JP2002152520A - Image encoder and encoding method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To obtain an image encoder in which generation of overrun can be prevented in decoding. SOLUTION: The image encoder 1 comprises means 4 performing arithmetic encoding of input image data for every raster, means 5 for affixing an end code to an arithmetic code, an overrun decision means 6 where a decision is made whether a code is decoded within a specified time for every raster, a raw data output means 3 outputting a code using input raster data, and means 2 for outputting a code generated from the arithmetic encoding means 4 or a code generated from the raw data output means 3 selectively using an output from the overrun decision means 6.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image encoding apparatus and method for encoding binary image data using arithmetic encoding.

[0002]

2. Description of the Related Art In a conventional printing apparatus provided with an image coding apparatus using arithmetic codes, all print images on one page are all encoded by arithmetic codes, and printing is performed while decoding codes at the time of printing. I was going.

[0003]

In such a conventional image coding apparatus, since all the printed images of one page are all encoded by arithmetic codes, when the printing apparatus is operated at a high speed, the compression ratio is reduced. In the case of a bad raster, there is a possibility that the decoding of the arithmetic code cannot catch up with the printing and the printing is not performed normally, resulting in an overrun.

Accordingly, an object of the present invention is to provide an image encoding apparatus and method capable of preventing occurrence of overrun during decoding.

[0005]

To solve this problem, an image coding apparatus according to the present invention comprises an arithmetic coding means for performing arithmetic coding on input image data for each raster, and an arithmetic coding means. An end code adding unit that adds a code, an overrun determination unit that determines whether decoding of a code is performed within a predetermined time for each raster, a raw data output unit that outputs a code using input raster data, And a code selection output means for selecting and outputting the code generated by the arithmetic coding means or the code generated by the raw data output means using the output of the overrun determination means.

Thus, the output code of the raw data output unit and the output code of the arithmetic coding unit are selectively output using the determination result of the overrun determination unit for each raster.
It is possible to prevent the occurrence of overrun at the time of decoding.

In order to solve this problem, an image coding method according to the present invention performs arithmetic coding on input image data for each raster, adds an end code to the arithmetic code, and performs code coding for each raster. It is determined whether or not decoding is performed within a predetermined time, a code output of raw data is performed using the input raster data, and an arithmetic code or The raw data code is selectively output.

[0008] Thus, since the raw data code or the arithmetic code is selectively output using the overrun determination result for each raster, it is possible to prevent the occurrence of overrun at the time of decoding.

[0009]

DESCRIPTION OF THE PREFERRED EMBODIMENTS The invention according to claim 1 of the present invention provides an arithmetic coding means for performing arithmetic coding on input image data for each raster, and an end code addition for adding an end code to the arithmetic code. Means, overrun determination means for determining whether code decoding is performed within a predetermined time for each raster, raw data output means for performing code output using input raster data, and output of the overrun determination means. Code selecting and outputting means for selecting and outputting a code generated by the arithmetic coding means or a code generated by the raw data output means using the determination result of the overrun determination means for each raster. Since the output code of the raw data output means and the output code of the arithmetic coding means are selectively output, it is possible to prevent the occurrence of overrun at the time of decoding.

According to a second aspect of the present invention, arithmetic coding is performed on input image data for each raster, an end code is added to the arithmetic code, and decoding of the code is performed within a predetermined time for each raster. A determination is made as to whether the decoding is to be performed, a raw data code is output using the input raster data, and an arithmetic code or a raw data code is selectively output from the determination result as to whether the decoding of the code is performed within a predetermined time for each raster. An image encoding method,
Since the raw data code or the arithmetic code is selectively output by using the overrun determination result for each raster, it is possible to prevent the occurrence of overrun at the time of decoding.

Hereinafter, an embodiment of the present invention will be described with reference to FIG.
This will be described with reference to FIG. In these drawings, the same members are denoted by the same reference numerals, and duplicate description is omitted.

FIG. 1 is a block diagram showing an image coding apparatus according to an embodiment of the present invention, FIG. 2 is a block diagram showing arithmetic coding means in the image coding apparatus shown in FIG. 1, and FIG. FIG. 4 is an explanatory diagram showing a reference pixel and a pixel to be encoded in the encoding device, FIG. 4 is an explanatory diagram showing a prediction table in the image encoding device in FIG. 1, and FIG. 5 is an arithmetic encoding unit in the image encoding device in FIG. FIG. 6 is a flowchart showing the arithmetic encoding process in FIG. 5, FIG. 7 is a flowchart showing the normalization process in FIG. 6, and FIG. 8 is a flowchart showing the code output process in FIG.

The image coding apparatus 1 of this embodiment includes a raw data output means 3 for performing code output using input raster data, and an arithmetic coding means 4 for performing arithmetic coding on input image data for each raster. Using an output of an end code adding means 5 for adding an end code to an arithmetic code, an overrun judging means 6 for judging whether decoding of a code is performed within a predetermined time for each raster, and an output of the overrun judging means 6. It comprises a code selection output means 2 for selectively outputting a code generated by the arithmetic coding means 4 or a code generated by the raw data output means 3.

The arithmetic encoding means 4 comprises an arithmetic encoder 4a, a prediction table 4b and an additional data generator 4c, as shown in FIG.

Here, the operation of the arithmetic coding means 4 will be described with reference to FIG.

First, various registers are initialized (S100). That is, the initialization is performed according to the following equations 1 to 11.

NUM = 0 (Equation 1) A = 0x100 (Equation 2) CT = 8 (Equation 3) CS = 0 (Equation 4) C = 0 (Equation 5) BUFFER = 0x00 (Equation 6) TEMP = 0 (Equation 6) 7) Amps = 0 (Equation 8) Alps = 0 (Equation 9) eflag = 0 (Equation 10) ecount = 0 (Equation 11) Here, NUM in Expression 1 is a register for counting the number of input pixels, and A in Expression 2 is A Is a register indicating the width of the effective area, and the initial value is 0x100 representing the width of the number line from 0 to 1, indicating that the decimal part is a number line having a resolution of 8 bits. Note that 0x indicates that the following numbers are in hexadecimal notation. CT in Equation 3 is a counter for code output processing, CS in Equation 4 is a carry propagation hold counter, C in Equation 5 is a 17-bit code register, BUFFER in Equation 6 is an 8-bit code output buffer, and 7, TEMP is a 9-bit temporary register, Amps in equation 8 is the number linear width of MPS, Alps in equation 9 is the number linear width of LPS, equation 10
Eflag indicates an end flag of the input image data, and ecount in Expression 11 indicates a code data output counter after detection of the end flag.

If initialization is performed in this way,
Next, arithmetic coding processing is performed (S101).

FIG. 6 is a flowchart showing the operation of the arithmetic coding process.

In the calculation encoding process, first, when the reference pixel and eflag do not stand, one pixel input to the encoding device is obtained and set as a pixel value PIX, and when the eflag stands, FIG. The data generated by the additional data generator 4c is referred to as a pixel value PIX. Then, the value of the reference pixel is given as an address from the prediction table 4b,
The value of MPS and the value of SRL are acquired as data (S1
04). The reference pixel is 8 pixels from the encoding target pixel as shown in FIG.
With reference to three pixels every other pixel, each pixel value is set as an address A2, A1, A0 in the prediction table 4b. FIG. 4 shows the structure of the prediction table 4b. Using the value of the reference pixel as an address, the address is from 0x0 to 0x7, the data width is 4 bits, 3 of which are SRL values, and 1 of which is MPS.
It is a table that is a value.

Next, the values of Alps and Amps are calculated from the obtained SRL (S105).

Then, the MPS value is compared with the PIX value (S106). If they match, the effective area width A is updated to the dominant symbol width Amps (S107). If they do not match, the effective area width A is updated. The value of the C register is updated to C = C + Amps while updating to the inferior symbol width Alps (S109).

Next, it is determined whether A is less than 1/2 by A <
It is determined whether 0x80 is true or false (S108). If it is turned off, normalization is performed to return A to １ ／ or more (S108).
110). Otherwise, the arithmetic coding process ends.

FIG. 7 is a flowchart showing the operation of the normalization processing in FIG.

In the normalization processing, A and C are shifted left by one bit, and the value of CT is subtracted by 1 (S111). next,
A code output process is performed (S112), and the processes of S111 and S112 are repeated until A becomes 0x80 or more (S112).
113).

FIG. 8 is a flowchart showing the operation of the code output process of FIG.

In the code output processing, first, it is evaluated whether or not the value of CT is 0 (S114). If not, the code output processing is terminated without performing the code output processing.
Is 0, a code output process is performed.

The code output is obtained by shifting C by 8 bits to the right and storing it in TEMP (S115).
An evaluation is made as to whether it is greater than F (S116).

If TEMP is greater than 0xFF, the value of (BUFFER + 1) is output as one byte,
Next, since the value of TEMP is greater than or equal to 0xFF, a carry occurs.
Update the value of FFER to the value of the lower 8 bits of TEMP,
If elfag is set, the value of ecount is incremented by 1 (S117).

On the other hand, if the value of TEMP is equal to 0xFF (S119), the number of holdings CS
Is incremented by one, and the code output is kept pending (S120).

If the value of TEMP is smaller than 0xFF (S119), the value of BUFFER is output by 1 byte, and then TEMP is 0xFF and no carry occurs. BUFFER
Is updated to the value of TEMP, and the value of ecount is incremented by 1 if elfag is set (S12).
1).

After outputting the respective codes according to the value of TEMP in this way, the value of C is updated to C & 0x7FFFF, the value of CT is updated to 8 (S118), and the code output process ends.

After the arithmetic coding process has been performed in this way, it is evaluated in FIG. 5 whether or not the count matches the set value of enum (S102). The arithmetic coding process by the converting means 4 is terminated, and if they do not match, the number of input pixels is incremented by one, and if the input pixel number NUM is equal to the set processing pixel number, an eflag is set (S103) and the process returns to S101. .

In FIG. 1, the end code adding means 5
After the arithmetic encoding unit 4 encodes one raster, the 3-byte codes 0xff, 0xff, and 0x00 are added to determine the end of one raster when decoding.

The overrun determining means 6 sets the data size per raster of the printing apparatus, the required time per raster, and the code size of the arithmetic code allowed per raster based on the operating frequency of the arithmetic coding means. Make a decision. The time Tprt (unit: ns) required for one pixel input of the printing apparatus is expressed by the time between rasters as T.
bd, raster data size is set to Tras (unit bit)
Then, it is obtained by the following Expression 12.

Tprt = Tbd / Tras (Equation 12) Next, the average processing cycle time Tar of one pixel of the decoder
(Unit: ns) represents the number of processing cycles of the pixel window as C
win, the number of normalization processing cycles is Cregl, the code output unit Csize (8 in the present embodiment, which is 8 bits), and the operation clock of the encoding device is Tclk (unit n).
s), the number of cycles required for the encoded output is Cout, and the required compression ratio is Rreq. Rreq that satisfies Tprt = Tar is obtained. Then, the threshold value Tth of the code used for the overrun determination is obtained by Tth = Tras / Rreq (Equation 14).

Using this Tth as a threshold, if the code size generated by the arithmetic coding means 4 for each raster is larger than Tth, it is determined that overrun has occurred, and if it is less than Tth, it is determined that no overrun has occurred.

The raw data output means 3 sends the raw raster data to the code selection output means 2 as a code.

The code selection output means 2 receives the overrun determination of the overrun determination means 6 and outputs the code output from the raw data output means 3 after the code indicating the raw data output in the raster with the occurrence of overrun. In the raster without overrun occurrence, the code output from the arithmetic coding means 4 is selected and output after the code indicating the arithmetic code output.

As described above, according to the present embodiment, the output code of the raw data output unit 3 and the output code of the arithmetic coding unit 4 are selectively output by using the determination result of the overrun determination unit 6 for each raster. Therefore, it is possible to prevent occurrence of overrun at the time of decoding.

[0041]

As described above, according to the present invention, the raw data code or the arithmetic code is selectively output by using the overrun determination result for each raster. This has an effective effect that prevention can be achieved.

[Brief description of the drawings]

FIG. 1 is a block diagram showing an image encoding apparatus according to an embodiment of the present invention.

FIG. 2 is a block diagram showing arithmetic coding means in the image coding apparatus of FIG. 1;

FIG. 3 is an explanatory diagram illustrating reference pixels and encoding target pixels in the image encoding device of FIG. 1;

FIG. 4 is an explanatory diagram showing a prediction table in the image encoding device of FIG. 1;

FIG. 5 is a flowchart showing an operation of an arithmetic coding unit in the image coding device of FIG. 1;

FIG. 6 is a flowchart showing an arithmetic coding process in FIG. 5;

FIG. 7 is a flowchart showing a normalization process in FIG. 6;

FIG. 8 is a flowchart showing a code output process in FIG. 7;

[Explanation of Codes] 1 Image coding device 2 Code selection output means 3 Raw data output means 4 Arithmetic coding means 5 End code addition means 6 Overrun determination means

Claims (2)

[Claims] 1. An arithmetic coding unit for performing arithmetic coding on input image data for each raster, an end code adding unit for adding an end code to an arithmetic code, and decoding of a code for each raster within a predetermined time. Overrun determination means for determining whether or not to be performed; raw data output means for performing code output using input raster data; code generated by the arithmetic coding means using the output of the overrun determination means or A code selection output means for selecting and outputting a code generated by the raw data output means. 2. An arithmetic coding is performed on input image data for each raster, an end code is added to the arithmetic code, and it is determined for each raster whether the code is decoded within a predetermined time. An image encoding method comprising: outputting a raw data code using data; and selectively outputting an arithmetic code or a raw data code based on a determination result as to whether the code is decoded within a predetermined time for each raster.