JPS61123276A - Picture data compression system - Google Patents

Abstract

PURPOSE:To compress efficiently a mixed picture between a character and a contrast picture by discriminating a character/line picture area or a contrast picture depending on the probability of generation of white/black pattern data comprising plural picture elements, separating the areas and applying MH coding to the contrast area after forecast conversion. CONSTITUTION:After a mixed picture between a character/line drawing and a contrast picture fed to an input terminal 1 is set to a register 2, the picture is fed to a shift register 3, an exclusive OR circuit 4 and a multiplexer 7. A means comprising a ROM 10, a register 11 and a line counter 10 discriminates whether an area is a character/line picture area or a contrast area. When the area is discriminated as the character/line picture area, the input picture data is fed to an MH coder 8 via the multiplexer 7 to apply MH coding, and when the area is discriminated as the contrast area, the input picture data is subjected to forecast conversion, the result is fed to the MH coder 8 via the multiplexer 7 as a smoothed forecast error signal (e) to attain MH coding.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to an image data compression method for compressing mixed image data of characters, line drawing data, and gray scale image data using a standard method.

[Conventional technology]

Characters, line drawings, etc. can be expressed in black and white binary, and M
Data can be compressed and transmitted using a standard method such as H encoding. As for grayscale images, they can be expressed as pseudo-halftones using black and white binary values using the dithering method, but since the grayscale levels are expressed by the density of black pixels, the black and white runs on the scanning line are separated. Therefore, data compression using standard methods such as MH encoding cannot be expected. To address these problems, various methods have been proposed that are capable of compressing dithered image data and are compatible with the standard method. For example, a predictive MH method (see "MH Coding of Dithered Images" Technical Report IE82-45 (1982), Institute of Electronics and Communications Engineers) has been proposed as a method using predictive conversion.

FIG. 5 shows a block diagram of the above-mentioned predictive MH method, with 41
is an input terminal, 42 is a predictor, 43.44.

is an adder, 45 is a delay circuit, 46 is an MH encoder, and 47 is an output terminal, which forms a predicted signal intersection of dithered image data Adder 43 adds data X and predicted signal father
In addition, the prediction signal intersection is subtracted from the dithered image data X, and a prediction error signal e is output. Adder 43 is modulo 2
This is an adder and performs the same function as an exclusive OR circuit.

In the adder 44, using the prediction error signal e from the adder 43 and the signal delayed by one pixel in the delay circuit 45,
A smoothed prediction error signal d is formed. This smoothed prediction error signal d becomes a signal whose level is inverted when the prediction error signal e is "1", is applied to the MH encoder 46, is converted into an MH code, and is sent out from the output terminal 47.

In the case of using a 4×4 dither matrix, the predictor 42 selects the reference pixel x1 four pixels before the pixel of interest xo on the scanning line in the case of − dimension, as shown in [a] of FIG.
For two-dimensional cases, the prediction is made using
As shown in b), the pixel of interest X is set using surrounding pixels x1 to x6 as reference pixels. It is used to make predictions.

FIG. 7 shows an example of the relationship among the multivalued image signal, dithered image signal X, prediction signal intersection, prediction error signal e, and smoothed prediction error signal d when one-dimensional prediction is used.

In the figure, a indicates a multi-valued image signal, b indicates a dithering threshold, and the dithered image signal X indicates a multi-valued image signal whose level is higher than the dithering threshold as a black pixel with diagonal lines,
The predicted signal intersection corresponds to a signal obtained by delaying the dithered image signal X by 4 pixels, and the prediction error signal e corresponds to the output signal of the exclusive OR of the dithered image signal The error signal 1 corresponds to a signal that is inverted for each prediction error signal e. Therefore, since the run length of this smoothed prediction error signal can be increased, efficient encoding can be performed in the MH encoder 46.

[Problem that the invention seeks to solve]

When using the one-dimensional prediction in the prediction MH method described above,
Since the reference pixel Xl is 4N pixels before the pixel of interest XQ, the correlation between the pixel of interest XQ and the reference pixel Xl is small in areas such as characters. That is, in areas such as characters, the correlation between adjacent pixels is large, but as the distance between pixels increases, the correlation decreases. Therefore, there is a drawback that the data compression ratio decreases. In addition, when two-dimensional prediction is used, prediction is performed using reference pixels suitable for both text and grayscale images, and transmission errors spread, so redundant bits are added to compressed data as an error countermeasure. There is a need to.

SUMMARY OF THE INVENTION An object of the present invention is to provide a one-dimensional encoding method that is compatible with standard methods and can obtain a high compression ratio when compressing data of a mixed image of characters and grayscale images.

Means for Solving Problem C] The present invention provides a method for determining whether a character, line drawing area or gray area is a text area, a line drawing area, or a gray area based on black and white pattern data consisting of a plurality of pixels other than all black and all white based on dithered image data of a light and shade image. a switching means for selecting the input image data when the determination means determines that it is a character or line drawing area, or selects the smoothed prediction error signal when determining that it is a grayscale image area;
For image data containing a mixture of text, line drawing areas, and gray area, the text and line drawing areas are MH encoded, and the gray area is predictively converted and then MH encoded. It is something to do.

[Effect]

After determining whether it is a text, m-picture area or a light and shade picture area based on the probability of occurrence of black and white pattern data consisting of multiple pixels, separating the text and line picture areas from the light and shade picture area, and predictively converting only the light and shade picture area. MH encoding is performed, and text and line drawing areas are directly MH encoded.

〔Example〕

FIG. 1 is a block diagram of an encoding circuit according to an embodiment of the present invention, where 1 is an input terminal, 2 is a register (REG), 3 is a shift register (SHR), 4.5 is an exclusive OR circuit,
6 is a register (REG), 7 is a multiplexer (MP
X), 8 is the MH encoder (CD), 9 is the output terminal, 10
is a read-only memory (ROM), 11 is a register (RBG), and 12 is a line counter (LCTR). The input terminal 1 receives data of a mixed image of characters, line drawings, and grayscale drawings, and after being set in the register 2, the data is transferred to the shift register 3. Exclusive logic circuit 4 and multiplexer 7 are added.

When a 4×4 dither matrix is used, the shift register 3 has a 4-bit configuration, and the exclusive OR circuit 4 receives the pixel of interest X from the register 2.

Reference pixel Xl from shift register 3 to 4 pixels before
will be added. The output signal of this exclusive OR circuit 4 becomes the prediction error signal e. Also, the register 6 is for giving a delay of one pixel, and the exclusive OR circuit 5
, the prediction error signal e and the smoothed prediction error signal d delayed by one pixel are added, and the smoothed prediction error signal d changes every time the prediction error signal e becomes "1".

Fig. 2 shows the gradation pattern when using a Bayer type 4 x 4 dither matrix, and the gradation pattern ranges from 0 to 16.
The gradation of is expressed by the number of black pixels. Furthermore, when the first line L1 to the fourth line L4 are summarized as a 1×4 pixel pattern, as shown in FIG.
~4 gradations. For example, when looking at the gradation 2 of the first line Ll, it corresponds to the gradation 3.4, 5.6, and 7°8 in FIG. 2, and when looking at the gradation 2 of the fourth line L4, Gradation levels 8.9, 10, 11.12.13 in Figure 2
It corresponds to In the grayscale image data, such 1×4 pixel pattern data is generated, and the all-black and all-white patterns of each line L1 to L4 can also occur in the text and line drawing areas. Because the nature is large,
It can be determined whether the area is a character, line drawing area, or gray area using the patterns of the levels lil to 3 of each line L1 to L4.

Read-only memory 10. The register 11 and the line counter 12 constitute a determining means for determining whether it is a character, line drawing area or gray area, and the shift register 3 outputs it as 4-bit parallel data, and the third
The data becomes the 1×4 pixel pattern data shown in the figure, and becomes part of the address signal of the read-only memory IJ10. The lower two bits of the line counter 12 are used in the read-only memory 1 for identifying which line among the first line L1 to the fourth line L4 of the 4×4 dither matrix.
It becomes part of the 0 address signal. Further, as internal history information, for example, 3 pinto is read out and set in the register 11, and the internal history information also becomes part of the address signal of the read-only memory 10.

The internal history information constitutes a kind of counter, and the 4-bit parallel data from the shift register 3 is input to the line counter 12.
The internal history information indicates the consecutive number of times corresponding to the 1 x 4 pixel gradation level 1 to 3 pattern in Figure 3 on the line indicated by the lower 2 bits from Since n is added to the read-only memory 10 as part of the address signal, the control information for the multiplexer 7 to select the smoothed prediction error signal d is read out. register 11
is set to

After it is determined that the area is a gray area, the 4-bit parallel data from the shift register 3 is applied to the gradations 1 to 3 of the 1×4 pixels in FIG. If the area does not correspond to the pattern m times in a row, it is determined that the area is a character or line drawing area. and,
Control information for multiplexer 7 to select image data from register 2 is read from read-only memory 10 by internal history information forming part of the address signal of read-only memory 10 , and the control information is transferred to register 11 . Set. Note that the aforementioned n and m can be set to 3, for example, and it is possible to determine whether the area is a character, line drawing area, or gray area using the 1×4 pixel gradation 1 to 3 pattern shown in FIG. can.

As mentioned above, it is determined whether it is a character, line drawing area or gray scale image area, and in the case of a character 2wA drawing area, the input image data is applied to the MH encoder 8 via the multiplexer 7, In the case of a grayscale image area, the input image data is subjected to predictive conversion and is applied as a smoothed prediction error signal to the MH encoder 8 via the multiplexer 7.
This is to perform MH encoding on image data in a gray area.

FIG. 4 is a block diagram of a decoding circuit according to an embodiment of the present invention, in which 21 is an input terminal for an MH code signal, 22 is an MH decoder (
DEC), 23 is a register (REG), 24 is an exclusive OR circuit, 25 is a register (REG), 26.27
is a shift register (SHR), 28 is an exclusive OR circuit, 29 is a multiplexer (MPX), 30 is an output terminal, 31 is a shift register (SHR), 32 is a read-only memory (ROM), 33 is a register (REG)
, 34 is a line counter (LCTR).

The received MH code signal is decoded by MH decoder 22 and sent to register 23 . It is added to the exclusive OR circuit 24 and shift register 27. The register 23 and the exclusive OR circuit 24 generate the decoded smoothed prediction error signal "
is converted into a prediction error signal e and added to the register 25. The shift register 26 is a 4-bit shift register, and the prediction error signal e and the prediction error signal e delayed by 4 bits are added to the exclusive OR circuit 28, and the prediction inverse conversion is performed. Then, the data of the gray area is outputted from the exclusive OR circuit 28 and applied to the multiplexer 29. Further, the shift register 27 is a 5-bit shift register for compensating for delays in the register 25 and shift register 26.

The data selectively output by the multiplexer 29 is sent out from the output terminal 30 as received restored image data, and is added to the shift register 31. This shift register 31 is a 4-bit shift register that converts the added 4-bit image data into 4-bit parallel data and adds the 4-bit parallel data as part of the address signal of the read-only memory 32. . Further, the lower two bits from the line counter 34 are added to the read-only memory 32 as an address signal indicating which line among the first line L1 to the fourth line L4.
The internal forward history information set in the read-only memory 32
is added as part of the address signal. The configuration including this read-only memory 32 is the same as the means for determining whether it is a character, line drawing area, or gray area in FIG. Control takes place.

Therefore, when it is determined that the area is a character or line drawing area, the image data decoded by the MH decoder 22 is transferred to the shift register 27.
When it is determined that it is a grayscale image area, the smoothed prediction error signal d is converted into a prediction error signal e, which is further subjected to predictive inverse conversion to become grayscale image data, and outputted through the multiplexer 29. be done.

Although the above-mentioned embodiment shows the case where a Bayer type dither matrix is used, the present invention can also be applied to cases where a halftone type dither matrix, a spiral type dither matrix, or the like is used. In addition, the image data is explained when the entire surface is dithered for an image containing a mixture of text, line art, and light and shade images, but the text and line art areas are binarized with a single darkness value, and only the light and shade areas are converted into dithered image data. The present invention can also be applied to data compression of image data.

〔Effect of the invention〕

As explained above, the present invention provides a read-only memory lO that determines whether it is a character, line drawing area, or gray area using black and white pattern data consisting of a plurality of pixels other than all black and all white based on dithered image data of a light and shade image. A determination means consisting of a line counter 12, etc., and this determination means selects the input image data when it is determined to be a character or &'i hundred area, or the smoothed prediction error signal when it is determined to be a gray area. It is equipped with a switching means such as a multiplexer 7 and an MH encoder 8, and performs M)I encoding on image data containing a mixture of characters, line drawing areas, and gray area, and enables efficient data compression. There is an advantage that

[Brief explanation of drawings]

Fig. 1 is a block diagram of an encoding circuit according to an embodiment of the present invention, Fig. 2 is an explanatory diagram of a gradation pattern based on a 4 x 4 Bayer dither matrix, and Fig. 3 is a 1 x 4 gradation pattern corresponding to a line. An explanatory diagram of the pattern, FIG. 4 is a block diagram of a decoding circuit according to an embodiment of the present invention, FIG. 5 is a block diagram of a conventional predictive MH method, and FIG. 6 is an explanation of reference pixels and target pixels of the predictive MH method. 7 are conversion explanatory diagrams of an example of the predictive MH method.

Claims (1)

[Claims] A determining means for determining whether the area is a character, line drawing area or a gray area based on black and white pattern data consisting of a plurality of pixels other than all black and all white based on dithered image data of a light and shade image; It is equipped with an MH encoder and a switching means that selects the input image data when the input image data is determined to be a grayscale image area, and a smoothed prediction error signal when it is determined that the image data is a grayscale image area. An image data compression method characterized in that data is MH encoded by the MH encoder.