JP2797950B2 - Predictive coding device - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a predictive coding apparatus for compressing image data of a facsimile apparatus, and more particularly to a predictive coding apparatus for compressing image data read from a color dot image.

[0002]

2. Description of the Related Art Conventionally, when predictive encoding processing is performed on image data read from a halftone image, as shown in the example of FIG. Efficient prediction has been performed by using the pixels A to F at a portion separated by the distance as reference pixels, and a high compression ratio has been realized. Generally, halftone dots having different angles corresponding to the number of color separations are used for the color halftone dots (15 °, 75 °, 90 °, 45 °).
etc).

[0003] When an attempt is made corresponding to the above-described method for color halftone images, as shown in FIG. 4, reference pixels A
.About.F must be moved to a position corresponding to the halftone dot angle .theta. / Halftone dot interval T, so that an identification unit for identifying the halftone dot angle and halftone dot interval from the image data is required.

FIG. 4 shows an example of identification of a halftone dot angle / halftone dot interval. FIG. 5 shows an example in which two types of halftone dot angles, θ1 and θ2, and two types of halftone intervals, T1 and T2, are identified. In the figure, observation pixels A ′ to D ′ are provided at the center of each halftone dot, a correlation between each observation pixel and the processing target pixel X is obtained, and a halftone dot angle and halftone dot interval of the pixel having the highest correlation are output. .

However, in such a method, since it is necessary to arrange observation pixels two-dimensionally, in the example of FIG. 5, data of at least L lines is required.
When the data amount is less than the line amount, there is a problem that accurate identification cannot be performed.

[0006]

As described above, in the conventional method, data of a plurality of lines is required for discriminating a dot angle and a dot interval.
There is a drawback that accurate identification cannot be performed.

SUMMARY OF THE INVENTION The present invention has been made in view of the above-mentioned conventional circumstances, and accordingly, an object of the present invention is to provide a novel predictive coding apparatus which can solve the above-mentioned drawbacks inherent in the prior art. Is to do.

[0008]

In order to achieve the above object, a predictive coding apparatus according to the present invention comprises a predicting section for predicting a state of a pixel at a point of interest based on input image data, and an output of the predicting section. In a predictive coding apparatus having a run length conversion unit for converting a signal into a code assigned according to a run length occurrence probability, a halftone dot angle and a halftone dot interval are automatically identified from the image data having a data amount of one line or less. And a prediction unit that performs highly efficient prediction based on the dot angle / dot interval information identified by the identification unit.

[0009]

Next, a preferred embodiment of the present invention will be described in detail with reference to the drawings.

FIG. 1 is a block diagram showing an embodiment of the present invention. One embodiment of the present invention is an identification unit 3 for identifying a halftone dot angle and halftone interval from the input image data 2 even with a data amount of one line or less, and a halftone angle / halftone interval signal output by the identification unit 3. 4 and a prediction unit 5 for predicting the value of the pixel of interest from the input image data 2 and an encoding unit 7 for converting the output signal of the prediction unit 5 into a code assigned according to the run-length occurrence probability. You.

Next, the operation of this embodiment will be described.

First, image data 2 is input to an input terminal 1. The image data 2 is input to the identification unit 3. The identification unit 3 identifies a dot angle and a dot interval from the input image data 2 and outputs a dot angle / dot interval signal 4. A method of identifying the halftone dot angle / halftone interval in the identification unit 3 will be described later in detail. The prediction unit 5 predicts a pixel of interest from the image data 2 and the dot angle / dot interval signal 4 and outputs a signal (prediction error signal 6) as to whether the predicted value matches the actual pixel value of interest. I do. The encoding unit 7 performs run-length encoding on the prediction error signal 6, and outputs a code 9 from an output terminal 8.

The details of the method for identifying the halftone dot angle and halftone dot interval in this embodiment will be described below.

FIG.6(A) is a dot angle θ1 and a dot interval T.
The example of the position of the center of each halftone dot in the case of 1 is shown.
Figure6(B) is a figure6Pixel on x-axis in case (A)
FIG. 7 is a schematic diagram showing the correlation strength between a pixel of interest and FIG. same
Like figure7(A) is the case of the dot angle θ2 and the dot interval T1.
3 shows an example of the position of the center of each halftone dot. Figure7
(B) is a figure7Pixels on the x-axis in case (A)
It is a schematic diagram showing the intensity of correlation between eye pixels.

[0015] FIG. 6 (B), the As apparent from comparison to FIG. 7 (B), the strength of the correlation between each pixel and the target pixel on the x-axis indicates the specific pattern depending on the angle. For this,
The angle can be specified from the pattern of the strength of correlation between each pixel on the x-axis and the pixel of interest.

[0016] FIG. 8 (A) dot angle θ1 · network point interval T
The example of the position of the center of each halftone dot in the case of No. 2 is shown.
Figure 8 (B) is a schematic view showing a pixel on the x-axis, the magnitude of correlation between the target pixel in the case of FIG. 8 (A).

[0017] FIG. 6 (B), the As is clear from comparison FIG 8 (B), the position with the pixels on the higher x-axis correlation between the target pixel is changed by the network point distance. For this reason, from the pattern of the strength of correlation between each pixel on the x-axis and the pixel of interest,
It is possible to specify a halftone dot distance.

Next, the details of the identification section 3 in this embodiment shown in FIG. 1 will be described.

The image data 2 input to the identification unit 3 is shifted by the shift register 10. The image data 2 can be generated by scanning the document according to a scanning line obtained by crossing a plurality of lines of the document. X is a pixel of interest, and ad are positions of observation pixels. The values of the observed pixels a to d are input to the value of the target pixel and XOR (exclusive OR) gates 11 to 14, respectively. XOR gate 1
Output values of 1 to 14 are up / down (U / D) respectively
It is counted by the counters 15-18. The up / down counters 15 to 18 count up when "H level" is input, and count down when "L level" is input.

Outputs of the up / down counters 15 to 18 are output from corresponding first and second comparators 19 and 20;
2, 23, 24; 25, 26, respectively. At the same time, the first comparators 19, 21, 23, 25 have the reference value N
However, the second comparators 20, 22, 24, 26 have reference values M
Are respectively input. The outputs of the comparators 19 to 26 are input to the judgment circuit 27.

When the correlation between the observed pixel and the pixel of interest is high, the probability that the value of the pixel of interest is uniquely determined according to the value of the observed pixel increases. That is, when the value of the observed pixel is “H level”, the value of the target pixel is also likely to be “H level”.
Alternatively, when the value of the observation pixel is “H level”, there is a high probability that the value of the pixel of interest becomes “L level”. Therefore, when the value of the observed pixel and the value of the pixel of interest are input to the XOR gate,
When the correlation between the observed pixel and the target pixel is high, the probability that the output of the XOR gate is biased to either “H level” or “L level” increases.

When the output from the XOR gate is input to an up / down counter (counting up by "H level" input and counting down by "L level" input), the correlation is high when a certain amount of image data is input. In this case, the absolute value of the counter output increases, and when the correlation is low, the absolute value of the counter output decreases.

In this embodiment, the outputs of the up / down counters are compared with the reference values M and N by comparators 19 to 26 (M> N). Thus, the counter output can be divided into three stages of M or more, N or more and less than M, and less than N. The comparison results from the comparators 19 to 26 are input to the judgment circuit 27. The determination circuit 27 uniquely determines the halftone dot angle and halftone dot interval from the comparison result of each comparator (which rank of the observed pixel is in the rank of M or more, N or more and less than M or less than N). , And outputs a dot angle / dot interval signal 4.

FIG. 2 is a diagram showing a specific example of the judgment circuit 27. As shown in FIG. 2, the decision circuit 27 is an AND circuit,
It is composed of a logic circuit formed by a combination of circuits and the like.

Next, an example of the judgment circuit 27 will be described in detail.

Two types of dot angles α 1 and α 2, dot interval D
1 and D2, an example of a judgment circuit for discriminating two types will be described. In FIG.
Table 1 shows the case where the correlation strength between the observed pixel at the position of c and d and the pixel of interest is (a), (b), (c), and (d).
Let's say

[0027]

[Table 1]

In Table 1, the magnitudes of the large, medium, and small correlations are such that the absolute value of the output of the U / D counter when a certain amount of data flows is large: M or more, medium: N or more and less than M, and small. : Less than N

The comparators 19 to 26 assume "H level" when the input is larger than the reference value and "L level" when the input is smaller than the reference value. Is as shown in Table 2.

[0030]

[Table 2]

In Table 2, the dot angle signal "L" = α1
, "H" =. Alpha.2 The halftone dot interval signal "L" = D1, and "H" = D2.

The decision circuit 27 shown in FIG. 2 has a relation shown in Table 2 constituted by a logic circuit. Although FIG. 2 does not show the case (d), in this case (a),
Except for the input conditions (b) and (c), the dot angle signal 4
1. Since all the halftone dot interval signals 42 become "H",
This is because the input condition in the case of (d) is satisfied.

The logic circuit shown in FIG. 2 is only an example, and it is needless to say that the logic circuit can be variously changed according to the input conditions (a) to (d).

The identification unit of this embodiment monitors the correlation strength between each pixel in the line including the target pixel and the target pixel, and specifies the halftone dot angle and halftone dot interval based on the principle described above. . For this reason, the halftone dot angle / halftone dot interval can be identified even with a data amount of one line or less, and the disadvantage that accurate identification cannot be performed until a plurality of lines are input as in the conventional method is improved. You.

[0035]

As described above, according to the present invention,
By monitoring the pattern of the strength of correlation between each pixel in the one line including the target pixel and the target pixel and identifying the halftone dot angle and halftone dot interval, the halftone dot angle / halftone can be obtained even with a data amount of one line or less. The effect of identifying the point interval and reducing the code amount is obtained.

[Brief description of the drawings]

FIG. 1 is a block diagram showing an embodiment of the present invention.

FIG. 2 is a block diagram illustrating a specific example of a determination circuit illustrated in FIG. 1;

FIG. 3 is a diagram illustrating an example of a reference pixel arrangement of a conventional predictive coding method.

FIG. 4 is a diagram illustrating an example of a reference pixel arrangement in a case where a conventional predictive coding method is associated with a color halftone dot.

FIG. 5 is a diagram showing an example of a conventional halftone angle / halftone interval discrimination method.

FIG. 6A shows an example of the position of each dot center in the case of a dot angle θ1 and a dot interval T1. FIG. 7B is a schematic diagram illustrating the correlation strength between the pixel on the x-axis and the pixel of interest in the case of FIG.

FIG. 7A shows an example of the position of the center of each halftone dot in the case of a halftone dot angle θ2 and a halftone dot interval T1. FIG. 8B is a schematic diagram illustrating the correlation strength between the pixel on the x-axis and the pixel of interest in the case of FIG.

FIG. 8A shows an example of the position of each dot center in the case of a dot angle θ1 and a dot interval T2. FIG. 9B is a schematic diagram illustrating the correlation strength between the pixel on the x-axis and the pixel of interest in the case of FIG.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 ... Input terminal 2 ... Input image data 3 ... Identifier 4 ... Dot angle / dot interval signal 41 ... Dot angle signal 42 ... Dot interval signal 5 ... Prediction unit 6 ... Prediction error signal 7 ... Encoding unit 8 ... output terminal 9 ... output code 10 ... shift register 11-14 ... XOR gate 15-18 ... up / down counter 19-26 ... comparator 27 ... judgment circuit

Claims (3)

(57) [Claims] 1. A prediction unit for predicting a state (1 or 0) of a pixel at a point of interest based on input binary image data, and converting an output signal of the prediction unit into a code assigned based on a run-length occurrence probability. A predictive encoding apparatus having a run-length transform unit that performs
Before being shifted by the shift register
The corresponding observation pixel and target pixel in the image data
Of exclusive OR gates equal to the number of observation pixels
Input and the corresponding gate output of the exclusive OR gate
Up / Down cards of the same number as the exclusive OR gate
Counter and a corresponding counter output of the counter and a first
The up inputting a reference value and a second reference value respectively
And the same number of first and second comparators as the
A halftone dot angle / halftone dot based on the outputs of the plurality of first and second comparators
A determination circuit for determining and outputting an interval signal; and an identification means for automatically identifying a halftone dot angle and a halftone dot interval from a data amount of one line or less of the image data, and a network identified by the identification means. A predictive coding apparatus comprising predicting means for performing highly efficient prediction based on point angle / halftone dot interval information. 2. The predictive encoding apparatus according to claim 1, wherein the image data is obtained by scanning the original in accordance with a scanning line obtained by crossing a plurality of lines of the original. 3. The predictive coding apparatus according to claim 1 , wherein said judging circuit is constituted by a logic circuit formed by an AND circuit, an OR circuit, and the like.