the context.

  In the image encoding apparatus according to the second embodiment of the present invention, the run-length encoding unit 1109 performs encoding using a predictive conversion value from the predictive conversion unit 1108 as an input and a wyle code that is a systematic run-length code. The wyle code is generated by using as a code a continuous number of prediction conversion values “0” between the prediction conversion value “1” and the next prediction conversion value “1”. FIG. 16 shows the wyle codes corresponding to the respective run lengths. The code is a binary display. A value indicated by * below the wylle code is assigned a value obtained by binary conversion of a value obtained by subtracting the lower limit of the range corresponding to the run length from the run length.

  In addition, the run-length encoding unit 1109 of the image encoding apparatus according to the second embodiment of the present invention can generate run lengths up to 16383 as the wheel code as shown in FIG. 16, but the run-length value is set to a preset run-length value. When it reaches, the run length up to 8 encoding target pixels at that time is output as a wile code, a reference pixel change code generation means 1110 generates a dummy reference pixel change code, resets the run length, and returns to zero.

  The reference pixel change code generation unit 1110 performs run-length encoding up to 8 pixels to be encoded when the periodic pixel distance acquired by the periodic pixel acquisition unit 1103 changes, and the run-length encoding is performed. Even when the predicted conversion value “1” at the end is not the last pixel of the 8 pixels to be encoded, the run length up to that is output as a wile code. Thereafter, the reference pixel change code shown in FIG. 17 is generated. At this time, the code is a binary display. Of the lower 6 bits indicated by * of the reference pixel change code, 1 bit is the last predicted conversion value of the 8 pixels to be encoded, and the remaining 5 bits are assigned a value of (periodic pixel distance value-1). . In addition, when the run-length encoding unit 1109 notifies the reference pixel change code generation unit 1110 that the run length has reached a preset run-length value, the reference pixel change code is not changed. A reference pixel change code is generated using data of the current reference pixel position.

  The decoding speed prediction unit 1111 is a unit that determines whether or not code generation is performed so that decoding is performed at a desired decoding speed, and predicts the decoding speed when a real-time decoding process is required. In the decoding speed prediction means 1111, the prediction conversion values sent from the prediction conversion means 1108 are put together for every 64 pixels to perform decoding speed prediction. The decoding speed prediction is obtained from the sum of the prediction conversion values “1” in 64 pixels, and a prediction decoding speed as shown in FIG. 18 is obtained. For example, when the number of predictive conversion values “1” is 44, the predictive decoding speed 11 is obtained. This is a value obtained by multiplying the actual predictive decoding speed by eight, and the actual decoding speed is 1.375 pixels. This is the prediction decoding speed of 64 pixel blocks. The decoding speed integrated value obtained by integrating the decoding speed is output.

  The encoding method switching unit 1112 outputs the output of the run-length encoding unit 1109 if the integrated decoding rate output from the decoding rate prediction unit 1111 is equal to or greater than a preset value at the end of one line of the image. The code is output as a code output. If the value is equal to or less than the set value, the input image data itself is output as the code of the line to ensure the decoding speed.

  In this way, image coding using the image coding apparatus according to the second embodiment of the present invention is performed.

As described above, a storage medium storing software program codes for realizing the functions of the above-described embodiments is supplied to a system or apparatus, and the computer (or CPU or MPU) of the system or apparatus is stored in the storage medium. It goes without saying that the object of the present invention can also be achieved by reading and executing the program code.

  In this case, the program code itself read from the storage medium realizes the novel function of the present invention, and the storage medium storing the program code constitutes the present invention.

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

  Further, according to the present invention, not only when the functions of the above-described embodiments are realized by executing the program code read by the computer, but the computer is operating on the basis of the instruction of the program code. It goes without saying that the case where the OS (operating system) or the like performs part or all of the actual processing and the functions of the above-described embodiments are implemented by the processing.

  Further, according to the present invention, a program code read from a storage medium is written in a memory provided in a function expansion board inserted into the computer or a function expansion unit connected to the computer, and then the program code is instructed. Based on this, it goes without saying that the CPU of the function expansion board or function expansion unit performs part or all of the actual processing and the functions of the above-described embodiments are realized by the processing.

  As described above, the present invention is an image encoding technique for reversibly compressing a pseudo gradation image, and includes reference pixel acquisition means for acquiring a reference pixel in the vicinity of the encoded pixel, and an appearance pattern of the acquired reference pixel. Prediction state storage means for storing the prediction state and the dominant symbol for each context, predominance symbol acquisition means for acquiring the preferential symbol stored in the prediction state storage means using the context, and coding pixel using the preferential symbol Prediction conversion means for performing prediction and outputting the success / failure data series, prediction transition table for transitioning the prediction state based on prediction success / failure, and prediction state transition means for performing transition of the prediction state based on the prediction transition table And run-length encoding means for performing run-length encoding on the prediction success / failure data. Measurement and run-length encoding the prediction success / failure data, so that highly accurate prediction conversion can be realized, and more than one pixel can be encoded at a time. Can be compressed.

The block diagram which shows the structure of the conventional image coding apparatus. Explanatory drawing which shows the flow of a process in the conventional image coding apparatus. 1 is a block diagram showing a configuration of an image encoding device according to Embodiment 1 of the present invention. Explanatory drawing which shows the flow of a process in the image coding apparatus in Embodiment 1 of this invention. Explanatory drawing which shows an example of a reference pixel and a floating pixel Explanatory drawing which shows an example of a prediction transition table 1 is a block diagram showing a hardware configuration for realizing an image encoding device according to Embodiment 1 of the present invention. FIG. 3 is a block diagram showing a modification of the hardware configuration for realizing the image coding apparatus according to Embodiment 1 of the present invention. Flowchart for realizing the image coding technique in Embodiment 1 of the present invention Explanatory drawing which shows the memory map of the storage medium which stores the various data processing program which can be read by the printing processing apparatus which concerns on Embodiment 1 of this invention. FIG. 2 is a block diagram showing the overall configuration of an image coding apparatus according to Embodiment 2 of the present invention. Explanatory drawing which shows the reference pixel acquired by the reference pixel acquisition means in Embodiment 2 of this invention. Explanatory drawing which shows the periodic pixel position of the periodic pixel acquisition means in Embodiment 2 of this invention. The block diagram which shows the prediction state memory | storage means in Embodiment 2 of this invention. Explanatory drawing which shows the prediction state memory | storage part in Embodiment 2 of this invention. Explanatory drawing which shows the run length code table in Embodiment 2 of this invention. Explanatory drawing which shows the reference pixel change code | symbol in Embodiment 2 of this invention. Explanatory drawing which shows the decoding speed prediction table in Embodiment 2 of this invention.

Explanation of symbols

DESCRIPTION OF SYMBOLS 101 Control means 102 Input means 103 Reference pixel acquisition means 104 Dominant symbol acquisition means 105 Appearance probability acquisition means 106 MPS / LPS exchange means 107 Register update means 108 Predictive state transition means 109 Normalization means 110 Predictive transition table 111 Output means 112 Storage means 301 control means 302 input means 303 reference pixel acquisition means 304 dominant symbol acquisition means 305 prediction conversion means 306 run-length encoding means 307 prediction state transition means 308 prediction transition table 309 output means 310 storage means (prediction state storage means)
701 External storage device 702 Central processing unit 703 Read only memory 704 Random access memory 801 External storage device 802 Central processing unit 803 Read only memory 804 Random access memory 1101 Image encoding unit 1102 Reference pixel acquisition unit 1103 Periodic pixel acquisition unit DESCRIPTION OF SYMBOLS 1104 Dominant symbol acquisition means 1105 Predictive state storage means 1106 Predictive state transition means 1107 Predictive transition table 1108 Predictive conversion means 1109 Run length encoding means 1110 Reference pixel change code generation means 1111 Decoding speed prediction means 1112 Coding method switching means

Claims (13)

Reference pixel acquisition means for acquiring reference pixels in the vicinity of the encoded pixel;
Prediction state storage means for storing a prediction state and a dominant symbol for each context that is an appearance pattern of the acquired reference pixel;
A dominant symbol acquisition means for acquiring a dominant symbol stored in the prediction state storage means using the context;
Predictive conversion means for performing prediction of an encoded pixel using the dominant symbol and outputting the success / failure data series;
A prediction transition table for transitioning the prediction state based on whether the prediction is successful;
Predicted state transition means for performing a transition of a predicted state based on the predicted transition table;
An image encoding apparatus comprising: run-length encoding means for performing run-length encoding on the prediction success / failure data. Comprising periodic pixel acquisition means for acquiring pixels having a high correlation with encoded pixels at a constant period;
The reference pixel of the reference pixel acquisition unit includes a floating pixel that is a pixel that can be replaced with another pixel, and the value of the floating pixel and the pixel obtained by the periodic pixel acquisition unit are switched. The image encoding device described. Run-length storage means for storing a run-length that is the length of the run that has just appeared;
Encoding method switching means for switching a method for encoding the run length,
The image encoding apparatus according to claim 1 or 2, wherein the encoding method switching means switches the encoding method based on a run length stored in the run length storage means. A run-length appearance frequency counting means that counts the appearance frequency of run-lengths that appear within a certain section;
Encoding method switching means for switching a method for encoding the run length,
The image coding according to claim 1 or 2, wherein the coding method switching means switches the coding method to be applied to the next section based on the appearance frequency counted by the run length appearance frequency counting means. apparatus. A reference pixel change code for generating a reference pixel change code that indicates a change position of a floating reference pixel and whether or not the change has occurred during a run when the floating pixel is changed in the periodic pixel acquisition means 3. The image encoding apparatus according to claim 2, further comprising generating means. 6. The image coding apparatus according to claim 5, wherein when the run length reaches a preset run length, the reference pixel change code generating means generates a reference pixel change code. In the prediction state storage means, there are n × m sets (n and m are integers) of the prediction state and the dominant symbol for each context.
3. The reference pixel acquiring unit according to claim 2, wherein the reference pixels other than the floating pixels generate the same context for n encoded pixel units, and predictive conversion of n × m pixels is performed simultaneously. Image encoding device. Decoding speed prediction means for performing decoding speed prediction of a code based on the number of prediction errors within a certain interval for the prediction success / failure data output from the prediction conversion means;
Coding method switching means for determining whether or not the code output data of the raster is run-length encoded data or input data itself from the integrated value of the decoding speed output from the decoding speed prediction means in predetermined image line units The image encoding apparatus according to claim 1, further comprising: A reference pixel acquisition step of acquiring reference pixels in the vicinity of the encoded pixel;
A prediction state storing step of storing a prediction state and a dominant symbol for each context of the acquired reference pixels;
A dominant symbol acquisition step of acquiring the dominant symbol stored in the prediction state storage step using the context;
A predictive conversion step of performing prediction of an encoded pixel using the dominant symbol and outputting the success / failure data sequence;
A state transition step for performing a transition of the prediction state based on the prediction transition table for transitioning the prediction state based on the prediction success and failure, and the prediction transition table;
A run-length encoding step for performing run-length encoding on prediction success / failure data. A periodic pixel acquisition step of acquiring pixels having a high correlation with the encoded pixel at a constant period;
The image encoding method according to claim 9, wherein the reference pixel in the reference pixel acquisition step includes a floating pixel, and the values of the pixel and the floating pixel obtained in the periodic pixel acquisition step are switched. A run-length storage step for storing a run-length that is the length of the run that has just appeared;
An encoding method switching step for switching a method for encoding the run length,
The image encoding method according to claim 9 or 10, wherein the encoding method switching step switches the encoding method based on the run length stored in the run length storing step. A run length appearance frequency counting step that counts the appearance frequency of run lengths that appear within a certain section;
An encoding method switching step for switching a method for encoding the run length,
The image coding method according to claim 9 or 10, wherein the coding method switching step switches the coding method to be applied to the next section based on the appearance frequency counted by the run length appearance frequency counting step. . A storage medium storing a control program describing the image encoding method according to claim 9, 10, 11, or 12.

Images