JPH0851542A - Image processing method and device therefor - Google Patents

Abstract

PURPOSE:To provide the image processing method and device with a wide dynamic range and excellent sharpness, granuality and gradation. CONSTITUTION:When an image signal G received by an image input device 11 is converted into an image signal G in compliance with an image output device 19, a harmonic component H in response to the image signal G is superimposed onto the image signal G and the image signal G with the harmonic component H superimposed thereon is sent to an image output device 19 in the image processing method. Furthermore, the image signal processing unit converting the image signal G received by the image input device 11 into the image signal G in compliance with the image output device 19 is provided with a harmonic component generating means 10 generating the harmonic component H corresponding to the image signal G and a harmonic component superimposing means 34 superimposing the harmonic component H onto the image signal G.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image signal processing method and apparatus capable of reproducing a visually pleasing image on the basis of human visual psychology in a color image copying machine or the like.

[0002]

2. Description of the Related Art Color reproducibility, gradation reproducibility, sharpness, and graininess are important characteristics in considering the quality of output images of color image copying machines. There have been proposed many image processing techniques for reproduction.

By the way, when the color output image is used as a means for transmitting information, the final receiving system of the image becomes the human visual system. Therefore, it can be said that it is desirable that the image processing described above be performed in image quality design in consideration of the perceptual characteristics of the visual system based on the visual psychological phenomenon.

This is, for example, disclosed in JP-A-3-2341.
As disclosed in Japanese Patent Publication No. 78, etc., it means that an image with a sharp effect can be obtained by performing the color correction processing in the L ^* a ^* b ^* space, which is a uniform color appearance space. There is.

However, with respect to spatial frequency processing which greatly affects the sharpness, graininess and gradation reproducibility of an image, generally, the high frequency of a filter for sharpness enhancement processing having a second derivative (Laplacian) characteristic. The one that tries to achieve both graininess and sharpness by cutting the band is used. On the other hand, as an example of deriving the relationship between image noise and gradation from visual characteristics, Honjo et al. Reported the effect of masking gradation step due to noise, so-called pseudo contour. Summary of the Fifth International Conference on Improvements in Non-Impact Printing Technology in San Diego (Paper Summaries of
the 5th Intern. Congr. on Advances in Non-Inpact
Printing Technologies (SPSE), pp.196) ". This basic concept is widely applied in the form of a dither method in halftone generation.

In the audio world in which psychoacoustic phenomena are taken into consideration, the effect of the secondary harmonic (chord) component by, for example, a vacuum tube audio amplifier having a secondary amplification characteristic may have a good psychoacoustic effect. It has been pointed out that in recent years, audio amplifiers that use MOSFETs with secondary amplification characteristics instead of transistors with exponential amplification characteristics have been announced, while in the field of image processing that deals with visual psychological phenomena. In the present situation, the processing that focuses on the harmonic component is not performed.

[0007]

As described above, as for the processing of high frequency components in the conventional image processing system, the image signal processing method based on the visual psychological phenomenon has not been proposed so much. In particular, few proposals have been made regarding harmonic components.

When the sharpness enhancement processing described above is performed by a discrete Laplacian filter in order to improve the sharpness of an image, image noise is also amplified and graininess is deteriorated when the sharpness is increased. Conversely, when the spatial frequency characteristic of the filter is lowered in order to reduce image noise, a so-called "sleepy" image with deteriorated sharpness is obtained. Also,
Several proposals have been made to switch or mix a plurality of discrete Laplacian filter coefficients according to the edge degree, saturation, etc. of the image, but in this case as well, in a region with low edge degree, the network generated by the image signal and the output device is generated. It is necessary to lower the spatial frequency characteristics in order to prevent point interference, which leads to a reduction in the dynamic range, unevenness and gradation step (pseudo contour).
It also brings about the result of revealing defects.

The present invention has been made in order to solve the above technical problems, and pays attention to the visual psychological characteristics of the harmonic components, and by superimposing them, sharpness, gradation, and granularity can be obtained. The present invention provides an image processing method and apparatus having good performance and a wide dynamic range.

The present invention also provides an image processing method and apparatus capable of further improving the characteristics by superimposing random noise in addition to the harmonic component.

[0011]

As shown in FIG. 1 (a), an image processing method according to the present invention uses an image signal G input from an image input device 11 and an image signal adapted to an image output device 19. Upon conversion into G, a harmonic component H corresponding to the image signal G is superimposed on the image signal G, and then the image signal G on which the harmonic component H is superimposed is sent to the image output device 19. It is designed to be sent out.

An image processing apparatus embodying this image processing method converts an image signal G input from an image input apparatus 11 into an image signal G suitable for an image output apparatus 19, as shown in FIG. In an image processing device for conversion processing, a harmonic component generation means 10 for generating a harmonic component H corresponding to the image signal G, and a harmonic for superimposing the harmonic component H on the image signal G in the image signal G. And a component superposing means 34.

[0013]

The operation of the technical means described above is shown in FIG.
It will be described based on.

In the figure, the harmonic component generating means 10
Generates a harmonic component H in an amount corresponding to the level of the image signal G. Then, the harmonic component tree-forming means 34 superimposes the generated harmonic component H on the color image signal G.

As a result, the superposed harmonic components have an edge enhancing effect at the edge portion of the input image, and prevent the dynamic range from being reduced due to excessive edge enhancement called ringing, and further, the gradation step difference. By generating a small amplitude distortion in the area where the spatial density change where (pseudo contour) is likely to occur is small,
It has the effect of canceling out gradation steps. Even for low-frequency noise that deteriorates graininess, it is possible to visually and psychologically reduce the noise level by adding a high-frequency component.

[0016]

DESCRIPTION OF THE PREFERRED EMBODIMENTS The present invention will be described in detail below with reference to the embodiments shown in the accompanying drawings.

[First Embodiment] FIG. 2 shows a schematic configuration of a first embodiment of a color image copying apparatus to which the present invention is applied.
In the figure, the color image copying apparatus is an image input apparatus 1
1, a gradation conversion circuit 12 for correcting the gradation of the input image signal, and the image signal as a lightness chromaticity signal L ^* , a ^* , b ^*
, A brightness / chromaticity conversion circuit 13 for separating the image signal, a sharpness correction circuit 14 for correcting the sharpness, a discrete Laplacian filter 32, and a lookup table 31 for non-linearly converting the image signal G.
And 33, a harmonic component superimposing circuit 34 that superimposes the obtained harmonic component on the image signal, a subtractive color mixture conversion circuit 15 that converts the image signal into subtractive color mixture information, a UCR circuit 16 that generates appropriate black information, and an image output. An image signal G is provided with a tone conversion circuit 17 for performing tone correction of the device, a binarization circuit 18 for binarizing a halftone image according to the image output device, and an image output device 19 for outputting an image. The corresponding harmonic component is superimposed on the image signal itself.

In this embodiment, the image input device 11 is a device for reading a color image and converting it into digital data, and specifically represents a color television camera, a color image scanner or the like. In this embodiment, each of the R, G and B components of the image signal is represented by 8 bits. Further, the gradation conversion circuit 12 is a look-up table for converting each component of R, G, B of the image signal into an equivalent brightness. Further, the lightness / chromaticity conversion circuit 13 converts the equivalent lightness-converted R, G, and B components into lightness L ^* and chromaticity a ^* , b ^* . Further, the sharpness correction circuit 14 performs a sharpness enhancement process using a discrete Laplacian filter.

3A shows the characteristics of the discrete Laplacian filter 32, FIG. 3B shows the look-up table 31 of the harmonic generation circuit according to the present invention, and FIG. 3C shows the look-up table 33. The contents are shown respectively.

The discrete Laplacian filter 32 has a size of 5.
It has a function of removing DC component at × 5, and each coefficient is -0.01133 -0.04848 -0.06587 -0.04848 -0.01133 -0.04848 -0.05350 -0.00918 -0.05350 -0.04848 -0.06587 -0.00910 -0.94732 -0.00910 -0.06587 -0.04848 -0.05350-. 0.00918 -0.05350 -0.04848 -0.01133 -0.04848 -0.06587 -0.04848 -0.01133 ・ ・ ・ (1). Here, the discrete Laplacian filter 32 converts the 8-bit input from 0 to 255 to 8-bit from 127 to 127 in order to prevent bit loss.

Furthermore, the look-up table 31 has a standardized third-order nonlinear output for an 8-bit input from -128 to 127. That is, the lookup table 31 is expressed by fl (x) = x ³ ÷ 128 ² (2) The look-up table 33 represents an inverse conversion of the look-up table 31.

Further, the harmonic component superimposing circuit 34, for the image signal G output from the sharpness correction circuit 14, outputs a weight α to the output of the look-up table 33 according to the image signal.
It is an adder circuit that adds the value multiplied by.

The effect in this circuit is described as follows.

The output from the discrete Laplacian filter 32 has no DC component. Look-up table 31 is 3
Since it is a second non-linear type, its output generates a third harmonic as shown in the following expression (3).

Cos (x) ³ = 3/4 cos (x) +1/4 cos (3x) (3) The second term of the equation (3) represents the generation of the third harmonic. The DC component is removed by the discrete Laplacian filter 32, and only the frequency component remains. As a result, only the frequency component including the harmonic component of the image signal is extracted.

Note that the harmonic component of various forms can be obtained in various forms such as when the order of the discrete Laplacian filter 32 and the lookup table 31 or 33 is exchanged, or when the lookup table 31 is inserted in the preceding stage of the discrete Laplacian filter 32. Extraction is possible. Further, as the higher harmonic wave in the present embodiment, the third higher harmonic wave is used because it is most effective in increasing the sharpness, but it can be easily generated for the higher order harmonic wave. Needless to say.

[Second Embodiment] The second embodiment will be described in detail below. FIG. 4 shows a schematic configuration of the second embodiment.

In the present embodiment, the same components as those in the first embodiment shown in FIG. 2 are designated by the same reference numerals and the description thereof will be omitted. In the present embodiment, the harmonic component generation circuit comprises a frequency component extraction circuit 41 and a harmonic component calculation circuit 42. The frequency component extraction circuit 41 is a circuit that executes discrete Fourier expansion in the vicinity of the pixel of interest, and the harmonic component calculation circuit 42 executes expansion in the spatial frequency dimension and discrete Laplacian filtering in the vicinity thereof to respond to the expansion magnification. It produces harmonics of different orders. FIG. 5A shows the harmonic component calculation circuit 42.
5B shows the spatial frequency distribution of the input image and FIG. 5B shows the spatial frequency distribution of the output image.

[Third Embodiment] The third embodiment will be described in detail below. FIG. 6 is a schematic diagram for explaining the principle of the third embodiment.

In the third embodiment, as shown in FIG. 6A, when the image signal G input by the image input device 1 is converted into the image signal G suitable for the image output device 19, The image signal G input by the image input device 1
Of the harmonic component H and the random noise N are switched or mixed according to the level of the image signal G and superimposed, and the image signal G on which the harmonic component H and the random noise N is superimposed is sent to the image output device 9. It is a thing.

FIG. 6B shows an image processing apparatus for implementing the above processing. This image processing apparatus is an image signal processing apparatus for converting an image signal G input by the image input apparatus 1 into an image signal G suitable for the image output apparatus 9, and the image signal G Harmonic component generating means 10 and random noise generating means 13 for generating a harmonic component H according to the above, harmonic component superimposing means 34 for superimposing the harmonic component H on the image signal G, and random noise superimposing means 2.
It is characterized by having 0.

The operation of the image processing apparatus described above is shown in FIG.
It will be explained based on.

In the figure, the harmonic component generating means 10
Is a harmonic component generation H of an amount corresponding to the image signal G level.
And the random noise generator 13 generates the random noise N in an amount corresponding to the level of the image signal G. Then, the harmonic component superimposing means 34 and the random noise superimposing means 20.
Superimposes the generated harmonic component H and random noise N on the color image signal G.

As a result, the superimposed harmonic components have an edge enhancing effect at the edge portion of the input image, and prevent the dynamic range from being reduced due to excessive edge enhancement called ringing. Further, the random number noise has an effect of making the gradation step inconspicuous by generating a noise having a minute amplitude in a region where the spatial density change is less likely to cause a gradation step (pseudo contour). The amount of random noise to be superimposed is, for example, at most about half of the gradation step.

FIG. 7 is a block diagram showing a concrete configuration example of the third embodiment.

In the figure, the image signal input from the image input device 1 has its sharpness detected by the sharpness detecting and switching means 2, and is divided into an image signal having high sharpness and an image signal having low sharpness. The sharpness correction means 3 corrects the sharpness of the image signal in the area determined to have a high sharpness, and the harmonic distortion generation means 4 superimposes the harmonic distortion according to the sharpness on the gradation level. It is sent to the halftone generating means 8 which outputs the halftone signal corresponding to the image output device 9 through the correcting means 7. On the other hand, the image signal in the area determined to have a low sharpness by the sharpness detection / switching means 2 is superposed with appropriate noise by the random number noise generating means 5 and the synthesizing means 6, and then the gradation correcting means 7, intermediate Key generation means 8
Sent to.

FIG. 8 is a block diagram showing a configuration example of the random noise generator 5. The image signal input to the random noise generating means 5 is an image signal of an area determined to have a low sharpness by the sharpness detecting and switching means 2. This image signal passes through the gradation conversion table 51, and the random number generating circuit 53 generates random noise according to the weighting by the filter circuit 52. The random frequency noise generated by the random number generation circuit 53 is corrected in spatial frequency characteristic by the spatial frequency characteristic correction circuit 54 and sent to the synthesizing means 6.

FIG. 9 illustrates the characteristics of the gradation conversion table 51. It is now assumed that the image output device 9 has the gradation reproduction characteristic y = f ₁ (x) as shown in FIG. The horizontal axis is the number of input stages, and various numbers of stages are selected according to the requirements of the equipment and system. In the example of the figure, it is set to 8 bits (256 stages). The vertical axis is the measured value of the output image, and units such as density (D), lightness (L ^* ), chromaticity (a ^* , b ^*, etc.) and saturation are used according to the purpose. FIG. 9 (B)
Is a target gradation reproduction characteristic of the image output device 9 y = f ₂
(X) is shown. FIG. 9C shows the characteristic y = f ₃ (x) of the gradation correcting means 7 at this time, and f ₂ (x) and f ₁ (f ₁ )
As a composite function of the inverse function f ₁ ^-1 1 (x) of (x), it is expressed as f ₃ (x) = f ₂ (fl ^-1 1 (x)) (4). In a system having a sufficiently large number of steps, the synthesized and reproduced tone reproduction characteristics match those in FIG. 9B,
In a system having a limited number of gradation steps, bit loss occurs as shown in FIG. Gradation conversion table 51
Provides a look-up table for correcting such an effective gradation characteristic.

The filter circuit 52 is a Laplacian (second derivative) filter and has a spatial frequency characteristic as shown in FIG. This is a characteristic used as an edge detection filter in image processing. Gradation conversion table 51
The signal input to the device is already sharpness detection and switching means 2
Therefore, the sharpness is determined to be low, and in the present embodiment, the sharp contour is used as an edge to catch a pseudo contour due to deterioration of the effective gradation characteristic. After the random number noise is generated by the random number generation circuit 53 according to the edge degree, that is, the pseudo contour degree by the filter circuit 52 and the granularity deterioration in the low spatial frequency region is prevented by the spatial frequency characteristic correction circuit 54, the output is the noise amount. , To the synthesizing means 6.

On the other hand, for a region judged to have a high sharpness by the sharpness detection / switching means 2, the harmonic distortion generation means 4 applies the harmonic distortion through the sharpness correction means 3 by the spatial frequency correction circuit. It is added and sent to the synthesizing means 6.

As the harmonic distortion generating means 4, a circuit similar to the harmonic component generating circuit 10 used in the first embodiment shown in FIG. 2 is used.

[0042]

As described above, according to the present invention, by superimposing a harmonic on the spatial frequency component of an image based on the human perception of noise, both improvement of image sharpness and reduction of false contour are achieved. It is possible to obtain a more visually preferable image.

Further, by generating random noise with a minute amplitude in a region in which the gradation difference (pseudo contour) is apt to occur and the spatial density change is small, the gradation step becomes inconspicuous.

[Brief description of drawings]

FIG. 1 is an overall configuration diagram showing a concept of the present invention.

FIG. 2 is a schematic overall configuration diagram of a first embodiment of the present invention.

FIG. 3 is a conceptual diagram showing an example of a discrete Laplacian characteristic and a table characteristic according to the first embodiment of the present invention.

FIG. 4 is a schematic overall configuration diagram of a second embodiment of the present invention.

FIG. 5 is a conceptual diagram showing characteristics of a harmonic component calculation circuit according to a second embodiment of the present invention.

FIG. 6 is a schematic overall configuration diagram of a third embodiment of the present invention.

FIG. 7 is a block diagram illustrating a specific configuration example of a third embodiment.

FIG. 8 is a block diagram showing a configuration example of a random noise generator.

FIG. 9 is an explanatory diagram showing characteristics of a gradation conversion table.

FIG. 10 is a graph showing a spatial frequency characteristic of a Laplacian filter.

[Explanation of symbols]

1 ... Image input device, 2 ... Sharpness detection and switching means,
3 ... Sharpness correction means, 4 ... Harmonic distortion generation means, 5 ... Random noise generation means, 6 ... Synthesis means, 7 ... Gradation correction means, 8
... halftone generation means, 9 ... image output device, 10 ... harmonic component generation means, 11 ... image input device, 12 ... gradation conversion circuit, 13 ... lightness / chromaticity conversion means, 14 ... sharpness correction circuit,
15 ... Subtractive color mixture conversion circuit, 16 ... UCR circuit, 17 ... Gradation conversion circuit, 18 ... Binarization circuit, 19 ... Image output device,
20 ... Random noise superimposing means, 31, 33 ... Look-up table, 32 ... Discrete Laplacian filter, 34 ... Harmonic component superimposing means, 41 ... Frequency component extracting circuit, 42 ...
Harmonic component calculation circuit, 43, 45 ... lookup table, 44 ... discrete Laplacian filter, 51 ... gradation conversion table, 52 ... filter circuit, 53 ... random number generation circuit,
54 ... Spatial frequency characteristic correction circuit

─────────────────────────────────────────────────── ─── Continuation of the front page (51) Int.Cl. ⁶ Identification number Office reference number FI technical display location H04N 1/46 G06F 15/68 310 J H04N 1/46 Z

Claims (10)

[Claims] 1. An image signal G obtained by adapting an image signal G input by an image input device (11) to an image output device (19).
When performing the conversion processing into the above, the harmonic component (H) of the image signal (G) input by the image input device (11) is superimposed according to the level of the image signal (G), and the harmonic component H An image output device (19)
An image processing method characterized in that the image is transmitted to 2. An image processing apparatus for converting an image signal (G) input by an image input apparatus (11) into an image signal (G) suitable for an image output apparatus (19), wherein the image input apparatus ( 11) Harmonic component generation means (10) for generating a harmonic component (H) of the image signal (G) input in
And an harmonic component superposing means (34) for superposing according to the level of the image signal (G). 3. The image processing apparatus according to claim 2, wherein the image signal (G) is a color image signal. 4. The harmonic component (H) is generated by a combination of two look-up tables (31), (33) which have an inverse function relationship with the discrete Laplacian filter (32) in the harmonic component generating means (10). The image processing apparatus according to claim 2 or 3, which generates the. 5. The image processing apparatus according to claim 2, wherein the harmonic component generating means (10) generates the harmonic component (H) by a combination of the spatial frequency adjusting means and the signal level converting means. . 6. The image processing apparatus according to claim 2, further comprising a spatial frequency component extracting means (41) and a harmonic component calculating means (42). 7. The harmonic component generating means (10) and the harmonic component superimposing means (34) respectively include a harmonic component (H).
The image processing apparatus according to claim 2 or 3, which particularly generates and superimposes a third harmonic component. 8. The image input device (1) when converting the image signal (G) input by the image input device (1) into an image signal (G) suitable for the image output device (9).
The harmonic component (H) and random noise (N) of the image signal (G) input at are switched or mixed according to the level of the image signal (G) and superposed, and the harmonic component (H)
And an image signal (G) on which random noise (N) is superimposed is sent to an image output device (9). 9. An image processing device for converting an image signal (G) input by an image input device (1) into an image signal (G) suitable for an image output device (9), wherein the image input device ( 1) a harmonic component generation means (10) for generating a harmonic component (H) of the image signal (G) input;
Image signal (G) input by the image input device (1)
Random noise generating means (13) for generating random noise (N) from the superimposing means (34) for superimposing a harmonic component (H) or random noise (N) depending on the level of the image signal (G). 20) An image processing device comprising: 10. The image processing apparatus according to claim 9, wherein the random noise generating means (13) generates random noise (N) by a combination of a discrete Laplacian filter, a look-up table and a random number generating means.