JPH037316B2 - - Google Patents

Description

[Detailed Description of the Invention] (a) Technical Field of the Invention The present invention relates to a data compression method, and particularly in ultra-high-speed facsimile, etc., it is possible to efficiently compress even complex images in which the number of change points on one scanning line exceeds a predetermined value. This invention relates to improvements in image data compression methods that can be processed.

(b) Prior Art and Problems As a conventional image data compression method, a Modified Read (hereinafter abbreviated as MR) method, which is an international standard, has been used. The MR method targets data for two adjacent scanning lines of image data obtained by raster scanning a document, and detects changing points on the encoding scanning line (scanning line to be encoded), that is, from white to black. , the position of the boundary point that changes from black to white is determined by the relative distance (hereinafter referred to as 'displacement') from a change point indicating the same change point on the reference scan line that precedes the encoded scan line, or on the encoded scan line. This is an encoding method that expresses the relative distance (hereinafter referred to as run length and abbreviated as RL) to the immediately preceding change point that shows the opposite change. Approximately 90% of all changing points in image data can be expressed as shifts, and high compression rates can be obtained by assigning short codes to these shifts.

An outline of the MR method will be explained using FIG. 1.

The mode of the changing point consists of three types: a pass mode shown in FIG. 5A, and a horizontal mode and vertical mode shown in FIG. In pass mode, after the change point a _{0 on the encoded scan line SA corresponding to the change point b 0} on the reference scan line SB is detected, the change point a ₀ on the reference scan line SB is detected before the next change point a ₁ is detected. This is a mode in which when change points b ₁ and b ₂ are detected in , these two change points b ₁ and b ₂ are not used as references for encoding.

The vertical mode is a mode in which the position of a changing point on the encoded scanning line SA is expressed as a relative position from the changing point on the reference scanning line SB, and is used when the relative position is within 3 pixels. In this mode, in FIG. 2B, for example, the position of a change point _a0 on the encoded scanning line SA is expressed by the number of pixels from the corresponding change point _b0 on the reference scan line SB.

The horizontal mode is used when the relative distance between the changing point on the reference scanning line SB and the corresponding changing point on the encoded scanning line SA is 3 pixels or more, such as a ₁ for b ₁ in b in the same figure. mode. In this mode, the change points a ₁ and a ₂ are represented by distances a ₀ , a ₁ , a ₁ , and a ₂ . As mentioned above, this distance is called RL (run length), and a one-dimensional modified Huffman (hereinafter abbreviated as MH) code is assigned to RL.

FIG. 2 is a main part block diagram showing a conventional MR type data compression circuit. In the figure, the input buffer section is a data compression processing section.

First, the read image data is transferred to one of the first, second, and third line memories 2, 3, and 4, for example, the first line memory 2 by the demultiplexer 1.
is input. Next, the image data of the next scanning line is sequentially input to the second and third line memories 3 and 4, and then the next image data is input again to the first line memory 2. In this way, image data is cyclically input to the three line memories. In this process, the two line memories not selected for inputting read data are used via multiplexer 5 for reference scanning lines and encoding scanning lines. That is, under the control of the input buffer control section, three line memories are cyclically rotated and one is used for reading and writing data, and during this time, the line memory into which image data was input just before is used as an encoding scanning line. , and the contents of the line memory that was input one line earlier are read out as the data of the reference scanning line.

In the data compression processing section, the MR mode detection section 7 detects the mode of the change point and counts the shift and RL while simultaneously shifting the image data on the reference scanning line and the encoded scanning line. Next, the MR code corresponding to the mode of the change point is read out by the MR code reading unit 8, and the MR code creating unit 9 performs parallel-to-serial conversion to create a series of variable length codes.

The data compression processing control section 10 controls pipeline processing in order to efficiently execute processing in three parts: the MR mode detection section 7, the MR code reading section 8, and the MR code creation section 9.

The processing cycle per line for the above three parts is as follows.

(Number of processing cycles of mode detection section) = (Number of image data shifts) + (Number of change points) + α ... (Number of processing cycles of MR code reading section) = (Number of change points) + β ... (Number of processing cycles of MR code creation section ) = (number of data of MR code) + (number of change points) + γ... However, in the above formula, α, β, and γ are the number of cycles each part has during pipeline processing.

As shown in FIG. 3, the number of data in the above MR code corresponds to two to three times the number of change points. That is, one change point is expressed by 2 to 3 bits.

~As is clear from the formula, when performing pipeline processing, if the number of change points per line is small, (number of shifts of image data >> number of change points)
Therefore, the processing time of the mode detection section 7 determines the time of the entire data compression processing section.

However, as the number of change points increases, the number of data in the MR code increases at a rate of two to three times the number of change points, so the processing time of the MR code creation section 7 will influence the processing time of the entire data compression processing section.

For example, if the number of image data shifts per line is approximately 3000 (corresponding to the number of pixels per line of a B4 original with a resolution of 12 lines/mm), the number of change points is approximately 1000, and the code data per change point is 3 bits. The number of processing cycles of the mode detection unit 7 and the MR code creation unit is both 4000 cycles + α or γ,
If the number of change points increases further, the processing time of the MR code creation unit 8 will become longer. In reality, due to the influence of time, this phenomenon occurs where the number of points changing is smaller.

When a complex image appears with more than 1000 points of change per line, for example, when the image is a character string, this state usually lasts for several dozen lines or more. Since the commonly used data compression circuit (FIG. 2) has a buffering capacity for only one line, the processing time of the data compression processing section is long and the processing is often not completed in time. Also, when the number of change points exceeds 1000, the number of image data shifts and the number of MR code data are almost the same.
That is, the compression ratio is close to 1, and there is a drawback that the compression effect is not improved.

(c) Purpose of the Invention The purpose of the present invention is to provide a data compression method that can shorten the processing time of a data compression processing unit and increase the data compression effect even when a complex image appears. It is in.

(d) Structure of the Invention The feature of the present invention is to detect the relative position of the change point of the image data on the next scanning line with respect to the change point of the image data on one scanning line, and to detect the relative position of the change point of the image data on the next scanning line based on the detected positional relationship. When encoding image data on a scanning line, if the number of change points or the number of encoded data on one scanning line exceeds a predetermined value, the image data on the next scanning line is converted to the image data on the one scanning line. The purpose is to replace and encode.

(e) Embodiments of the Invention The present invention takes advantage of the fact that the MR code for a shift of 0 is one bit, and replaces the next line data with the previous line data, so that all change point modes have a shift of 0. The compression processing time is shortened and the compression effect is improved.

An embodiment of the present invention will be described below with reference to FIGS. 4 to 6.

FIG. 4 is a main part block diagram showing the configuration of a data compression circuit used in an embodiment of the data compression method according to the present invention.

The data compression circuit shown in the figure has a change point counter 11 and a code data counter 12 newly added compared to the conventional circuit, and the multiplexer 5' can select the same line memory as a reference scanning line and an encoding scanning line. The difference is that

When a value greater than a certain value is detected in the change point counter 11 or code data counter 12 during data compression processing of the current line, for example, the change point
If the number of points exceeds 1000, or if the amount of encoded data becomes larger than the amount of image data, this will be notified to the input buffer control unit so that the data in the next line's image will be the same as the data in the current line. , controls the first to third line memories 2, 3, and 4. As a result, for example, when the above state is detected in lines a, c, and e in the image data of Fig. 5a, the image data compressed and reproduced by this method will be as shown in Fig. 5b. The above lines b, d,
f is the same as the lines a, c, and e immediately above each.

Such an operation is performed as follows. That is, as shown in the first line (row of encoding processing line a) in FIG. 5d, the image data of line a stored in the third line memory 4 is Encoding processing is performed using the image data of the line immediately before line a (not shown) stored in line a as a reference scanning line, and the image data of the next line b
image data is read and stored in the first line memory 2. During this operation, the change point counter 11 or code data counter 12 detects that the change point in line a or the code data amount exceeds a predetermined value, and the data compression processing control unit 10 notifies the input buffer 6 of this fact. Ru. Then, the input buffer control unit 6 uses the third line memory 4 holding the image data of line a as the encoded scanning line instead of the first line memory 2 described above as the image data of line b. Control. As a result, the encoded scanning line and the reference scanning line are the same, and therefore the transmitted data on line b becomes an MR code in which all change points are represented by 1 bit, reducing the data compression processing time and compressing the data. The processing effect also increases. Similar operations are subsequently performed when encoding lines d and f.

According to the present embodiment, the expressions ~ in the conventional system are as follows. In other words, since the next line is the same as the previous line, (number of processing cycles of mode detection section) = (number of shifts in image data) + (number of change points) ... (number of processing cycles of MR code reading section) = (change Number of points) + β' Processing cycle number of MR code creation unit processing) = 2 × (number of change points) + γW...Then, in the MR code creation unit, the number of code data is the same as the number of change points, and the code '1' with a deviation of 0 continues. Therefore, since the processing time is short, the mode detection section does not have to wait for processing (α=0), and the processing time as a whole is shortened.

FIG. 6 is a main part block diagram showing a data compression circuit as another embodiment of the present invention. 14 and a multiplexer 15 are attached.

In other embodiments, when the number of change points or the number of code data in the current line data is larger than a predetermined value,
The next line is not compressed, but instead the generator 14 outputs patterns of '1' corresponding to the number of detected change points on the current line via the multiplexer 15.

This method eliminates the need to process data for one line, so the input buffer has buffers for two lines, leaving the data for the current line. Therefore, according to this embodiment as well, the compression processing time is significantly shortened, and the fear that the processing will not be completed in time is eliminated.

(F) Effects of the Invention As explained above, according to the present invention, when the number of change points on one scanning line or the number of encoded data exceeds a predetermined value when encoding image data,
By replacing the image data on the next scanning line with the image data on the one scanning line for encoding, the data compression processing time is shortened and the compression rate is also significantly improved. Further, since the encoding processing method is compatible with the standard without changing the image data at all, it is possible to easily add the processing device according to the present invention to an existing compression device.

[Brief explanation of the drawing]

Figure 1 is a diagram schematically showing the change point mode of the MR encoding system, Figures 2 and 3 are diagrams showing the difficulties of the conventional data compression system using the MR system, and Figure 2 is a diagram showing the data compression circuit. Main part block diagram, Figure 3 is a diagram showing the relationship between the number of change points and the amount of code data, Figure 4 is a diagram showing the relationship between the number of change points and the amount of code data.
5 and 5 are diagrams showing an embodiment of the present invention, FIG. 4 is a block diagram of the main part of the data compression circuit used in the above embodiment, and FIG. 5 is an example of a reproduced image according to the embodiment. The pattern diagram shown in FIG. 6 is a main part block diagram showing the configuration of a data compression circuit used in another embodiment of the present invention. In the figure, 1 is a demultiplexer, 2 to 4 are line memories, 5 and 5' are multiplexers, 6 is an input buffer, 7 is an MR mode detection section, 8 is an MR code detection section, 9 is an MR code generation section,
11 is a change point counter, 12 is a code data counter, numeral 11 is an input buffer section, and numeral is a compression processing section.

Claims (1)

[Claims] 1. When detecting the relative position of the changing point of image data on the next scanning line with respect to the changing point of image data on one scanning line, and encoding the image data on the next scanning line from the detected positional relationship. , when the number of change points or the number of encoded data on the one scanning line exceeds a predetermined value, the image data on the next scanning line is replaced with the image data on the one scanning line for encoding. Data compression method.