JPS5810963A - Picture compensation processing method - Google Patents

Abstract

PURPOSE:To obtain an excellent compensated picture, by referencing a noted picture element and picture elements around the noted picture element and performing an MTF compensation to the noted picture element only when the change in density at a limited small area is a specific value or over, in the picture compensation processing method for multi-value quantized picture. CONSTITUTION:An input picture signal obtained by sampling an original picture with a scanner is quantized into 8-value at a quantizer 100 and stored in an input buffer 101 temporarily. An MTF compensation circuit 105 fetches a noted picture element Q7 at that time and surrounding picture elements Q3, Q6, Q8 and Q11 from the input buffer 101 to output the MTF compensation level Q7' of the noted picture element. An adder 102 fetches the noted picture element Q7 and the surrounding picture elements Q1-Q6, Q8-Q13 from the buffer 101 and obtains the density level sum of the picture elements including the noted picture element and that for three picture elements not including the noted picture element and the discrimination of the change in the density is made at a comparison discrimination circuit 103. When the result of discrimination is effective, the MTF compensation level Q7' is given to an output buffer 106 and if not, the level Q7 is given.

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to image correction for sharpening multilevel quantized images.

Digital copying machines scan the original, sample the pixel density, and then further (=multilevel quantization) the pixel density.
A multi-level quantized image is obtained from the second step, and the necessary processing is performed on it.

M T F (Mod
ulation Transfer'Functi
on) is generally around 40*(6tp/wx) when a CCD image sensor is used, so (-MTP correction is often applied to improve image quality).

Currently, MTF correction is generally performed using a digital filter. FIG. 1(a) shows an example of the filter. The level X' after TF correction by this filter becomes X'=3xX-0,5x (B+D+B+G). Actually, this value X' is normalized before being output.

By the way, MTF correction is originally intended to reduce the halftone level of an image and make the image clearer. Therefore, it is extremely effective to use line images such as character images twice. On the other hand, when MTF correction is applied to halftone images such as photographs, undesirable phenomena such as white spots are easily removed.

° In fact, in multi-level quantized images obtained by reading and quantizing originals in the reading section of digital copying machines, even if the image parts have the same density, there are variations in the density level of about ±1000 lbs., including quantization noise. It is normal to see (density unevenness). Therefore, when MTF correction is applied, granular noise occurs or edge portions at halftone levels: two-white omissions occur, so that it is often impossible to obtain a good halftone image.

For example, when MTF correction is applied to the 8-value quantized image shown in FIG. 2 using the filter shown in FIG. 1(a), the corrected image becomes as shown in FIG. 3. If you compare and observe Figures 2 and 3, you will understand that a phenomenon occurs in the large white areas.

Considering the above-mentioned characteristics of MTF correction, it has been considered that in digital copying machines and the like, it is possible to specify in advance whether or not to perform MTF correction depending on the type of document to be handled. However, it is normal for general documents to contain a mixture of 4-character image parts and halftone image parts, and it is difficult to specify the type of original (= therefore, the presence or absence of MTF correction in advance) in most original documents. MTF correction cannot be applied to the data, and there is no point in providing an MTF correction function.

Therefore, it is an object of the present invention to be applicable to images containing a mixture of text and photographs, and to reduce the effects of variations in reading sensitivity and quantization noise of images to be processed.
An object of the present invention is to provide an image correction processing method capable of reproducing high-quality images.

If you look closely at a multilevel quantized image, you will notice that in general (=, the density level changes rapidly in text and line drawing parts, but the density changes in halftone images such as photographs) are much more gradual. The present invention utilizes such properties to
This is an attempt to selectively apply TF correction.

That is, the main features of the image correction processing method according to the present invention are 1. A pixel of interest (pixel of interest) of a multilevel quantized image and its surrounding pixels are referred to, and a limited small area (-) containing the pixel of interest is The density change is checked, and only when a sudden density change exceeding a certain degree is detected, the pixel of interest (=M
The purpose is to perform TF correction.

A specific example of the image correction process according to the present invention will be described below.

Now, when focusing on a certain pixel of an n-value quantized image obtained by a known method, let the pixel of interest be Ql in FIG. 4, and the surrounding pixels be Ql-Q6 and Qs to Q13.

In order to determine whether or not to perform the target pixel Q71-MTF correction, the following T1-T12 for pixels Q1 to Q13 are applied.
Calculate the value of

T+ , == Q2 +Q3 +Q4T2 =Q6
+Q7 +Qs T3・-Q10+ Qll + Q12T4 = Q
2 +Q6 +Q10 Ts -Q3 +Q7 +Q11 T6 = Q4 +Qs +Q12 T7=Qro-+ Q6 + Q3 Ts =Q1o+Q7+Q4 T9 = Q11+Q8 +Qs Too = Ql +Q6 +Q11 T11= Q2 +Q? +Ql2.

Tl2 = Q3 +Q8 + Q13 Tl s above
T2 e T3 is the sum of the density levels of three pixels arranged in the horizontal direction, T4 * T5 * T6 is the sum of the density levels of three pixels arranged in the vertical direction, T7 . T8゜T9 is the sum of the density levels of three pixels arranged diagonally to the right, T
10 * T11 e T12 are respectively diagonal left direction C
This is the sum of the density levels of three pixels in a row.

Next, the absolute value of the difference between the sum of density levels of pixels including the pixel of interest Q7 and the sum of density levels of three pixels not including the pixel of interest is calculated for each direction, and the following determination is made. This is T (
(is the threshold level.

(i) I T2-T11≧'rH or 1T2T3
If 1≧TII, a = 1 (ii) ITs T41≧TH or 1T5-T
If 61≧T II, a = 1 Tsukuda) 1T8-T71≧TH or I Ts -T9
If 1≧Tll, a = 1 4v) 1T1t Ttol≧T'H or 1Tu
-T121≧TH, then a = 1 (φ) If none of the above conditions (1) to 2 hold true, a = Q If the result of this determination is 0, the target pixel Q7 will not undergo MTF correction, Output at the same level.

If a = 1, target pixel Q71: MTF correction is applied and output.

Incidentally, the MTF correction may be performed using a digital filter as shown in FIG. 1(a) as in the conventional case.

As an example, set the pixel density to level O (white) to level 7 (black).
Let us consider the correction processing of the image shown in FIG. 2 which has been quantized with 1=8 values. The average level difference/pixel of the image is set to 3, that is, the above judgment threshold T H is set to 9, and the first MTF correction filter is
Let us use the one shown in Figure (a). Then, the corrected image of the image in FIG. 2 becomes as shown in FIG. 5.

Comparing FIG. 2 and FIG. 5, it can be seen that no MTF correction is performed in portions where the density changes are gradual, and undesirable phenomena such as white spots do not occur. On the other hand 1. It can be seen that the MTF correction is performed on the parts where the density changes sharply, and the parts become clearer.

In addition, in the above example, to determine whether the MTF correction of the pixel of interest is appropriate,
Although a total of 12 sets of three pixels in each of the horizontal, vertical, and diagonal directions were considered, it is also possible to reduce or even increase the number of sets. In other words, the number of pixels surrounding the pixel of interest to be referenced is not necessarily limited to the above example. However, the formula for determining whether MTF correction is appropriate is not limited to the above example, and another formula may be used as long as it correctly reflects density changes. Furthermore, the MTF correction filter is not limited to the one mentioned above.

Next, an example of an apparatus for performing image correction according to the present invention will be described with reference to FIG.

An input image signal obtained by sampling an original image using a known scanner (not shown) is subjected to, for example, eight-level quantization by a quantizer 100, and is temporarily stored in an input sofa 101. The MTF corrector 105 calculates the current pixel Q7 of the 8-level quantized image and the surrounding pixels Q3 . Q6
．． Qst Qll (MTF by the filter in Figure 1)
(1) is taken in from the input buffer 101, the MTF correction level of the pixel of interest is determined, and Q7 is output.

The adder 102 selects the pixel of interest Q7 and the surrounding pixels Q1 to Q7.
Q6, Qs to Q13 are taken in from the input buffer 101, and the above-mentioned value of T1-T13 is determined. This T1-T
Using the value No. 13, the comparison/judgment unit 103 makes a determination regarding the density change described above, and the determination result (a) is sent to the output level selector 104.

When the judgment result of the comparison judgment circuit 103 is valid (a=1), the output level selector 104 selects the MTF correction level sent from the MTF corrector 105 and outputs the output cover sofa 10 from Q7'.
If not, select the pixel level of interest Q7 sent from the input cover sofa 101 and send it to the output buffer 10.
Send to 6. Thus, via the output buffer 106,
A signal of a corrected image of the 8-value quantized image is output.

The image correction processing method of the present invention is as described above, and the correction processing can be applied regardless of the type of image, such as a character image, a photographic image, or an image that is a mixture of both. The effect of the present invention is remarkable because it is possible to minimize the influence of density variations and quantization noise due to variations in sensitivity, and to obtain a good corrected image.

[Brief explanation of the drawing]

Figure 1 is a diagram explaining an example of an MTF correction filter, Figure 2 is a diagram explaining an example of an MTF correction filter.
The figure shows an example of an 8-value quantized image, the third figure shows an image obtained by simply performing MTF correction on the image in figure 2, and figure 4 explains a specific example of image correction according to the present invention. diagram for,
FIG. 5 is a diagram showing an image obtained by correcting the image in FIG. 2 by the method of the present invention, and FIG. 6 is a block diagram showing an example of an apparatus for executing the image correction process of the present invention. 100... Quantizer, 101... Input buffer, 102... Adder, 103... Comparison judge,
104...Output level selector, 105...M
TF corrector, 106...output buffer. Figure 1 Figure 3

Claims (1)

[Claims] 1. Referring to the pixel of interest (pixel of interest) in the multi-level quantized image and its surrounding pixels, examine density changes in a limited small area including the pixel of interest, and detect density changes that are more rapid than a certain degree. Only when the pixel of interest is detected is M T
F (Modulation'l'transfer
Function) An image correction processing method characterized by performing correction.