JP4160887B2 - Image processing apparatus and image processing method - Google Patents

Description

  The present invention relates to an image processing apparatus and an image processing method for inputting a digital image signal and performing a predetermined image quality improvement process.

  Conventionally, an image processing method is known in which an image is decomposed into a plurality of frequency bands by wavelet transform, a signal is controlled for each band, only a certain band is emphasized by multiplication, and inverse transform is performed (for example, Patent Documents). 1). That is, in this image processing method, an input image signal is subjected to wavelet transform and decomposed into a plurality of frequency band signals, predetermined image processing is performed on at least one frequency band signal, and inverse conversion is performed. Note that the predetermined image conversion is a multiplication of a predetermined number.

  On the other hand, it is also possible to perform filter processing for each frequency band and for each direction in real space. For example, in real space filter processing, since there is a degree of freedom in setting the frequency band of the filter, it is advantageous to perform filter processing with a limited line memory. In particular, by performing spatial frequency correction independently for each direction and for each band, emphasis can be made only in a required direction, and an improvement in the sharpness of characters on a halftone dot can be expected.

  Furthermore, the present invention relates to a method of obtaining a character image on a halftone dot with good sharpness while preventing an increase in hardware scale by emphasizing the high frequency band independently for each direction and emphasizing the low frequency band with a single filter. Have been proposed.

JP-A-6-274614

  However, in the conventional image processing apparatus shown above, when wavelet transform is used, it is decomposed into components for each direction by decomposing into two-dimensional wavelet coefficients. However, in order to obtain signal components in the frequency band using the wavelet transform, a large amount of line memory is required, resulting in an increase in hardware scale. It was.

  In addition, when independent emphasis processing for each direction is performed on high-frequency components, especially in dense pattern parts (fine character parts that require high resolution) on white ground (on non-halftone dots), where defects tend to catch human eyes. However, there is a problem in that density unevenness and line discontinuity due to an emphasis difference in each direction are noticeable.

  The present invention has been made in view of the above, and does not increase the scale of hardware, and by combining filtering processing controlled for each direction and filtering processing by a single filter, sharp characters in halftone dots are sharpened. An object is to improve the image quality of a dense pattern portion on a white ground (on a non-halftone dot) at the same time as obtaining an excellent image.

In order to solve the above-described problems and achieve the object, the invention according to claim 1 is directed to an image processing apparatus that performs a desired spatial frequency correction on an input image signal, and a directional filter having directionality is provided. The first spatial frequency correction means for performing spatial frequency correction for each direction of the image signal, and a single filter having a frequency characteristic different from that of the directional filter. Calculated by the second spatial frequency correction means for collectively performing the spatial frequency correction on the image, the feature quantity calculation means for calculating the feature quantity of the image signal for each directionality, and the feature quantity calculation means. It was in accordance with the feature quantity of each directional, with corresponding, and a third spatial frequency correction means for controlling the output of the directional for the first spatial frequency correction means And wherein the door.

According to the first aspect of the present invention, for example, the first spatial frequency correction means performs low frequency emphasis independently in each direction, so that necessary directions for low frequencies indispensable for character emphasis are provided. It is possible to emphasize only the component, and the second spatial frequency correction means performs high-frequency emphasis at once in all directions by a single filter. In the dense pattern area on the white ground (on the non-halftone dots), the same intensity can be emphasized in all directions. In addition, by controlling the filter processing according to the feature amount of the image, it is possible to ensure graininess and sharpness. As a result, it is possible to suppress deterioration in image quality such as density unevenness due to an emphasis difference for each direction and line breakage at an edge intersection, which occurs when high-frequency independent control is performed in each direction.

  The invention according to claim 2 is characterized in that the first spatial frequency correction means and the second spatial frequency correction means are arranged so as to process the image signals in parallel.

  According to the second aspect of the present invention, the first spatial frequency correction means and the second spatial frequency correction means can perform parallel processing on input image signals.

  The invention according to claim 3 is characterized in that the filter frequency characteristic of the second spatial frequency correction means is a high frequency characteristic with respect to the filter frequency characteristic of the first spatial frequency correction means.

  According to the third aspect of the invention, the first spatial frequency correction means performs the spatial frequency correction independently for each direction on the low frequency side, and the second spatial frequency correction means performs the high frequency side in all directions. By performing uniform spatial frequency correction, the sharpness on the halftone characters is ensured, and the image quality of the dense pattern portion on the white characters (on the non-halftone dots) is also ensured.

The invention according to claim 4, characteristic quantity calculated by the feature calculating unit is characterized in that an edge of the image.

According to the fourth aspect of the present invention, the feature amount calculated by the feature amount calculation means is used as the edge amount of the image, so that the edge of the image is not dulled.

The invention according to claim 5 is characterized in that a second feature amount calculating means for calculating the feature amount of the image signal for each directionality and a maximum for calculating the maximum value of the feature amount acquired for each direction. And a fourth spatial frequency correction unit that controls an output of the second spatial frequency correction unit according to a maximum value of the feature amount for each direction.

According to the invention of claim 5 , the spatial frequency correction is performed on the image subjected to the spatial frequency correction by the second spatial frequency correction unit according to the maximum value acquired along the directionality of the directional filter. Thus, for example, when the high-frequency component of the image is corrected by the second spatial frequency correction unit, it is possible to ensure the feature amount of the image.

The invention according to claim 6 is characterized in that the feature amount is calculated using a halftone correction value indicating an edge amount of an image and a local halftone dot degree.

According to the sixth aspect of the present invention, by extracting the edge amount as the feature amount of the image, for example, it is possible to secure the image although complicated kanji and resolution are required.

The invention according to claim 7 is characterized in that the feature amount is a value obtained by subtracting the halftone dot correction value from an edge amount of the image.

According to the seventh aspect of the present invention, it is possible to suppress image deterioration by setting the halftone correction value to a value obtained by subtracting the halftone correction value from the edge amount of the image.

According to an eighth aspect of the present invention, there is provided a halftone dot correction amount calculating unit for calculating a halftone dot correction amount for each direction, and a maximum value of the halftone dot correction amount for each direction calculated by the halftone dot correction amount calculating unit. Halftone dot maximum value calculating means for calculating, and the fourth spatial frequency correcting means, when the halftone correction amount acquired for each direction is less than or equal to a predetermined value, the image feature amount for each direction The output of the second spatial frequency correcting means is controlled in accordance with the maximum value of.

According to the eighth aspect of the present invention, the high-frequency emphasis on the halftone image is suppressed, and the processing for ensuring the sharpness is performed on the other image areas, thereby over-enhancing the halftone image at high frequency. In this case, it becomes possible to avoid the deterioration of the graininess caused by the undulations remaining in the halftone dots.

According to a ninth aspect of the present invention, in the image processing method for executing a desired spatial frequency correction on an input image signal, a directional filter having directionality is used for each direction with respect to the image signal. And a second step of performing spatial frequency correction collectively in all directions of the image signal using a single filter having a frequency characteristic different from that of the directional filter. If, for each direction, and a third step of calculating a feature value of the image signal, in accordance with the feature quantity of each directional calculated in the third step, the corresponding of the first step And a fourth step of controlling the output of the directionality to be performed.

According to the ninth aspect of the present invention, for example, in the first step, low frequency emphasis is performed independently in each direction, so that necessary direction components for low frequencies indispensable for character emphasis are provided. In the second step, high-frequency emphasis is performed in a single filter all at once in a second step. In a dense pattern portion (on a non-halftone dot), the same intensity can be enhanced in all directions. In addition, by controlling the filter processing according to the feature amount of the image, it is possible to ensure graininess and sharpness. As a result, it is possible to suppress deterioration in image quality such as density unevenness due to an emphasis difference for each direction and line breakage at an edge intersection, which occurs when high-frequency independent control is performed in each direction.

The image processing apparatus according to the present invention (Claim 1) performs low-frequency emphasis independently in each direction by, for example, the first spatial frequency correction means, so that low-frequency indispensable for character emphasis is provided. It is possible to emphasize only the necessary direction component, and the second spatial frequency correction means enables high-frequency enhancement to be performed in all directions in a single filter, so that the high-frequency component is particularly important. in the dense pattern portion of Do affecting white ground (on the non-dot), both all directions can emphasis of the same strength, also by controlling the filtering in accordance with the feature amount of an image, the graininess and sharpness Securement becomes possible. As for the high frequency, it is possible to suppress deterioration in image quality such as density unevenness due to an emphasis difference for each direction and line breaks at edge intersections, which occur when independent control is performed in each direction. That is, while suppressing the increase in hardware scale, it is possible to achieve both of ensuring sharpness for characters on halftone dots and ensuring the resolution of dense pattern portions on white ground (on non-halftone dots).

  The image processing apparatus according to the present invention (Claim 2) can perform parallel processing on the input image signal by the first spatial frequency correcting means and the second spatial frequency correcting means. As a result, the processing speed can be improved.

  In the image processing apparatus according to the present invention (claim 3), the first spatial frequency correction unit performs spatial frequency correction independently for each direction on the low frequency side, and the second spatial frequency correction unit performs the high frequency side. Since uniform spatial frequency correction is performed in all directions, the sharpness of characters on halftone dots is ensured, and the image quality of dense pattern portions on white characters (on non-halftone dots) is also ensured.

Further, the image processing apparatus according to the present invention (claim 4 ) uses the feature amount calculated by the feature amount calculation means as the edge amount of the image, so that the edge of the image is not dulled and resolution deterioration is avoided. There is an effect that can be done.

The image processing apparatus according to the present invention (Claim 5 ) has the spatial frequency according to the maximum value acquired along the directionality of the directional filter for the image subjected to spatial frequency correction by the second spatial frequency correction means. By performing the correction, for example, when the high frequency component of the image is corrected by the second spatial frequency correction means, it becomes possible to secure the feature amount (edge) of the image, so that the resolution and sharpness are improved. There is an effect that an excellent image can be obtained.

The image processing apparatus according to the present invention (Claim 6), since extracts the edge amount as the feature amount of the image, for example, when it is complicated Chinese characters and resolution is required can be secured image There is an effect.

In addition, since the image processing apparatus according to the present invention (claim 7 ) has a value obtained by subtracting the halftone dot correction value from the edge amount of the image, the image deterioration can be suppressed.

Further, the image processing apparatus according to the present invention (claim 8 ) suppresses high-frequency emphasis on the halftone image and performs processing for ensuring sharpness in other image areas. When emphasized, the effect of avoiding deterioration of graininess due to the undulations remaining in the halftone dots can be avoided.

In the image processing method according to the present invention (claim 9 ), for example, in the first step, low frequency emphasis is performed independently in each direction, so that it is necessary for low frequency that is indispensable for character emphasis. It is possible to emphasize only directional components, and in the second step, high-frequency emphasis is performed at once in all directions by a single filter. In dense patterns on white ground (on non-halftone dots), the same intensity can be enhanced in all directions. In addition, by controlling the filter processing according to the feature amount of the image, it is possible to ensure graininess and sharpness. As a result, it is possible to suppress deterioration in image quality such as density unevenness due to an emphasis difference for each direction and line breakage at an edge intersection, which occurs when high-frequency independent control is performed in each direction. That is, while suppressing the increase in hardware scale, it is possible to achieve both of ensuring sharpness for characters on halftone dots and ensuring the resolution of dense pattern portions on white ground (on non-halftone dots).

  DESCRIPTION OF EXEMPLARY EMBODIMENTS Hereinafter, preferred embodiments of an image processing apparatus and an image processing method according to the invention will be described in detail with reference to the accompanying drawings. The present invention is not limited to this embodiment.

(First Embodiment 1)
FIG. 1 is a block diagram showing the configuration of an image processing apparatus and system according to an embodiment of the present invention. As shown in the figure, this system is roughly divided into an image input device 100 such as an image scanner, a personal computer, and a digital camera, an image processing device 110 described later, and an image output device that performs print output on recording paper such as a laser printer. 120. Various corrections by the image processing apparatus 110 are executed on a microcomputer system or a personal computer equipped with a CPU.

  The image processing device 110 includes a scaling processing unit 101 that scales an image input by the image input device 100 to a specified magnification, a filtering processing unit 102 that performs desired spatial frequency correction on the scaled image, and a spatial A γ conversion processing unit 103 that performs γ conversion so that the frequency-corrected image has a desired density characteristic, and a halftone processing unit 104 that performs halftone processing such as dither processing and error diffusion processing on the density characteristic converted image. It is equipped with.

  When the image input by the image input apparatus 100 is a color image, color correction means for realizing a function of converting an input image of R, G, B signals into output data of CMYK signals is also provided.

  Note that, here, the scaling process by the scaling process unit 101, the γ conversion process by the γ conversion process unit 103, and the halftone process by the halftone process unit 104 are well-known techniques, and thus description thereof is omitted. A filtering process performed by the filtering processing unit 102, which is a feature of the embodiment, will be described below with reference to FIGS.

  In the first embodiment, an example will be described in which filter processing is performed using a low frequency filter in four directions of vertical, horizontal, right diagonal, and left diagonal and one high frequency filter. For the feature amount calculation, different processing is performed for low frequency and high frequency.

  FIG. 2 is a block diagram showing the configuration of the filter processing unit according to the first embodiment of the present invention. As shown in the figure, the filter processing unit 102 includes a smoothing filter 201 (see FIG. 3) having a filter matrix for smoothing processing, and a low frequency for correcting the spatial frequency of the low frequency component in the vertical direction (main scanning direction). Filter (vertical) 202 (see FIG. 4-1), low-frequency filter (horizontal) 203 (see FIG. 4-2) for correcting the spatial frequency of the low-frequency component in the horizontal direction (sub-scanning direction), diagonally right ( Low-frequency filter (diagonal right) 204 (see FIG. 4-3) that corrects the spatial frequency of the low-frequency component in the upward direction (rightward direction), and spatial frequency correction of the low-frequency component in the diagonally left direction (upward direction to the left) A low frequency filter (left oblique) 205 (see FIG. 4-4) is provided.

  Further, as shown in the figure, the filter processing unit 102 includes a low-frequency feature value calculation (vertical) unit 206, a low-frequency feature value calculation (horizontal) unit 207, a low-frequency feature value calculation (right diagonal) unit 208, a low-frequency feature. An amount calculation (diagonal left) unit 209 is provided.

  Further, as shown in the figure, the filter processing unit 102 includes a high-frequency feature amount calculation (vertical) unit 210 that calculates image features at high frequencies in the vertical direction (main scanning direction), and image features at high frequencies in the horizontal direction (sub-scanning direction). A high-frequency feature amount calculation (horizontal) unit 211 that calculates a high frequency feature amount calculation (right diagonal direction) unit 212 that calculates an image feature at a high frequency in the diagonally right direction (upward direction to the right) A high-frequency feature amount calculation (diagonal to the left) 213 for calculating image features in FIG.

  Further, as shown in the figure, the filter processing unit 102 converts the enhancement coefficient calculation unit 214 that converts the enhancement coefficient by the low frequency feature amount calculation (vertical) unit 206 and the enhancement coefficient by the low frequency feature amount calculation (horizontal) unit 207. Enhancement coefficient calculation unit 215 for converting, enhancement coefficient calculation unit 216 for converting the enhancement coefficient by the low frequency feature quantity calculation (right diagonal) unit 208, and enhancement coefficient calculation for converting the enhancement coefficient by the low frequency feature quantity calculation (left diagonal) unit 209 217, a maximum value calculation unit 218 that calculates the maximum value of the high-frequency edge amount in each direction, an enhancement coefficient calculation unit 219 that converts the enhancement coefficient after the maximum value calculation, and all directions (main scanning, sub-scanning using a single filter) A high-frequency filter 220 that performs high-frequency filter processing in all directions (scanning, right-up direction, and left-up direction) is provided.

  In addition, the filter processing unit 102 includes a multiplication unit 221 that multiplies the output value of the low frequency filter (vertical) 202 and the enhancement coefficient calculation unit 214, a low frequency filter (horizontal) 203, and an enhancement coefficient calculation unit 215, as illustrated. Multiplier 222 for multiplying output values, multiplier 223 for multiplying output values of low frequency filter (diagonal right) 204 and enhancement coefficient calculation unit 216, output values of low frequency filter (diagonal left) 205 and enhancement coefficient calculation unit 217 Multiplication unit 224 that multiplies the output values of the enhancement coefficient calculation unit 219 and the high frequency filter 220, an addition unit 226 that adds the output values of the smoothing filter 201 and the addition unit 227, a multiplication unit 221 and an addition unit 228. The output values of the adder 227, the adder 228 for adding the output values of the adder 229 and the adder 229, and the output values of the adder 223 and the adder 230 are added. Adding unit 229, an addition section 230, for adding the output value of the multiplier 224 and the multiplying unit 225.

  The smoothed filter 201 having the filter matrix shown in FIG. 3 smoothes the image after the scaling process. Although the description is omitted because it is a well-known technique, smoothing of an image uses a filter matrix of an appropriate size, and calculates the density value of the target pixel by the average value of the density values of the target pixel and surrounding pixels. It is a process to replace.

  The smoothing filter 201 is a filter shown in FIG. 3, and smoothes image data. Further, the low frequency filter (vertical) 202 is the filter shown in FIG. 4A, the low frequency filter (horizontal) 203 is the filter shown in FIG. 4B, and the low frequency filter (right diagonal) 204 is shown in FIG. 3 and a low frequency filter (left diagonal) 205 is a filter shown in FIG. 4-4.

  Next, FIG. 5 shows a processing block diagram of the low-frequency feature value calculation (vertical) unit 206. As shown in FIG. 5, in the first embodiment, the low frequency feature amount calculation (vertical) unit 206 is the low frequency edge amount calculation unit 231 in the vertical direction itself. Here, the filter used for the low-frequency edge amount calculation (vertical) unit 231 is the filter shown in FIG. The feature amount calculated by the low frequency feature amount calculation (vertical) unit 206 in this way is converted into an enhancement coefficient by the enhancement coefficient calculation unit 214. This conversion is, for example, as shown by a solid line in the graph showing the relationship between the edge amount and the enhancement amount shown in FIG. By multiplying this enhancement coefficient by the output value of the low-frequency filter (vertical) 202, the low-frequency vertical enhancement amount is calculated.

  The amount of enhancement is calculated in the same way for horizontal, right diagonal, and left diagonal. The low frequency filter in each direction is shown in FIGS. 4-2 to 4-4, and the low frequency edge amount calculation filter in each direction is shown in FIGS. 6-2 to 6-4. It is. 6A is a vertical low frequency edge calculation filter, FIG. 6B is a horizontal low frequency edge calculation filter, FIG. 6C is a right diagonal low frequency edge calculation filter, and FIG. This is a low-frequency edge calculation filter in the left diagonal direction, and uses these to calculate the edge amount in each direction.

  On the other hand, the high frequency filter 220 is a filter shown in FIG. Unlike a low-frequency filter, it has a characteristic that emphasizes all directions in a lump.

  Next, the configuration of the high-frequency feature value calculation (vertical) unit 210 is shown in the block diagram of FIG. As shown in FIG. 9, in this embodiment, the high frequency feature amount calculation (vertical) unit 210 includes a high frequency edge amount calculation unit 232, a halftone correction amount 233, and a subtraction unit 234. Here, the vertical direction of the high-frequency edge amount calculation unit 210 is a filter illustrated in FIG. The halftone dot correction amount calculation unit 233 performs the following calculation with reference to the pixel shown in FIG.

Halftone correction amount Δ = | A−B | + | B−C | + | C−D | + | D−E | + | E−F | + | F−G |
・ ・ ・ ・ ・ ・ (Formula 1)
Here, A, B,... Are the sum of three adjacent pixels with the same symbol.

  Since the halftone dot correction amount Δ (vertical) calculated by the halftone dot correction amount calculation unit 233 in this way is an amount indicating the halftone dot quality of the image, the high frequency edge calculated by the high frequency edge amount calculation unit 232 is used. By subtracting from the amount, a high-frequency edge amount (vertical) in which a halftone dot-like portion is removed, that is, a high-frequency feature amount (vertical) is calculated.

  The amount of high frequency emphasis is calculated in the same way for horizontal, right diagonal, and left diagonal. The high-frequency edge amount calculation filters in each direction are those shown in FIGS. 10-2 to 10-4, and the pixels referred to in the calculation of the halftone correction amount in each direction are shown in FIGS. 11-2 to 11-. This is shown in FIG.

  Next, the maximum value calculation unit 218 calculates the maximum value of the high-frequency edge amount in each direction, and the enhancement coefficient calculation unit 219 converts it into an enhancement coefficient. The conversion at this time is the conversion indicated by the dotted line in FIG. By multiplying this enhancement coefficient by the output value of the high frequency filter 220, a high frequency enhancement amount is calculated.

  Finally, the output value of the smoothing filter 201, the low frequency enhancement amount in four directions, and the high frequency enhancement amount are added and output.

  That is, according to the first embodiment, as shown in the flowchart of FIG. 12, an image signal supplied from the outside is input (step S11), and each directional filter is used for each low frequency direction. The first spatial frequency correction is executed (step S12), and for the high frequency component, the second spatial frequency correction is performed in a lump direction using a single filter (step S13). Thereafter, these images are emphasized, combined and output.

  Therefore, according to the first embodiment, since the low frequency is emphasized independently in each direction, only the necessary direction component can be emphasized for the low frequency indispensable for character emphasis. By emphasizing only necessary direction components, for example, in a character image on halftone dots, it is possible to emphasize only in the edge direction of characters without unnecessarily enhancing halftone dots adjacent to the characters, and an output result with high sharpness can be obtained. I can expect.

  In addition, high-frequency emphasis is performed in a single filter in all directions at once, so a dense pattern part on a white ground (on non-halftone dots), where high-frequency components are particularly important (for example, complex kanji and resolution are required) The edge can be enhanced with the same intensity in all directions. As a result, even when the high frequency is controlled independently in each direction, it is possible to suppress the phenomenon of density unevenness due to the emphasis difference for each direction and the break of the line at the edge intersection.

  In this embodiment, for the purpose of suppressing image quality deterioration due to halftone dots also reacting as edges in high-frequency feature quantity calculation, a halftone dot correction amount indicating halftone dotness is calculated and subtracted from the edge amount. However, if there is no particular problem with image quality deterioration due to halftone dots, such as the edge amount calculation coefficient to be used, the halftone correction amount need not necessarily be calculated. In this example, a simple halftone dot correction amount is calculated. However, if a mechanism for detecting halftone dots with higher accuracy is used as necessary, higher image quality can be achieved.

  The edge amount calculation unit may have a configuration other than that shown in this example. In this example, the edge amount is calculated using one filter for each band and direction. For example, as an edge amount calculation process that takes the maximum value of the primary differential filter and the secondary differential filter for each band and direction. Also good.

  Also, the conversion from the edge amount to the enhancement coefficient may be performed differently for each direction, for example, the enhancement in the diagonally right and diagonal directions is set to be weaker than the vertical and horizontal directions.

(Second embodiment)
In the second embodiment, an example will be described in which filter processing is performed by a low-frequency filter, an intermediate-frequency filter, and a single high-frequency filter in four directions of vertical, horizontal, right diagonal, and left diagonal. Regarding the feature amount calculation, common calculation is performed for medium frequency and high frequency, and calculation is separately performed for low frequency.

  FIG. 13 is a block diagram illustrating a configuration of a filter processing unit according to the second embodiment of the present invention. As shown in the figure, the filter processing unit 102 according to the second embodiment includes a smoothing unit 301 having a filter matrix, a low frequency enhancement unit 302, and a medium / high frequency enhancement unit 300.

  The middle / high frequency emphasizing unit 300 includes a middle frequency filter (vertical) 303 for correcting the spatial frequency of the middle frequency component in the vertical direction, and a middle frequency filter for correcting the spatial frequency of the middle frequency component in the horizontal direction, as shown in the figure. (Horizontal) 304, Medium frequency filter (right diagonal) 305 that corrects the spatial frequency of the middle frequency component in the right diagonal direction, Medium frequency filter (left diagonal) 306 that corrects the spatial frequency of the middle frequency component in the diagonal left direction It has.

  The medium / high frequency emphasizing unit 300 includes a medium / high frequency feature amount calculation (vertical) unit 307, a medium / high frequency feature amount calculation (horizontal) unit 308, and a medium / high frequency feature amount calculation (right diagonal) unit, as illustrated. 309, a medium / high frequency feature amount calculation (left diagonal) unit 310 is provided.

  Further, as shown in the figure, the medium / high frequency enhancement unit 300 includes an enhancement coefficient calculation unit 311 that converts an enhancement coefficient by the medium / high frequency feature amount calculation (vertical) unit 307, and a medium / high frequency feature amount calculation (horizontal) unit 308. Enhancement coefficient calculation unit 312 for converting enhancement coefficient, enhancement coefficient calculation unit 313 for converting enhancement coefficient by medium / high frequency feature amount calculation (right diagonal) unit 309, enhancement coefficient by middle / high frequency feature amount calculation (left diagonal) unit 310 An enhancement coefficient calculation unit 314 that converts the maximum value, a maximum value calculation unit 315 that calculates the maximum value of the medium / high-frequency edge amount in each direction, an enhancement coefficient calculation unit 316 that converts the enhancement coefficient after the maximum value calculation, and the heel direction of the high-frequency component Are provided with a high-frequency filter 317 as a single filter for correcting all of them.

  Further, as shown in the figure, a multiplication unit 318 that multiplies the output value of the medium frequency filter (vertical) 303 and the enhancement coefficient calculation unit 311, and a multiplication that multiplies the output value of the medium frequency filter (horizontal) 304 and the enhancement coefficient calculation unit 312. Unit 319, multiplication unit 320 that multiplies the output value of medium frequency filter (diagonal right) 305 and enhancement coefficient calculation unit 313, multiplication unit 321 that multiplies the output value of medium frequency filter (diagonal left) 306 and enhancement coefficient calculation unit 314 , The multiplication unit 322 that multiplies the output values of the enhancement coefficient calculation unit 316 and the high frequency filter 317, the addition unit 325 that adds the output values of the smoothing unit 301 and the addition unit 326, and the outputs of the low frequency enhancement unit 302 and the multiplication unit 327 An adder 326 for adding values, an adder 327 for adding the outputs of the multiplier 318 and the multiplier 328, an adder 328 for adding the outputs of the multiplier 319 and the adder 329, and a multiplier Adding section 329 for adding the output value of the adder 330 and the part 320, an addition section 330, for adding the output of the multiplication unit 321 and the multiplication unit 322.

  Next, the operation of the filter processing unit configured as described above will be described. First, the smoothing unit 301 smoothes image data in the same manner as described in the first embodiment. The low frequency emphasizing unit 302 and the medium / high frequency emphasizing unit 300 emphasize the low frequency and the medium / high frequency of the image, respectively, and the output from these and the smoothed data are added and output by the adding unit 325.326. Hereinafter, the low frequency emphasizing unit 302 and the middle / high frequency emphasizing unit 300 will be described.

  FIG. 14 is a block diagram showing a configuration of the low frequency emphasizing unit 302 in FIG. A region 350 surrounded by a dotted line is a portion where the low frequency vertical component is emphasized. The low frequency filter (vertical) 351 is a filter shown in FIG. The edge amount calculation unit (low frequency vertical) 352 is a filter shown in FIG. 6A. The edge amount calculated by this filter is converted into an enhancement coefficient by the enhancement coefficient calculation unit 353. A multiplication unit 354 multiplies the enhancement coefficient and the output value of the low frequency filter (vertical) 351 to calculate the low frequency vertical enhancement amount.

  Also, the amount of emphasis is calculated by the same configuration and method in the horizontal, right diagonal, and left diagonal directions. The low frequency filter (horizontal) includes a low frequency filter (horizontal) 355, an edge calculation unit 356, an enhancement coefficient calculation unit 357, and a multiplication unit 358. The low frequency filter (right diagonal) is a low frequency filter (right diagonal) 359. , An edge calculation unit 360, an enhancement coefficient calculation unit 361, and a multiplication unit 362. The low frequency filter (left diagonal) is a low frequency filter (left diagonal) 363, an edge calculation unit 364, an enhancement coefficient calculation unit 365, and a multiplication unit 366. As described above, the same low-frequency emphasis processing is performed for each direction, and these are combined.

  The low frequency filter in each direction is shown in FIGS. 4-2, 4-3, and 4-4, and the edge amount calculation filter in each direction is shown in FIGS. It is what is shown in FIGS. 6-4. The enhancement values in all four directions are added to obtain the output value of the low frequency enhancement unit 302.

  Next, the processing of the middle / high frequency emphasizing unit 300 shown in FIG. 13 will be described. Configuration of the medium frequency emphasizing unit, that is, a medium frequency filter (vertical) 303, a medium frequency filter (horizontal) 304, a medium frequency filter (right diagonal) 305, a medium frequency filter (left diagonal) 306, and a medium / high frequency feature amount in FIG. Calculation (vertical) unit 307, medium / high frequency feature value calculation (horizontal) unit 308, medium / high frequency feature value calculation (right diagonal) unit 309, medium / high frequency feature value calculation (left diagonal) unit 310, enhancement coefficient calculation unit 311 ˜314, the maximum value calculation unit 315, and the enhancement coefficient calculation unit 316 are the same as the operations of the respective blocks in FIG. 2, and only the characteristics (coefficients) of the respective filters are different. The intermediate frequency filters in each direction are shown in FIGS. 15-1 to 15-4, and the edge amount calculation filters in each direction are shown in FIGS. 16-1 to 16-4.

  That is, FIG. 15-1 is a vertical medium frequency filter, FIG. 15-2 is a horizontal medium frequency filter, FIG. 15-3 is a right diagonal medium frequency filter, and FIG. 16-1 is an edge calculation filter for a medium frequency filter in a vertical direction, FIG. 16-2 is an edge calculation filter for a medium frequency filter in a horizontal direction, and FIG. 16-3 is an edge for a medium frequency filter in an oblique right direction. Calculation filter, FIG. 16-4 is an edge calculation filter for a medium frequency filter in the diagonally left direction.

  The high frequency emphasizing unit may perform the same processing as the high frequency emphasizing process of the first embodiment after the maximum value is calculated using the above-described edge amount calculation result for medium frequency filter emphasis. The maximum value calculation unit 315 calculates the maximum value of the medium / high frequency edge amount calculation result in each direction, and the enhancement coefficient calculation unit 316 converts the result into an enhancement coefficient. By multiplying this enhancement coefficient by the output value of the high frequency filter 1317, the amount of high frequency enhancement is calculated. Here, the high frequency filter is a filter shown in FIG.

  Therefore, according to the configuration of the second embodiment, low frequency and medium frequency emphasis is performed independently in each direction, so that images with good resolution and sharpness are obtained for halftone dots and white ground characters. be able to. In addition, by emphasizing higher frequencies than the medium frequency with a single filter, density unevenness due to the difference in emphasis in each direction when the control is performed independently in each direction, or line breaks at the edge intersections. The phenomenon can be suppressed.

  In this example, for the filter in the diagonal direction, both the enhancement filter and the edge calculation filter are processed separately in the right diagonal direction and the left diagonal direction. However, this is not particularly limited, and both are processed as one diagonal filter. Alternatively, conversely, it is effective to further increase the direction from four directions.

  In the first embodiment and the second embodiment, edge enhancement and smoothing of each frequency band are performed in parallel. However, smoothing is not necessarily performed in parallel with edge enhancement. It can be done before or after emphasis.

(Third embodiment)
In the third embodiment, the method for emphasizing high-frequency components is changed as compared with the first embodiment.

  FIG. 17 is a block diagram illustrating the configuration of the high-frequency feature value calculation unit and each calculation unit according to the third embodiment of the present invention. Since the high-frequency feature value calculation units 210 to 213 are the same, the configuration of the high-frequency feature value calculation unit 210 is illustrated here. The high frequency feature amount calculation unit 210 includes a high frequency edge amount calculation (vertical) unit 401, a halftone dot correction amount calculation (vertical) unit 402, a high frequency edge amount calculation (horizontal) unit 403, a halftone dot correction amount calculation (horizontal) unit 404, High-frequency edge amount calculation (right diagonal) unit 405, halftone correction amount calculation (right diagonal) unit 406, high-frequency edge amount calculation (left diagonal) unit 407, halftone correction amount calculation (left diagonal) unit 408, subtraction unit 409- 412.

  In the configuration of FIG. 17, specifically, the maximum value in each direction of the halftone correction amount calculation results of the high frequency feature amount calculation units 210 to 213 shown in FIG. If the maximum value is equal to or greater than a predetermined value, it is determined that the dot is a halftone dot portion, and high frequency enhancement is not performed. If the maximum value is less than the predetermined value, it is determined that it is a non-halftone part, and enhancement processing similar to that in the first embodiment is executed.

  This is intended to avoid the generation of an image with poor graininess due to the undulation of the halftone dot when the halftone dot is excessively emphasized at a high frequency. For this reason, the configuration of this example is configured to obtain a sharp image with respect to the other image regions while suppressing defects caused by high-frequency enhancement up to the halftone dot portion.

  Therefore, according to each of the embodiments described above, the resolution and sharpness of the characters on the halftone dot provided by the filter control for each direction are combined by combining the filter processing controlled for each direction and the filter processing using a single filter. While obtaining a good image, it is possible to obtain a good image without density unevenness or line break even in a dense pattern on a white ground (on a non-halftone dot).

  In addition, an increase in hardware scale can be suppressed by using a filter that emphasizes all directions collectively for emphasizing the high frequency band. Further, by controlling the filtering process according to the feature amount of the image, it is possible to obtain an image having both graininess and sharpness.

  As described above, the image processing apparatus and the image processing method according to the present invention are useful for image processing in an image forming apparatus such as a digital copying machine and a facsimile apparatus, and an image input apparatus such as an image scanner and a digital camera. The present invention is suitable for an image processing apparatus and an image processing method for performing suitable filter processing on a character document such as a catalog or a flyer, and a general document such as a character document on halftone dots, or a computer system for performing these processes.

1 is a block diagram illustrating a configuration of an image processing apparatus and system according to an embodiment of the present invention. It is a block diagram which shows the structure of the filter process part concerning the 1st Embodiment of this invention. It is explanatory drawing which shows the filter matrix of the smoothing filter concerning embodiment of this invention. It is explanatory drawing which shows the filter matrix of the low frequency emphasis filter (vertical direction) concerning the 1st and 2nd embodiment of this invention. It is explanatory drawing which shows the filter matrix of the low frequency emphasis filter (horizontal direction) concerning the 1st and 2nd embodiment of this invention. It is explanatory drawing which shows the filter matrix of the low frequency emphasis filter (right diagonal direction) concerning the 1st and 2nd embodiment of this invention. It is explanatory drawing which shows the filter matrix of the low frequency emphasis filter (left diagonal direction) concerning the 1st and 2nd embodiment of this invention. It is a block diagram which shows the structure of the low frequency feature calculation part in FIG. It is explanatory drawing which shows the filter matrix of the low frequency edge amount calculation filter (vertical direction) concerning the 1st and 2nd embodiment of this invention. It is explanatory drawing which shows the filter matrix of the low frequency edge amount calculation filter (horizontal direction) concerning the 1st and 2nd embodiment of this invention. It is explanatory drawing which shows the filter matrix of the low frequency edge amount calculation filter (right diagonal direction) concerning the 1st and 2nd embodiment of this invention. It is explanatory drawing which shows the filter matrix of the low frequency edge amount calculation filter (left diagonal direction) concerning the 1st and 2nd embodiment of this invention. It is a graph which shows the relationship between the edge amount concerning embodiment of this invention, and the enhancement amount. It is explanatory drawing which shows the filter matrix of the high frequency emphasis filter concerning the 1st and 2nd embodiment of this invention. It is a block diagram which shows the structure of the high frequency feature-value calculation part in FIG. It is explanatory drawing which shows the filter matrix of the high frequency edge amount calculation filter (vertical direction) concerning the 1st Embodiment of this invention. It is explanatory drawing which shows the filter matrix of the high frequency edge amount calculation filter (horizontal direction) concerning the 1st Embodiment of this invention. It is explanatory drawing which shows the filter matrix of the high frequency edge amount calculation filter (right diagonal direction) concerning the 1st Embodiment of this invention. It is explanatory drawing which shows the filter matrix of the high frequency edge amount calculation filter (left diagonal direction) concerning the 1st Embodiment of this invention. It is explanatory drawing which shows the example of calculation of the high frequency halftone correction amount (vertical direction) concerning the 1st Embodiment of this invention. It is explanatory drawing which shows the example of calculation of the high frequency halftone correction amount (horizontal direction) concerning the 1st Embodiment of this invention. It is explanatory drawing which shows the example of calculation of the high frequency halftone correction amount (right diagonal direction) concerning the 1st Embodiment of this invention. It is explanatory drawing which shows the example of calculation of the high frequency halftone correction amount (left diagonal direction) concerning the 1st Embodiment of this invention. It is a flowchart which shows the procedure of the spatial frequency correction | amendment concerning the 1st Embodiment of this invention. It is a block diagram which shows the structure of the filter process part concerning the 2nd Embodiment of this invention. It is a block diagram which shows the structure of the low frequency emphasis part in FIG. It is explanatory drawing which shows the filter matrix of the intermediate frequency filter (vertical direction) concerning the 2nd Embodiment of this invention. It is explanatory drawing which shows the filter matrix of the intermediate frequency filter (horizontal direction) concerning the 2nd Embodiment of this invention. It is explanatory drawing which shows the filter matrix of the intermediate frequency filter (right diagonal direction) concerning the 2nd Embodiment of this invention. It is explanatory drawing which shows the filter matrix of the intermediate frequency filter (left diagonal direction) concerning the 2nd Embodiment of this invention. It is explanatory drawing which shows the filter matrix of the edge calculation filter (vertical direction) for medium frequency filters concerning the 2nd Embodiment of this invention. It is explanatory drawing which shows the filter matrix of the edge calculation filter for horizontal frequency filters concerning the 2nd Embodiment of this invention (horizontal direction). It is explanatory drawing which shows the filter matrix of the edge calculation filter for middle frequency filters concerning the 2nd Embodiment of this invention (right diagonal direction). It is explanatory drawing which shows the filter matrix of the edge calculation filter for medium frequency filters concerning the 2nd Embodiment of this invention (left diagonal direction). It is a block diagram which shows the structure of the high frequency feature-value calculation part concerning each of the 3rd Embodiment of this invention, and each calculation part.

Explanation of symbols

DESCRIPTION OF SYMBOLS 100 Image input device 102 Filter processing part 110 Image processing apparatus 202 Low frequency filter (vertical)
203 Low frequency filter (horizontal)
204 Low frequency filter (right diagonal)
205 Low frequency filter (left diagonal)
206 Low-frequency feature quantity calculation (vertical) 207 Low-frequency feature quantity calculation (horizontal) section 208 Low-frequency feature quantity calculation (right diagonal) section 209 Low-frequency feature quantity calculation (left diagonal) section 210 High-frequency feature quantity calculation (vertical) Unit 211 high-frequency feature amount calculation (horizontal) unit 212 high-frequency feature amount calculation (right diagonal) unit 213 high-frequency feature amount calculation (left diagonal) units 214 to 217, 219 enhancement coefficient calculation unit 218 maximum value calculation unit 220 high-frequency filters 221 to 225 Multiplier 226 to 230 Adder 231 Low frequency edge amount calculator 232 High frequency edge amount calculator 233 Halftone correction amount calculator 234 Subtractor 300 Medium / high frequency enhancer 415 Maximum value calculator 416 Halftone dot determiner

Claims (9)

In an image processing apparatus that performs desired spatial frequency correction on an input image signal,
First spatial frequency correction means for performing spatial frequency correction for each direction of the image signal using directional filters each having directionality;
Second spatial frequency correction means that collectively performs spatial frequency correction in all directions of the image signal using a single filter having a frequency characteristic different from that of the directional filter;
Feature amount calculating means for calculating the feature amount of the image signal for each directionality;
Third spatial frequency correction means for controlling the output of the corresponding directivity of the first spatial frequency correction means according to the feature quantity for each directionality calculated by the feature quantity calculation means;
An image processing apparatus comprising:   The image processing apparatus according to claim 1, wherein the first spatial frequency correction unit and the second spatial frequency correction unit are arranged to process the image signal in parallel. 3. The image processing apparatus according to claim 1, wherein the filter frequency characteristic of the second spatial frequency correction unit is a high frequency characteristic with respect to the filter frequency characteristic of the first spatial frequency correction unit. . The image processing apparatus according to claim 1 , wherein the feature amount calculated by the feature amount calculation unit is an edge amount of an image. Feature amount calculating means for calculating the feature amount of the image signal for each directionality;
Maximum value calculating means for calculating the maximum value of the feature amount acquired for each direction;
Fourth spatial frequency correction means for controlling the output of the second spatial frequency correction means in accordance with the maximum feature value for each direction;
The image processing apparatus according to claim 1, 2, or 3. The feature amount, the image processing apparatus according to claim 1 or 5, characterized in that is calculated using the dot correction value indicating the edge amount and the local halftone degree of image. The image processing apparatus according to claim 6 , wherein the feature amount is a value obtained by subtracting the halftone dot correction value from an edge amount of the image. Halftone correction amount calculating means for calculating a halftone correction amount for each direction;
Halftone dot maximum value calculating means for calculating the maximum value of the halftone dot correction amount for each direction calculated by the halftone dot correction amount calculating means;
With
The fourth spatial frequency correction means, when the halftone correction amount acquired for each direction is equal to or less than a predetermined value, the second spatial frequency correction means according to the maximum value of the feature amount of the image for each direction. The image processing apparatus according to claim 5 , wherein the output of the image processing apparatus is controlled. In an image processing method for executing a desired spatial frequency correction on an input image signal,
A first step of performing spatial frequency correction for each direction on the image signal using directional filters each having directionality;
A second step of collectively performing spatial frequency correction in all directions of the image signal using a single filter having a frequency characteristic different from that of the directional filter;
A third step of calculating a feature amount of the image signal for each directionality;
A fourth step of controlling the output of the corresponding direction of the first step according to the feature amount for each direction calculated in the third step;
An image processing method comprising:

Images