JPS6195669A - Picture processing method - Google Patents

Abstract

PURPOSE:To execute an arbitrary shading processing by processing a picture based on a profile signal obtained from a sharp signal and an un-sharp signal. CONSTITUTION:The sharp signal S and the un-sharp signal U are inputted to a subtractor 1. The obtained profile signal P is inputted to comparators 4 and 6, which compare positive and negative slice levels L<+> and L<->. The obtained positive and negative profile signals Pb<+> and Pb<-> are added by an adder 10 to go to the profile signal Pb, which is inputted to a subtractor 11, and the profile signal Pb can be obtained with calculation such as P-Pb. A multiplying table 12 has a shading coefficient (m) and outputs an adjusted shading signal m.Pa.

Description

DETAILED DESCRIPTION OF THE INVENTION [Industrial Application Field] The present invention relates to an image signal processing method in image scanning and recording processing, and particularly to a method for performing image signal blurring and emphasis processing on contour portions.

"Conventional Example" and "Problems to be Solved by the Invention" When recording an original image using an image scanning recording device, it is necessary to make the image of a specific object included in the original image clearly appear as if it stands out from the background. In some cases, contour enhancement processing as shown in Japanese Patent Publication No. 49-40088, Japanese Patent Application Laid-Open No. 56-8140, and Japanese Patent Application No. 58-14621 filed by the present applicant is performed. In other words lol! Contour enhancement processing is performed to obtain a signal Dt- in which the contours of a specific image are emphasized by performing the calculation Dds+k(S-ty) (1) from the sharp signal S and the unsharp signal U as shown in Figure J1. It is being done.

Here, the sharp signal S is the image signal of the pixel whose outline should be emphasized, and the unsharp signal U may or may not include the image signal of the sharp signal S of multiple pixels around the sharp signal 8.) is the average value of the image signals.

On the other hand, in image scanning recording processing, there may be cases where it is not necessary to emphasize the contours included in the original image, or it may be better to use K to blur them. To obtain this blurred image, the sharp signal S or unsharp signal U may be used as is, or it may be better to further blur the signal.

Furthermore, when an original image is recorded at high magnification, noise caused by particulate pigments in the original image tends to appear in the recorded image, and there is a need to remove this type of noise.

The present invention has been proposed in view of the above-mentioned conventional circumstances, and its purpose is to provide a contour processing method that can perform blurring processing to a degree of inheritance on the first K contour. Another object of the present invention is to provide an image processing method that can perform not only the blurring process but also the edge enhancement process. Furthermore, it is an object of the present invention to provide an image processing method capable of reducing noise appearing in a recorded image due to particulate pigment in the original image.

"Means for Solving the Problem" In order to achieve the above object, the present invention takes the following means. That is, first, the unsharp signal U is subtracted from the sharp signal S to obtain the contour signal P, and then the contour is blurred from the sharp signal S and the contour signal P according to the following equation (3). A signal R2 is obtained from the image.

R2-8-kP (3) (k is coefficient) Here, the degree of blurring differs depending on the coefficient ko1.

Furthermore, by taking the following measures, it is possible to perform the contour blurring process and the contour enhancement process at the same time. That is, the contour signal Pa within a band separated from the contour signal P by the positive slice level A/L+ and the negative slice level NL-.
and the contour signal Pb outside this band, and from the above two signals Pα, Pb and the sharp signal, R1-8-m-Pa+n-Pb (4) (m, n
is a coefficient) The signal R1 obtained by the calculation is taken as an image signal.

In the above process, by adjusting the positive and negative slice levels, it is possible to adjust the blur level 9 emphasis level,
Further, by adjusting the coefficients m and n, the degree of blurring and the degree of emphasis can be adjusted.

Furthermore, in this case, this noise can be reduced by setting the slice level to a level higher than the level of the contour signal PN obtained from noise generated due to particulate pigment in the original image at high magnification.

"Embodiment" and "Operation" FIG. 2 shows an embodiment of the outline blurring process which is the basis of this invention. First, by subtracting the unsharp signal U from the sharpened signal 8, a contour signal P can be obtained.

Pnaw 8-U ・・・・・・(2) (ascension) Next, multiply the Jyap signal 8 by a coefficient of an arbitrary value to the Joki subtraction signal P [subtracting k-P] Therefore, it is possible to obtain an image signal 82 whose outline is blurred. The degree of blurring can be adjusted by the magnitude K of the coefficient.

FIG. 1 provides a detailed explanation of the invention. As shown in FIG. 2 above, first, by subtracting the unsharp signal U from the sharp signal 8, the contour signal P is obtained.

P-8-U (2) (Elevation) Next, a positive slice level A/L+ and a negative slice level L- are provided for the contour signal PK to increase the band between the above two levels. The contour signal within the band is extracted as Pa, and the contour signal outside the band is extracted as Pb, and both are multiplied by the blurring coefficient m1 and the emphasis coefficient nt, and calculations are performed based on the following (4) 2.

R1-8-m-Pa+n-Pb = (4) (ascension) where two slice levels /I/L+, L- and contour signal P
The relationship between the positive and negative peaks 1lIP+ and P- is L+≧P+
By setting L-≦P- (slice level A/L+, making the absolute value of L- sufficiently large), the emphasis signal Pb becomes zero, and the signal is exactly the same as shown in FIG. 2 above, that is, the outline is blurred. An image signal (see formula 8) can be obtained.

R'2mS-m-Pa-8-m-PxS-k ・P
``''(5) Conversely, by setting both of the two slice reps/I/L+lL- to zero, the blurring signal Pa becomes zero, and only contour enhancement can be performed (see equation 1).
.

R3-8+n-Pb-8+n-P-8+n (S-U)=S+k (S-U)
−・・・−(6) Swiss level L+, L- is 0<L+
<P+.

When 0>L->P-, the degree of blurring and emphasis can be obtained according to the size (Equation (4)) Furthermore, the degree of edge blurring and edge emphasis is determined by the blurring coefficient m and the emphasis coefficient n. It can also be changed depending on the size.Especially m
If n is set to -0 and n is changed appropriately, only the outline is emphasized by K, and by set n=0 and m is changed appropriately, only the outline is blurred.

Furthermore, the positive value /l/L+ must be made larger than the noise N shown in the upper part of Figure 1, which occurs due to granular pigments in the original image at high magnification, and the negative value. By making the slice level L- larger in absolute value than the level of the contour signal PN (that is, by applying only blurring processing to the granular pigment)
The noise N can be suppressed to a small level like the signal N' shown in FIG.

FIG. 8 is a circle showing an example of an apparatus for implementing the present invention, and FIG. 4 is a diagram showing its main signal waveforms. The sharp signal S and the unsharp signal U are input to a subtracter 1,
In the subtracter 1, the calculation according to equation (2) is performed, and the resulting contour signal P is input to the comparator 4. A positive nufide level L+ set by a control bag Wt (not shown) K such as a further gOPU is input to the comparator 4 from the foot 2, and ? , the slice level L+ is injected at timing t from the OPU, and the positive 7-fiber level A/L+
The above contour signal P is compared, and when P≧L+, K
sets the selection signal 8EL1, which is the output of the comparator 4, to rE[J level, and when P<L+Z, K sets the selection signal SE,
Set L1 to I”LJVbeA/.

On the other hand, the contour signal P and the positive slice level signal L+ are
The signal P'b+ is also input to the subtracter 5, where the calculation PL+ is performed, and the resulting signal P'b+ is input to the selector 8 mentioned above.

A zero level signal is also input to the selection 118.
The selector 8 receives the selection signal 5 input from the comparator 4.
When ELI is at the rHJ level (that is, when P≧L+), the above signal P'b+ and the selection signal 5ELI
When is at rLJ level (i.e. when P(L+)) KF
i A zero level signal is output. These two (output) signals or the signals combined in series are the positive contour signal Pb+
becomes.

The contour signal P is also input to the comparator 6, and the negative level /L/L- set by 0PU (not shown) is input to the comparator 6 from the foot 8, and the timing t from OPU is input to the comparator 6. The slice level /L/L- is programmed in. Then, the comparator 6 compares the contour signal P with the negative space ratio NL- and 2>(When P≦L-, the selection signal 81!L2 which is the output of the comparator 6 is outputted.
level, and when it is P)L-, the selection signal f9
Set EL2 to "LJ level /L/.

On the other hand, the contour signal P and the negative slice level signal L'''' are also input to the subtracter 7, and the subtracter 7 outputs the signal P−L−
The calculation is performed, and the resulting signal P'b-
is input to the selector 9. This selector 9 also receives a zero level signal and outputs a 1 signal P'b- when the selection signal BEL2 inputted from the comparator 6 which is 1 liter old is at the rHJ level, and also outputs a 1 signal P'b-. 5EL2 is rL
When the level is J, a zero level is output, and these two
The two output signals (P'b-, 0) are combined in series to obtain the negative contour signal Pb-.

Positive and negative contour signals Pb+Pb- obtained as above
are input to an adder 10, where they are added together to obtain a contour signal Pb.

The contour signal Pb obtained by K in this way is input to the subtracter 11, and the contour signal P is also input to the subtracter 11.
A calculation Pb is performed to obtain a contour signal Pa, which is input to the multiplication tape 1v12. Multiplication tape/
In I/12, the blurring coefficient m having the characteristics as shown in FIG.

Furthermore, the contour signal Pb, which is the output of the adder circuit 1O, is also input to the multiplication tape #18, and an emphasis coefficient n is applied to the contour signal Pb by the multiplication table 18 having the characteristics shown in FIG. 7(0). The multiplication is adjusted to match the nine emphasis belief fn-Pbt.
obtain.

A sharp signal S is also input to the subtracter 14.
The above adjusted blur signal m-Pa from the sharp signal S
The resulting value is input to the adder 15.

The adjusted emphasis signal n-Pb is input to the adder 15, and the adjusted emphasis signal n·Pb and the value obtained by subtracting the blur signal m·Pa from the sharp signal S are added and output. An image signal R1 (Equation (4)) is obtained.

The delay circuit shown in FIG. 8 is for adjusting the two signals input to the 4 to 15 arithmetic circuits so that they have the same timing.

As described in the explanation of FIG.
By adjusting 'L+ and L-, the degree of blurring and the degree of emphasis can be adjusted. In the circuit shown in FIG. 8, the slice levels A/L+ and L- are 0PUK as described above.
It is therefore loaded onto foot 2.8. Therefore, the above L+,
The calculation for determining the value of L- can be adjusted to correspond to the magnitude of the sharp signal), or the magnitude of the sharp signal level A/L+, L- can be changed to correspond to the desired effect.

Slice rep jvL+, L- need not be kept constant;
Rather, it is preferable to use positive and negative levels depending on the magnitude of the sharp signal. This requirement can be realized by using a look-up tape Jv16 as shown in FIG. 5 in place of the latch 2.8.

That is, latches 2 and 8 are connected to L+ as shown in FIG.
-f+(S), L--f-■) is replaced with the look-up table memory 16 that stores the relationships, ! When a sharp signal S with a high level and brightness is input to the lookup table memory 16, K is input with sharp signal S, which has a large absolute value, and a sharp signal S with a low level. Sometimes, the levels NL0 and L-, which have small absolute values, are output (the broken line portion of the P signal in FIG. 6). Do this!
Level slice @L+ according to the density level of the image signal.

L- is obtained, and image processing of blur emphasis according to the image level is realized. An example of the output waveform is shown in FIG.

It goes without saying that the unsharp signal U may be used as the signal input to the lookup table memory 16.

Furthermore, if the degree of blurring and the degree of emphasis of the outline is changed depending on the brightness of the image, it is possible to reduce the degree of blurring and emphasis of the outline of bright image parts and increase the degree of dark parts, for example.

To realize this request, the multiplication tape /I/
12 and 18 may be replaced with the circuits shown in FIG. 7(a), respectively. That is, in FIG. 7, the lookup table memory 12-1.18-1, FIG. 7(d)
, coefficients α and β that change according to the unsharp signal U are output as shown in (e). In the example of FIG. 7, the larger (brighter) the level μ of the unsharp signal U becomes, the smaller the values of α and β become. These coefficients α and β are multiplier 1
According to 2-2.18-2, each multiplication tape/I/1
The output m −Pa of 2.18 is multiplied by n−Pa.

As a result, both the edge blurring signal component α·m-Pa and the edge enhancement signal component β·n-Pbd become signals that depend on the unsharp signal level. In the data example of lookup tapes /I/12-1 and ll-1 in FIG. 7, the larger the unsharp signal U, the smaller the edge blurring component and edge enhancement component. In summary, the brighter the image, the weaker the blurring effect (for contour signal components lower than the currently set slice level) and the weaker the enhancement effect (for contour signal components higher than the slice level p), and the darker the image, the stronger they are. Work like. Figure 8 shows the contour enhancement component β・n.
- A diagram illustrating how Pb changes depending on the unsharp signal U is shown. For simplicity, in FIG. 8, the level sphere values and L- are set to zero. In the example of Naka, the contour emphasis component of the bright part is suppressed, so n
The positive contour signals, which had a difference in the -Pb signal, are almost at the same level in the R1 signal. In this manner, by using the circuit shown in FIG. 7, it is possible to realize contour enhancement characteristics that correspond to the image signal level.

The reason why the unsharp signal U is used as the image signal here is to give the above-mentioned effect also to the minor side contour portion of the contour signal P. It is also possible to use the sharp signal S as the image signal, but in that case it has almost no effect on the minus side contour signal portion.

The above explanation has been given by taking the edge enhancement component β・n・Pb as an example, but the exact same effect can be achieved by using the edge blurring component α・m−PaK.
Needless to say, you can also get it. Furthermore, both can be made to coexist based on equation (4). However, in that case, m-Pa is changed to α・m-Pa, and n-Pb is changed to β・n
-Pb, respectively (7)2R1=S-a・
It is necessary to use m-Pa+β·n-Pb = (7).

Incidentally, as shown in FIGS. 8 and 7, the multiplication table 12 may be configured to multiply the positive contour signal Pa+ and the negative contour signal Pa- by separate blurring coefficients m+ and m-, respectively. Figure 7).

Similarly, the multiplication tape A/18 has separate blurring coefficients n'' for each of the positive contour signal Pb+ and the negative contour signal Pb-,
n- and multi may be a combination of two.

See FIG. 7(c) The look-up table memory 16) and multiplication table 1 of FIG.
By preparing multiple sets of 2.18 (or the circuit shown in Fig. 7 used in place of these) and switching them for each fixed color range of the original image, it is possible to, for example, change the image outline in a specific corner range. It is possible to obtain effects such as emphasizing the color of the image or blurring the outline of the image in another specific color range. With this function, for example, by extracting and blurring the skin of a person, it is possible to obtain a duplicate image with smoother skin. For example, Japanese Patent Publication No. 50-14845 is used to extract color components.

Also, depending on the area and hue of the original image, 7 rice levels L+, L
It goes without saying that the −1 coefficients m, n, α, and β may be set.

"Effects" As explained above, this invention makes it possible to obtain a recorded image with a blurred outline based on the outline signal obtained from the sharp signal and the unsharp signal. This is effective when you want to blur the image. Further, the contour signal is sliced at all positive and negative slice levels to obtain a contour signal within the level band and a contour signal outside the level band, and it is possible to perform contour blurring and emphasis processing on both signals. Furthermore, since the magnitudes of the blurring signal and the emphasis signal can be adjusted by multiplying and combining gA numbers of appropriate magnitudes, the above effect can be further enhanced. Furthermore, by setting the suffix level to be higher than the level of the contour signal generated by the particulate pigment in the original image, it is possible to reduce the noise caused by the particulate pigment in the original image.

[Brief explanation of the drawing]

Figure 1 is a diagram explaining the details of this invention, Figure 2 is a diagram explaining a basic embodiment of this invention, Figure 8 is a diagram showing an apparatus for carrying out this invention, and Figure 4 is its main FIG. 5 is a diagram showing a look-up table memory in place of the latch used in this invention. FIG. 6 is a diagram showing an embodiment in which the null level is changed. FIG. 7 is a diagram showing blurring and edge enhancement. FIG. 8 is a diagram for explaining the embodiment of FIG. 7. In the figure, 1...subtractor, 2.8...latch circuit,
4...Comparator, 5...Subtractor, 6...
Comparator, 7... Subtractor, 8.9... Selector, 10... Adder, 11... Subtractor, 12.18... Lookup table memory, 14.
...Subtractor, 15...Adder, 16...Lookup table memory, S...Sharp signal,
U...unsharp it, P...contour signal,
Pa...contour signal, Pb...contour signal, k
, m, n... coefficient, 几1. R2...processed image signal. Fig. 2-s-kp Fig. 4 5EL1 Jitte SEL 2'+ Fig. 5 Fig. 7 (a) 1st

Claims (10)

[Claims] (1) Photoelectrically scan the image to generate sharp signal S and unsharp signal U, and from sharp signal S to unsharp signal U
Using the contour signal P obtained by subtracting R_2=S-kP
An image processing method characterized by blurring a contour part to a degree according to the magnitude of a coefficient k by calculation (where k is a coefficient). (2) Photoelectrically scan the image to obtain a sharp signal S and an unsharp signal U, subtract the unsharp signal U from the sharp signal S to obtain a contour signal P, and from this contour signal P a positive slice level L^+ Extract the contour signal Pa within the band of the negative slice level L^- and the contour signal Pb outside this band, and calculate the following to the contour part by calculating R_1=S-m・Pa+n・Pb (m and n are coefficients). An image processing method characterized by performing blurring and/or emphasis according to the magnitudes of coefficients m and n. (3) By making the absolute values of slice levels L^+ and L^- sufficiently large, the contour signal Pb is set to zero, and R_2
2. The image processing method according to claim 2, wherein the contour portion is blurred to a degree corresponding to the magnitude of the coefficient m by calculating =S-m·Pa. (4) To multiply the contour signal Pa by the coefficient m, use Pa as an input and use a lookup table memory with m×Pa as the output, and to multiply the contour signal Pb by the coefficient n, use Pb as the input and n The image processing method according to claim 2, characterized in that a look-up table memory whose output is ×Pb is used. (5) The image processing method according to claim 2, wherein the two slice levels L^+ and L^- are determined according to the magnitude of the sharp signal S. (6) The image processing method according to claim 2, wherein the slice levels L^+ and L^- are determined according to the magnitude of the unsharp signal U. (7) Contour blurring component m・Pa and contour enhancement component n・P
b is a coefficient α(
By multiplying S) and β(S), the contour blurring component α(S)・m according to the level of the sharp signal S is obtained.
- The image processing method according to claim 2, characterized in that Pa and edge enhancement components β(S)·n·Pb are obtained. (8) Contour blurring component m・Pa and contour enhancement component n・P
By multiplying b by coefficients α(U) and β(U) that change depending on the level of the unsharp signal U,
Contour blurring component α according to the level of unsharp signal U
(U)・m・Pa and contour enhancement component β(U)・n・Pb
An image processing method according to claim 2, characterized in that: (9) Each has a plurality of circuit parts that determine slice levels L^+ and L^-, and circuit parts that determine the size of outline blurring and emphasis components, and has a plurality of circuit parts that determine the slice levels L^+ and L^-, and has a plurality of circuit parts that determine the size of the outline blurring and emphasis components, and has a plurality of circuit parts that determine the slice levels L^+ and L^-. Claim 2, characterized in that the above circuit is switched and used for the remaining area.
Image processing method described in section. (10) Slice levels L^+, L^-, coefficients m, n,
The image processing method according to claim 7 or 6, characterized in that α and β are set based on an original image area, hue, etc.