JP2848569B2 - Image data binarization method and image processing apparatus - Google Patents

Description

Description: BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a binarization method for image data and an image processing apparatus using the binarization method.

[Prior Art] Conventionally, as this kind of pseudo halftone processing method, an error diffusion method and an average density approximation method (Japanese Patent Laid-Open No. 57-1043693) are known.

In the former, the multi-valued pixel data of the target pixel is
This is a binarization method of binarizing, distributing an error between the binarization level and the multi-valued pixel data of the target pixel to a pixel near the target pixel at a predetermined distribution ratio, and adding them. The latter uses the already binarized binary data in the vicinity of the target pixel to convert the target pixel to black or white.
In this method, the weighted neighborhood average value of the two values is calculated, and the average value closer to the target pixel level is selected to binarize the target pixel.

The former requires a two-dimensional operation on multi-valued data,
It has the disadvantage that it cannot be hardened inexpensively due to its large amount of processing, and the latter has a low-frequency texture peculiar to this method for an image having a gradual density change, and the gray scale that can be substantially reproduced The number is extremely inferior to the former.

[Problem to be Solved by the Invention] Therefore, an average density preservation method in which a single average density is obtained, the image is binarized using the value as a threshold value, and a binarization error correction process is added. A new binarization method to be called has been considered.

However, when binarizing an image having a uniform density level over a large area, such as a CG image, the binarization is performed in a very regular pattern in which black or white lines are arranged particularly at a low node or a high node. There are drawbacks that are very unnatural.

The present invention eliminates the above-mentioned conventional drawbacks and provides image data having a uniform density level over a wide area, such as a CG image, for realizing high-quality binary reproduction without a regular pattern. Provided is an image processing apparatus that realizes high-quality binarization reproduction without a binarization method and a regular pattern.

[Means for Solving the Problem] To solve the problem, the image data of
The binarizing method pseudo-randomly changes an average value obtained based on binarized data in a predetermined range that has already been binarized in the vicinity of the pixel of interest, and calculates the pixel of interest by the pseudo-randomly changing average value. Is characterized in that the multi-valued image data is binarized.

Further, the image processing apparatus of the present invention is an image processing apparatus for binarizing multi-valued image data by an average value based on binarized data in a predetermined range which has already been binarized, and generates a pseudo random number. It is characterized by comprising random number generating means and average value correcting means for correcting the average value based on the pseudo random number.

Hereinafter, an embodiment of the present invention will be described in detail with reference to the accompanying drawings.

<Example of Configuration of Image Processing Apparatus> FIG. 1 is a block diagram showing the configuration of the image processing apparatus of the present embodiment.

The multi-valued image data 100a such as image data read from a document input from the image data input unit 100 is binarized by the binarization processing unit 200 as pseudo halftone processing, and is displayed on the display screen by the image data output unit 300. Alternatively, it is output as a printed material or the like. Here, as described in detail below, the binarization processing unit 200 includes an average density calculation processing unit 200a that calculates an average density as a binarization threshold and an error diffusion processing unit 200b that performs error correction processing. The integrated processing is performed. Hereinafter, in this embodiment, this processing method is referred to as an average density storage method.

In this embodiment, the binarization processing unit 200 further includes a random number generation unit 200
The random number from the random number generation unit 200c prevents a regular pattern of the binarization result.

<Principle of the average concentration preservation method> Regarding the principle of the average concentration preservation method of this method, FIGS. 2A to 2C
This will be described with reference to the drawings.

This method uses the input multi-valued data f (i, j) shown in FIG. 2A.
When it is desired to binarize (0-255), as shown in FIG. 2B, a plurality of binary data B (i, j) located in the vicinity thereof and already binarized are The weighted average value m (i, j) is obtained by using the weight mask R (x, y) prepared in advance.
, And binarized using the average value m (i, j) as a threshold, and the average value m (i, j) and the input image data f (i, j).
This is a method of correcting adjacent input multi-valued data to be binarized from now on with a difference value from j) and storing the density.

If shown in the example shown in FIGS. 2A to 2C, B (i, j) = 1 when f (i, j) + E (i, j)> m (i, j)
When f (i, j) + E (i, j) ≦ m (i, j), B (i, j) = 0
To be However, E (i, j) = E ₁ (i, j) + E ₂ (i, j) The above principle will be described with reference to FIG.

The weight mask used is, as shown in FIG.
The sum is set to 255 for the pixel. Therefore,
m (i, j) is 2 as it is as a threshold value between 0 and 255
Can be used for pricing. Now, the threshold value m (i, j) for binarizing the multi-value data 20 at the target pixel position in FIG.
If the value data B (i, j) is used, B (i, j) = (37 × 0 + 27 × 1) + (16 × 0 + 27 × 0 + 37 × 1 + 27 × 0 + 15 × 1) + (6 × 0 + 16 × 0 + 27 × 0 + 15 × 0 + 5 × 1) = 84 Therefore, B (i, j) is binarized to 0, the error that occurs is 20−84 = −64, and E ₁ (i, j + 1) = E ₂ (i + 1, j) )
= −32.

Due to this error diffusion, f (i + 1, j) is 25 to 25-3
2 = −7 and f (i, j + 1) is corrected from 30 to 30−32 = −2.

<Principle of the binarization processing of this embodiment> In this embodiment, when the above processing is performed on an output multi-valued image such as a CG (computer graphics), uniform and constant multi-valued data in a planar shape is obtained. In order to solve the disadvantage that the binarization results in a binary image having an unnatural texture that is orthogonal to the regular pattern, the dither signal P is generated in accordance with the output from the pseudorandom number generator during the binarization according to the above equation. _N (i, j) is assigned.

That is, when f (i, j) + E (i, j) + P _N (i, j)> m (i, j),
(I, j) = 1 f (i, j) + E (i, j) + P _N (i, j) ≦ m (i, j), B
(I, j) = 0... E ₁ (i, j + 1) = 1/2 {f (i, j) + E (i, j) + P _N (i, j) −m (i, j)} = E ₂ (I + 1, j) where E (i, j) = E ₁ (i, j) + E ₂ (i + j).

The pseudo-random number generator used in this embodiment is a known M-sequence code (SWGolomb, “Shift-Register Sequences” Holden).
-Day, Inc., San Francisco, 1967) and 400 A4 manuscripts
In order not to generate texture based on the periodicity when processing at dpi, one cycle is set to 2 ²⁵ -1 and ± δ corresponding to 0 and 1 of the 1-bit output (δ is 2 to 4 levels)
With a density level range of

<Example of Configuration of Binarization Processing Unit> An example of a hardware configuration of the binarization processing unit according to the present embodiment will be described as a fourth example.
This will be described in detail with reference to the drawings.

First, the binary data binarized by the comparator 10 is inputted to a D-type flip-flop (hereinafter referred to as D / F / F) 3i and a line memory 2, and is sequentially subjected to ten D / F / Fs 3a to 3j and By shifting to the line memory 1, each output terminal becomes the third
The twelve binary data of the weight mask area shown in the figure are output simultaneously.

The average density calculating unit 8 which receives the output outputs the average value m (i, j) based on a table conversion method if the average density is calculated and stored in advance in a ROM in which the average density is calculated using a weight mask. ) Is obtained. The output is input to one input terminal of the comparator 10 to be a threshold,
The subtraction according to the equation is performed by the subtractor 9. Its output is the error RO
As shown in formula in is inputted to the M12 said ROM, E ₁ and E ₂ 2
Then, E ₁ (i, j) is input to the error memory 14 and held for a delay corresponding to about one line until binarization of the next line.
On the other hand, E ₁ (i + 1, j) is added to the adder 1 together with E ₁ (i + 1, j) output from the same error memory 14 already delayed and held for one line.
3, the input data f (i + 1, j) is corrected.

The pseudo-random number generation unit 19 surrounded by a broken line is composed of 25 1-bit input / output DF / F 17 _{-1 to} 17 _-25 and three exclusive ORs (EXO).
_{R) 18 -1, 18 -2,} 18 composed of _a-3 and a multiplexer 16. The outputs of the 25 D / F / Fs are prevented from being all 0 by a preset circuit (not shown).
Therefore, the M-sequence code for one cycle T = 2 ²⁵ -1 is DF / F
_{Available at 17-25} output terminals. The multiplexer 16 converts the value into a predetermined number of bits of ± δ (± 2 to 4 in this embodiment) according to the state of the pseudo random number of 0 and 1 of the same 1 bit. F (i) together with the correction data
+1 and j).

This random number generation unit is applied to an input image with an average value of 0 without generating periodicity when binarizing an A4 document.
Even when viewed as the average density, the image is not disturbed. In particular, when the CG image having a uniform and the same level in the form of a plane is binarized, a binary image having irregular regularity can be obtained. The data corrected by the random number is input to the subtracter 9 and the comparator 10 at the timing adjusted by the D / F / F 11, and the next pixel is binarized.

The above processing is executed repeatedly for each pixel in synchronization with the pixel clock applied to all the F / Fs (not shown).

Although the <another configuration example of the binarization processing unit> The example was granted dither signal P _N of ± [delta] based on the random number to the input image data, the process proceeds P _N (i, j) to the right side in the equation Similarly, the same effect can be obtained even when the equation is satisfied, that is, when the threshold value m (i, j) is added.

FIG. 5 is an embodiment obtained by adding P _N in the threshold value side, the output dither signal P _N of the embodiment similar to the pseudo-random number generator 19, the average density m (i at the adder 15, j). This embodiment has the advantage of reducing the load on the adder 13.

<About the value of ± δ> The amplitude δ of the dither signal is generally reduced in accordance with the value of the input data f, that is, by decreasing the amplitude f when f is small. Can be constant.

<Another Example of _PN Generator> In the present embodiment, the synchronization is set to be large compared to the data amount at the time of processing A4 original at t = 2 ²⁵ -1. For example, T = 2 ¹⁰ −
If the hardware scale is reduced to about 1 and the D / F / F 17 is preset with other data such as the absolute address of the pixel of interest or the image multi-value data itself during binarization of the image, It is possible to artificially lengthen the period T.

[Effects of the Invention] According to the present invention, for an image having a uniform density level over a wide area, such as a CG image, image data 2 for realizing high-quality binary reproduction without a regular pattern.
An image processing apparatus that realizes high-quality binarization reproduction without a binarization method and a regular pattern can be provided.

In other words, despite the small amount of computing devices, ED
It is possible to provide a pseudo halftone processing device capable of improving image quality as compared with the conventional method. Also, by giving a simple pseudo random number, a binary image in which the regularity of the CG image is disturbed can be obtained.

[Brief description of the drawings]

FIG. 1 is a block diagram showing the configuration of the image processing apparatus of the present embodiment, FIGS. 2A to 2C are diagrams for explaining an average error saving method, FIG. 3 is a diagram for explaining average density calculation, and FIG. FIG. 5 is a diagram illustrating a hardware configuration of a binarization processing unit according to the present embodiment. FIG. 5 is a diagram illustrating a hardware configuration of a binarization processing unit according to another example. In the figure, 100... Image data input section, 200... Binarization processing section, 200a... Average density calculation processing section, 200b... Error diffusion processing section, 200c... Random number generation section, 300. Department.

Continuation of the front page (56) References JP-A-63-209370 (JP, A) JP-A-63-102473 (JP, A) JP-A-63-155952 (JP, A) JP-A-58-186265 (JP) JP-A-53-136424 (JP, A) JP-A-1-276969 (JP, A) (58) Fields investigated (Int. Cl. ⁶ , DB name) H04N 1/40-1/409 H04N 1/46 H04N 1/60

Claims (2)

(57) [Claims] An average value obtained based on binarized data in a predetermined range which has already been binarized in the vicinity of a pixel of interest is pseudo-randomly changed, and the average value which changes pseudo-randomly is used as the average value of the pixel of interest. Binarizing the multi-valued image data. 2. An image processing apparatus for binarizing multi-valued image data with an average value based on binarized data in a predetermined range which has already been binarized, comprising: random number generating means for generating pseudo random numbers; An image processing apparatus comprising: an average value correcting unit that corrects the average value based on the pseudo random number.