CN1288599C - Method for transforming image dimension - Google Patents

Abstract

The present invention relates to a method for transforming image dimension. In the method for transforming image dimension, after the original image is converted to a half-tone image, the weighting amount of each ink dot is calculated by a weighting window, and the weighted calculation of the temporary storage weight of each ink dot is carried out by an error diffusion algorithm so as to generate the half-tone image after the transformation of the dimension. The method for transforming image dimension comprises the following steps: the data of the original image is converted to the half-tone image, and the half-tone image is formed by the arrangement of a plurality of ink dots; the predetermined transformation of the dimension of the weighting window is carried out, and the weighting amount of each ink dot in the half-tone image is calculated; the weighted calculation of the weight of each ink dot is carried out so as to generate the half-tone image after the transformation of the dimension.

Description

The method of picture size conversion

Technical field

The present invention relates to a kind of method of picture size conversion, be applied to the treatment scheme of the picture size conversion of half tone image.

Background technology

Along with making rapid progress of science and technology, digital camera has replaced traditional simple eye camera or idiot camera, and the main tool when becoming the modern and photographing.Because the user is after having taken image, image file in the digital camera can be sent in the personal computer, utilize printer or multifunctional paper feeding machine to carry out the output of image (Hard copy) again, and the size of image output also can be according to the different demand of user, and amplifies or dwindle.Therefore, the mode of utilizing printer or multifunctional paper feeding machine to carry out image (Hard copy) output, during compared to traditional processing, the user also needs film is taken the mode of photo flushing shop output, and sizable convenience is arranged.

Please refer to shown in Figure 1ly, in general printer or multifunctional paper feeding machine, the treatment scheme of its picture size conversion is:

In step 11, earlier with image document by the copy (copy) or the processing of scanning (scan), (suppose that its picture size is n * m) with the RGB that obtains this image document original size (three primary colors of light: R is red, G is green, B blue) image; Then, in step 12, with this RGB picture size conversion, (at this moment, its picture size has become N * M) to produce the RGB image of new size; Then, in step 13, the RGB image to new size carries out CMYK (printing three primary colors: C viridescent, M fuchsin, Y Huang and K (key) black) color conversion process again, to obtain the CMYK image in the step 14; Then, in step 15, again the CMYK image is carried out shadow tone (Half toning) and handle, to obtain the half tone image in the step 16; At last, in step 17,, promptly finish the flow process of entire image conversion and printing again with this half tone image output printer or multifunctional paper feeding machine prints.

And in the process of above-mentioned image transformation and subsequent color conversion process, be at the beginning the time, the RGB image that is about to original size carries out the conversion of picture size, then carries out follow-up cmyk color conversion, shadow tone again ... wait processing.

Yet, when picture size is amplified, need each pixel after amplifying is all carried out cmyk color conversion, shadow tone ... wait processing, therefore, the processing time of the time that is spent on follow-up treatment scheme than original image size is length.

When if picture size is amplified twice, then total treatment scheme need have more three times of times in the processing computing of color conversion and shadow tone.So, when the picture size amplification multiple is higher, need of the processing computing of cost more time in color conversion and shadow tone.

Summary of the invention

Because above-mentioned problem, fundamental purpose of the present invention is to provide a kind of method of picture size conversion, can effectively improve conversion efficiency.

To achieve these goals, the invention provides a kind of method of picture size conversion, comprise the following step: an original image information is converted to a half tone image, and this half-tone picture similarly is to be arranged by a plurality of ink dots to form; One weight allocation window through a default size change over, is calculated in this half tone image the respectively weight allocation of this ink dot again; And utilize error diffusion algorithm that the weight of this ink dot is respectively implemented ranking operation, to produce this half tone image behind the size change over.

The method of aforesaid picture size conversion, wherein, an original image information is converted in the step of a half tone image at this, when this original image information is a RGB image, then need this original image information is carried out the cmyk color conversion process, pass through halftone process again, so that this original image information is converted to this half tone image.

The method of aforesaid picture size conversion, wherein, an original image information is converted in the step of a half tone image at this, when this original image information is a CMYK image, then with its process halftone process, so that this original image information is converted to this half tone image.

The method of aforesaid picture size conversion, wherein, the zone that this weight allocation window is covered is corresponding to the position of this ink dot output in this half tone image.

The method of aforesaid picture size conversion, wherein, the size of this weight allocation window depends on the size and the amplification coefficient of this original image information window.

The method of aforesaid picture size conversion, wherein, at this weight allocation window being calculated in this half tone image the step of the weight allocation of this ink dot respectively and this again through a default size change over utilizes error diffusion algorithm that the weight of this ink dot is respectively implemented to carry out following steps between the step of ranking operation with this half-tone picture after producing size change over: the adding up to calculate the weight allocation of each pixel of weighted value of carrying out ink dot according to the degree of covering of this weight allocation window.

The invention has the beneficial effects as follows that the method is carried out cmyk color conversion, shadow tone to each pixel ... wait the time of handling required cost after can effectively saving down and earlier picture size being amplified in the prior art again.

And be after reducing the picture size amplification, produce the fuzzy situation of image frame, therefore, the present invention continues to use the halftoning method of error-diffusion method, weight to each ink dot is implemented ranking operation, the grey scale average value and the original image of present output pixel neighborhood are compared, and the error diffusion that will compare the back gained arrives down a bit, GTG value by following some original images deducts this error, make down after any original GTG value dims or lighten according to error amount, do the processing of quantification again, so that the weight of each point is handled and ink dot distributes more flexible, image frame is more level and smooth.

For making purpose of the present invention, structural attitude and function thereof are had further understanding, existing conjunction with figs. is described in detail as follows.

Description of drawings

Fig. 1 is in general printer or the multifunctional paper feeding machine, the treatment scheme of its picture size conversion;

Fig. 2 is the schematic flow sheet of the inventive method;

Fig. 3 is the method flow diagram of picture size conversion of the present invention;

Fig. 4 is original half tone image;

Fig. 5 is the result after Fig. 3 process weight allocation window calculation;

Fig. 6 is the ultimate principle of error diffusion algorithm;

Fig. 7 is the order of each pixel with the error diffusion algorithm computing;

Fig. 8 is the result of each pixel error diffusion calculation back gained; And

Fig. 9 is the half tone image after the original image of gained after the error diffusion calculation amplifies.

Wherein, description of reference numerals is as follows:

The A ink dot

The B ink dot

The C ink dot

A ' weight allocation window

B ' weight allocation window

C ' weight allocation window

Poly-11 RGB images of step

Step 12 RGB picture size convergent-divergent, the RGB image of generation new size

The RGB image of step 13 new size carries out the cmyk color conversion process

Step 14 CMYK image

Step 15 CMYK image carries out shadow tone (Half toning) to be handled

Step 16 half tone image

Step 17 printer output

Step 21 RGB image

Step 22 RGB image carries out the cmyk color conversion process

Step 23 CMYK image

Step 24 CMYK image carries out shadow tone (Half toning) to be handled

Step 25 half tone image

Step 26 through default size change over, is calculated the weight allocation of each ink dot in the half tone image with the weight allocation window again

Step 27 is implemented ranking operation with error diffusion algorithm to the weight of each ink dot, with the half tone image behind the generation size change over

Step 28 printer output

Step 31 is carried out the cmyk color conversion with the RGB image document, to produce the CMYK image document

Step 32 is carried out halftone process with the CMYK image document, and to produce half tone image, this half-tone picture similarly is to be arranged by several ink dots to form

Step 33 through a default size change over, is calculated the weight allocation of each ink dot in this half tone image with the weight allocation window again

Step 34 is implemented ranking operation with error diffusion algorithm to the weight of each ink dot, with the half tone image behind the generation size change over

Embodiment

As shown in Figure 2, be the schematic flow sheet of the inventive method, the step 21 of the method to step 25 is that general printer or multifunctional paper feeding machine are to the treatment of picture flow process; In step 21 and step 22, be that earlier (original image is of a size of n * m) and carries out the cmyk color conversion, and (size of images is n * m) to produce CMYK image document in the step 23 with the RGB image document of an original image; Then, again this CMYK image document is carried out halftone process in step 24, (size of images is n * m), and this half-tone picture similarly is to be arranged by several ink dots to form to produce half tone image in the step 25.

Then, step 26 with the weight allocation of each ink dot in this half tone image of weight allocation window calculation (at this moment promptly, size of images is enlarged into N * M), and the zone that this weight allocation window is covered is corresponding to the position that ink dot output is arranged in the original half tone image; Step 27 is implemented ranking operation with error diffusion algorithm to the weight of each ink dot again, promptly produces the half tone image behind the size change over; At last, in step 28 again with this half tone image output printer or multifunctional paper feeding machine prints.

In the explanation of above-mentioned steps, the size of weight allocation window depends on the size of the window size and the amplification coefficient of original image information, to change the dispersion degree of each ink dot.

Picture size transform method of the present invention can be applicable to multifunctional paper feeding machine/scanner/image output devices such as printer, so that a RGB image is first through after cmyk color conversion, the halftone process, produce a half tone image, then, again with the weight allocation of each ink dot in the weight allocation window calculation half tone image, then, with error diffusion algorithm the weight of each ink dot is implemented ranking operation again, can produce the half tone image behind the size change over, thus, can shorten the required time of whole image dimension conversion process.

Please refer to shown in Figure 3ly, be the method flow diagram of picture size transform method of the present invention, further method of the present invention is described as follows:

At first, the RGB image document is carried out the cmyk color conversion, to produce CMYK image document (step 31), digital picture in the general computing machine all is with rgb color mode treatment image, and when image will be exported (Hard copy), then needing image transitions be the color mode of CMYK, required ink color coupling with output the time.

Therefore, utilize modes such as copy/scanning earlier, obtain original RGB image document, again this RGB image document is carried out the conversion of cmyk color pattern, to produce the CMYK image document.

Then, the CMYK image document is carried out halftone process, to produce half tone image, this half-tone picture similarly is to be arranged by several ink dots to form (step 32).

Shadow tone is a kind of continuous GTG to be converted to the method for binary picture, and binary picture can allow people's eyes produce the illusion of the average GTG block of approximate black and white ratio through certain special arrangement.

And the treatment technology of halftone utilizes the density that black dots distributes on the unit area exactly, produces the illusion with different GTGs.So this step is that (step 31) middle CMYK image document that produces is obtained the half tone image of this image document through the halftone image processing techniques.

As shown in Figure 4, be the half tone image of certain image gained after the halftone image processing techniques, wherein, A, B, three lattice points of C are the positions that will stamp ink dot, and will be example with this half tone image after this, to explain.

Then, the weight allocation window through a default size change over, is calculated the weight allocation (step 33) of each ink dot in this half tone image again.

At first, the user sets a weight allocation window (Weight Allocation Window) earlier, the size of this weight allocation window depends on the size and the amplification coefficient of original image information window, for example: the original image information window is of a size of 3 * 3, and desire is when amplifying 2 times with entire image, and this weight allocation window size is 6 * 6.

And each pixel weight distribution calculation is that the weight allocation window is covered corresponding to the position that has ink dot to occur in the original half tone image, and the degree of covering according to each weight allocation window carries out adding up of weighted value again.

Please refer to shown in Figure 5ly, is Fig. 4 through the result after the weight allocation window calculation, and the weight allocation calculation specifications of each pixel are as follows:

The zone that weight allocation window (scope A ', B ', C ' that black dotted lines institute frame is lived) is covered, be corresponding to the position that has ink dot (A, B, C) to occur in the original half tone image, and the numeral in the lattice point (pixel) in this weight allocation window all is defined as 1; And the zone that has two weight allocation windows to repeat to cover, the numeral in its lattice point (pixel) all is defined as 2 (1+1=2); And the zone that has three weight allocation windows to repeat to cover, the numeral in its lattice point (pixel) all is defined as 3 (1+1+1=3) ..., by that analogy.

With error diffusion algorithm the weight of each ink dot is implemented ranking operation, with the half tone image (step 34) behind the generation size change over, error diffusion algorithm (Error Diffusion Method) is when handling certain picture point, the front error amount that processed images point had spread must be taken into account in the lump, make comparisons with a fixing critical value (Threshold Value) more at last, the energy of the half tone image that the method produces will concentrate on the part of high frequency, therefore, for the mankind's vision system, can produce visual effect (it is more level and smooth that image color is felt) preferably.

Please refer to shown in Figure 6ly, be the ultimate principle of error diffusion algorithm, and the meaning of each symbol representative be described as follows:

x _{I, j}: original image is in (i, j) locational GTG value;

b _{I, j}: (i, j) binary pixel of position;

v _{I, j}: (i, j) position is through the image gray-scale level value after the error correction;

T: critical value;

e _{I, j}: (i, j) quantization error of position;

H (i, j): the error diffusion wave filter;

Binary pixel b _{I, j}Value be by v _{I, j}Make comparisons with a fixing critical value t and obtain, if v _{I, j}Value greater than critical value t, then b _{I, j}Be output as 1, on the contrary, if v _{I, j}Value less than critical value t, then b _{I, j}Be output as 0.

And e _{I, j}Be b _{I, j}With v _{I, j}Between quantization error, (i after processing j), can feed back to next pixel through error diffusion wave filter h for this error.Error diffusion wave filter h (i, j) effect is that the error of current pixel position neighbor point is made weighted mean, if error diffusion wave filter h is (i, j) output valve is greater than zero, grey scale average value and the original image of then representing present output pixel neighborhood are bright excessively by comparison, otherwise, then represented dark.

And be the difference that remedies this image appearance, then with error diffusion to down a bit, deduct this error by the GTG value of some original images down, make down after any original GTG value dims or lighten according to error amount, do the processing of quantification again, so that the binary picture b that is exported _{I, j}Whole GTG value be similar to original image x _{I, j}Whole GTG value.

And above-mentioned notion can be represented with following mathematical expression:

$v_{i,j}=x_{i,j}-\underset{m,n}{\mathrm{\Σ}}{h}_{m,n}{e}_{i-m,j-n}$

$b_{i,j}=\left\{\begin{array}{ccc}1& \mathrm{if}& v_{i,j}>t\\ 0& & \mathrm{otherwise}\end{array}\right.$

e _i，j＝b _i，j-v _i，j

Whether please refer to shown in Figure 7ly, is the order of each pixel with the error diffusion algorithm computing, right by a left side before this, more from top to bottom, went to calculate each pixel and will export.And the order of error diffusion algorithm computing can have difference according to user's demand, only exposes arithmetic speed order of operation faster among this embodiment.

Please refer to shown in Figure 8ly, is the result of each pixel error diffusion calculation back gained, the step of following specification error diffusion calculation:

At first, set earlier critical value t by the weight allocation window peaked 1/2nd (this example is 9/2=4.5) of energy accumulated weight, therefore, critical value t is 4.5.And each pixel is added the value (shown in Fig. 5 is through the value after the weight allocation window calculation) of itself for the value of a pixel on it.

With first row (from left side number) second pixel, its value is the value 1 that 1 (value of a last pixel) adds itself, and therefore, its value is 2 (less than critical value t:4.5), therefore, and its binary pixel b _{I, j}Be zero (not exporting); With first row (left side number) the 3rd pixel, its value is the value 1 that 2 (values of a last pixel) add itself, and therefore, its value is 3 (less than critical value t:4.5), therefore, and its binary pixel b _{I, j}Be zero (not exporting) ...; With the 5th pixel of first row (left side number comes), its value is the value of a pixel on 4) add value 1 own, therefore, its value is 5 (greater than critical value t:4.5), therefore, its binary pixel b _{I, j}Be 1 (will export); And with the 6th pixel of first row (left side number comes), its value is the value 1 that 5 (values of a last pixel) add itself, (the specification error value is that the weight allocation window is of a size of 3 * 3=9) and equals-3 (less than critical value t:4.5) to deduct the error amount 9 that some diffusion comes again, therefore, its binary pixel b _{I, j}Be zero (not exporting) ..., by that analogy.

Please refer to shown in Figure 9ly, promptly is the half tone image after original image by error diffusion calculation back gained amplifies, by among the figure as can be known: the method for this picture size conversion is not only to utilize duplicating of lattice point and image after obtaining amplifying as prior art.Whether the method combines the calculating of weight allocation window and error diffusion algorithm, will export to judge each pixel, therefore, can make image after the amplification more similar in appearance to original image, and output effect more level and smooth (smooth).

Above-mentioned flow process is that change in size with general RGB image document is to explain; Certainly, the method of the disclosed picture size conversion of the present invention can directly apply in the picture size conversion of half tone image, only need utilize the weight allocation of each ink dot in this half tone image of weight allocation window calculation earlier, with error diffusion algorithm the weight of each ink dot is implemented ranking operation again, can produce the half tone image behind the size change over.

The above is preferred embodiment of the present invention only, is not to be used for limiting practical range of the present invention; All equalizations of doing according to claims of the present invention change and modify, and all are included in the claim of the present invention.

Claims (6)

1. the method for a picture size conversion is characterized in that, comprises the following step:

One original image information is converted to a half tone image, and this half-tone picture similarly is to be arranged by a plurality of ink dots to form;

One weight allocation window through a default size change over, is calculated in this half tone image the respectively weight allocation of this ink dot again; And

Utilize error diffusion algorithm that the weight of this ink dot is respectively implemented ranking operation, to produce this half tone image behind the size change over.

2. the method for picture size conversion as claimed in claim 1, it is characterized in that, an original image information is converted in the step of a half tone image at this, when this original image information is a RGB image, then need this original image information is carried out the cmyk color conversion process, pass through halftone process again, so that this original image information is converted to this half tone image.

3. the method for picture size conversion as claimed in claim 1, it is characterized in that, an original image information is converted in the step of a half tone image at this, when this original image information is a CMYK image, then with its process halftone process, so that this original image information is converted to this half tone image.

4. the method for picture size conversion as claimed in claim 1 is characterized in that, the zone that this weight allocation window is covered is corresponding to the position of this ink dot output in this half tone image.

5. the method for picture size conversion as claimed in claim 1 is characterized in that, the size of this weight allocation window depends on the size and the amplification coefficient of this original image information window.

6. the method for picture size conversion as claimed in claim 1, it is characterized in that, at this weight allocation window being calculated in this half tone image the step of the weight allocation of this ink dot respectively and this again through a default size change over utilizes error diffusion algorithm that the weight of this ink dot is respectively implemented to carry out following steps between the step of ranking operation with this half-tone picture after producing size change over: the adding up to calculate the weight allocation of each pixel of weighted value of carrying out ink dot according to the degree of covering of this weight allocation window.

Images