JP3796235B2 - Image compression and decompression method - Google Patents

Description

  The present invention relates to a method for compressing and decompressing image data, and more specifically to a method for compressing and decompressing image data in a JBIG standard compression and decompression chip.

  The original image after sampling and digitization shows a very large data capacity. It is very troublesome to image, store, and transmit images having such a large data capacity. The image compression technique is for displaying the same image as small data to shorten the data transmission time.

  Basic data in image compression serves to remove duplicate data from an image. The duplication of data in images has the following causes.

  (1) As a cause of duplication of image data, first, there is duplication of coding, and an image file may be displayed with a code that is too short.

  (2) Another cause of image data duplication is that adjacent pixels have a very high correlation, and therefore, inter-pixel duplication may occur when data of a specific pixel is extracted from others.

  JBIG (Joint Bi-level Image Group) is a black and white (BW) image compression standard established by CCITT in 1993. The method used on this basis is to express each pixel in the bilevel image by one pixel. The process of full image compression can be achieved with a single CODEC (Compression and Depression) chip of JBIG standard. As JBIG technology is almost established, it has been widely used for printing and faxing.

FIG. 1 is a flowchart of an image data CODE method based on the JBIG standard.
When the CODE chip performs data processing, all steps in the figure need to be performed in each cycle T in order to compress and decompress each bit in the data.

  Initially, the system performs a memory access (step 10) and reads data from the memory. This step needs to be done for either data compression or data decompression. The accessed data is combined to form a context (Context, CX), which is used to collate a table and obtain a response state (State, ST).

  Step 20 causes the matching table (Look-Up-Table, LUT) to find the LSZ value that responds with the ST obtained in Step 10. Thereafter, A ′ = A−LSZ is subtracted in step 30, and the LSZ value obtained in step 20 is subtracted from the default value A to obtain the value of A ′.

  Step 40 selects A = A 'or A = LSZ and determines which continues the next operation. Thereafter, another subtraction of C′C−A + LSZ is performed (step 50). However, this time, a new C 'is searched by subtracting A from C using another default value C, and the LSZ value is added.

  Thereafter, step 60 performs C = C 'or C = C, which is a C value selection, to determine which continues the next operation. Step 70 is a renormalization (renormalization) process, and A is renormalized between 0.5 and 1. Step 80 renormalizes C and enlarges it.

  Finally, in step 90, byte out / byte in is performed, and the compressed and decompressed data is output to the memory. This completes the flowchart of the image data CODEC.

  For a 100 megahertz central processing unit (CPU), the work cycle is 10 ns. However, it takes about 10.5 ns to encode one bit of image data from memory access (step 10) to byte out / byte in (step 90). As for decoding, since the circuit becomes more complicated, about 11.5 ns is required for the entire processing.

  From the prior art described above, it has been found that the time required for the CPU to complete the compression process or the decompression process is longer than the work cycle. Therefore, this causes a delay in image data processing, resulting in a reduction in the printing speed of the printer.

  If the time required for compression and decompression of image data in each cycle cannot be shortened, the printing speed cannot be increased even if a fast CPU is used.

  Referring to the above description, an object of the present invention is to provide an image data compression and decompression (CODEC) method for a JBIG standard CODEC chip.

  This compression and decompression method is a method of compressing and decompressing image data into three parts: memory access, numerical operation, renormalization and byte-out / byte-in. Each step requires one work cycle.

  According to the image compression and decompression method of the present invention, 3-bit data compression and data decompression can be performed simultaneously in a work cycle. Therefore, the CODEC speed of the entire image can be increased at least twice. The method disclosed herein can be implemented on a high speed CPU to increase the data processing speed.

  The image compression and decompression method of the present invention is based on the data obtained by the 0.35 procedure simulation, and the CODEC method described below effectively increases the speed of the image data compression and decompression operation at least twice. Is possible. Therefore, this method can also be applied to the structure of a high-speed CPU.

  The present invention may be fully understood from the detailed description given below by way of example only, but the detailed description is not intended to limit the invention.

  The image compression and decompression (CODEC) method disclosed herein is implemented by a JDEC standard CODEC chip. FIG. 2 is a flowchart thereof. FIG. 2 illustrates the steps that need to be performed when compressing or decompressing one bit of image data.

  First, in the first cycle (n-1), a memory access is performed (step 100). If the data is compressed or decompressed, the image data needs to read the CODEC chip from memory and translates the reference pixel around the pixel to be restored from context (CX) to response status (ST).

  As shown in FIG. 3, when compressing or decompressing the data image within each pixel (white square), the surrounding pixels are taken as the reference pixel CX (gray square).

  In the second cycle n, as a next step, certain numerical operations are performed on the ST data obtained in the first cycle. The numerical calculation includes subtraction and selection of a collation table (LUT). The ST data and more portable symbol (MPS) obtained in step 100 are used to find a responding LSZ by collating the table (step 110).

  Subsequently, the system performs subtraction for two different parameters, the first default parameter A and the second default parameter C (steps 121 and 122). The first default parameter A is subtracted by the LSZ value to obtain a first subtraction value A ′ = A−LSZ. The second default parameter C is subtracted by the first default parameter A, and the LSZ value obtained by the LUT is added (step 110) to obtain C ′ = C−A + LSZ which is the second subtraction value.

  Thereafter, the system performs two different parameter parts, the new first default parameter A and the new second default parameter C (steps 131 and 132). The new first default parameter A is selected between the LSZ value and the first subtraction value A ′. The new second default parameter C is selected between the second default parameter C and the second subtraction value C ′. The new first default parameter A and the new second default parameter C selected in steps 131 and 132 become the first default parameter A and the second default parameter C necessary for the numerical operation at the next bit. Used.

  Finally, in the next cycle, the third cycle (n + 1), the new first default parameter A, and the new second default parameter C are renormalized (steps 141 and 142). The image data is output after being subjected to numerical operation and renormalization (step 143).

  In steps 141 and 142, the first default parameter A and the second default parameter C are renormalized. The byte-out / byte-in in step 143 is output for saving the compressed or decompressed image data to another memory through the JBIG project CODEC chip.

  Referring to FIG. 4, a process in which the present invention is used to restore the 10th bit of image data will be described. (N-1) In the work cycle, the ninth bit of data is not completely restored, so that the tenth bit cannot be restored.

  In this case, the ninth bit can be temporarily set to 0 or 1 (CX = 0 or 1). Then, the system can restore the 10th bit. In work cycle n, the ninth bit is restored. Therefore, the CX value of the ninth bit is known and used to execute correct restoration of the image data.

  With reference to FIG. 5, the process of processing three pixels simultaneously in each work cycle will be described. The prior art work cycle without this pipeline improvement is long compared to the invention. This method can efficiently shorten the work cycle of each image data processing, and increases the speed at which image data is compressed or restored.

Various modifications will be apparent to those skilled in the art and are deemed to be within the spirit and scope of each claim recited in the claims below.

The flowchart figure of the image data compression and decompression method of the JBIG standard of a prior art. The flowchart figure of the compression and decompression | restoration method which this invention disclosed. Schematic of the process of compressing or decompressing image data. Schematic of the process of restoring the 10th bit in the image. Schematic of the process of parallel processing 3 pixels simultaneously in a work cycle.

Explanation of symbols

10 Memory access 20 Verification table (LUT)
30 subtraction 40 selection 50 subtraction 60 selection 70 A renormalization 80 C renormalization 90 byte out / byte in 100 memory access 110 collation table (LUT)
121 Subtraction 122 Subtraction 131 Selection 132 Selection 141 A Renormalization 142 C Renormalization 143 Byte Out / Byte In

Claims (3)

An image compression and decompression method using the Joint Bilevel Image Group (JBIG) image compression standard,
It consists of a first cycle, a second cycle, and a third cycle,
The first cycle reads image data, converts a plurality of reference pixels around a pixel to be restored from a context (CX) format to a state (ST) format,
The second cycle performs a numerical operation of image data, and at this time, a first step of obtaining an LSZ value using ST in a collation table (LUT) for the numerical operation;
A second step of subtracting the LSZ value from the first default parameter to obtain a first subtraction value and selecting a new first default parameter between the LSZ value and the first subtraction value ;
A third step of subtracting the first default value from the second default parameter and adding the LSZ value to obtain a second subtraction value and selecting a new second default parameter from the second default parameter and the second subtraction value And including
The third cycle renormalizes the new first default parameter and the new second default parameter, and outputs processed image data.
In the second cycle, the second step and the third step are performed in parallel,
Image compression and decompression method.
2. The image compression and decompression method according to claim 1, wherein the step of converting a plurality of reference pixels around a pixel to be restored from the context (CX) format to the state (ST) format in the first cycle is compressed from memory. And an image compression and decompression method comprising: reading image data into a decompression (CODEC) chip and processing the image data.
The image compression and decompression method according to claim 1, wherein converting a plurality of reference pixels around a pixel to be restored from a context (CX) format to a state (ST) format during a first cycle of reading image data. In order to restore image data, the image compression and restoration method is characterized in that a pixel before the pixel restored to 0 or 1 is set to obtain CX = 0 and 1.

Images