JP2017038167A - Image processing apparatus, image processing method, and program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To reduce a mosquito noise while maintaining the image quality of extended image data.SOLUTION: There is provided an image forming apparatus 10 processing extended image data obtained by extending compressed image data, the image forming apparatus comprising: a white background detection part 302 that detects a white background image area in the extended image data; an edge detection part 303 that detects an edge part in the extended image data; a background removal parameter determination part 304 that determines the value of a background removal parameter to be applied to each of areas of the extended image data for each of the areas on the basis of results of detection performed by the white background detection part 302 and the edge detection part 303; and a background removal processing part 305 that performs background removal processing on the extended image data by using the value of the background removal parameter determined by the background removal parameter determination part 304.SELECTED DRAWING: Figure 3

Description

  The present invention relates to an image processing apparatus, an image processing method, and a program.

Conventionally, when image data is transmitted / received to / from a third party using a network or the like, processing for compressing the image data has been performed in order to reduce the data capacity and enable easy transmission / reception. .
However, when irreversible compression is performed, when the compressed image data is decompressed, so-called mosquito noise occurs in a portion with a lot of high-frequency components around the edges of characters and lines (a portion where the color of the background is abruptly different). However, if the decompressed image data is output as it is, there is a problem that the output image deteriorates.

As a method for solving this problem, a background removal processing method according to the compression rate is already known. However, in this background removal processing method, when image data compressed at a high compression rate is decompressed, there is a problem that the gradation of the highlight portion is skipped and a pseudo contour is likely to occur.
For this reason, when image data with a low mosquito noise generation rate, that is, with low high-frequency components is compressed at a high compression rate, the effect of removing the mosquito noise can be achieved by performing background processing according to the high compression rate. There was a problem that the adverse effect that pseudo contours are likely to occur is greater.

Therefore, as a technique for solving such a problem, Patent Document 1 discloses a method of switching a background removal processing parameter using a parameter at the time of irreversible compression. In this method, background removal is performed by selecting an optimum background removal parameter obtained in advance from parameters used during lossy compression such as a quantization matrix.
According to this method, it is possible to reduce mosquito noise while minimizing adverse effects such as generation of pseudo contours.

The background removal processing method described in Patent Document 1 is effective when parameters for irreversible compression can be acquired, such as when image data is compressed and decompressed in the same computer system.
However, there are cases where parameters for lossy compression cannot be acquired, as in the case where image data compressed by another computer system is obtained via e-mail, USB memory, or the like. In such a case, there is a problem that the background removal processing method described in Patent Document 1 cannot be applied.

  The present invention solves such a problem and can reduce mosquito noise while maintaining the image quality of the decompressed image data even when the parameters used for the compression processing cannot be referred to when decompressing the compressed image data. The purpose is to do so.

  In order to achieve the above object, an image processing apparatus according to the present invention detects a white background image area in the decompressed image data in an image processing apparatus that processes decompressed image data obtained by decompressing compressed image data. White background detecting means, edge detecting means for detecting edge portions in the decompressed image data, and background removal parameters to be applied to each area of the decompressed image data based on detection results of the white background detecting means and the edge detecting means. And a processing means for performing background removal processing on the decompressed image data using the background removal parameter value determined by the determination means.

  According to the above configuration, when decompressing compressed image data, it is possible to reduce mosquito noise while maintaining the image quality of the decompressed image data.

1 is a block diagram illustrating a hardware configuration of an image forming apparatus that is a first embodiment of an image processing apparatus according to the present invention; FIG. FIG. 2 is a block diagram illustrating a schematic configuration of functions for forming an image on a sheet based on compressed image data among the functions of the image forming apparatus illustrated in FIG. 1. It is a block diagram which shows in more detail the function structure of the mosquito noise reduction process part shown in FIG. FIG. 3 is a diagram schematically illustrating a state of a part of decompressed image data handled by the image forming apparatus. It is a figure which shows an example of the filter matrix for edge detection. It is a figure which shows an example of the content of the gamma table for noise reduction pixels. It is a figure which shows an example of the content of the gamma table for non-noise reduction pixels. 3 is a flowchart showing a procedure of mosquito noise reduction processing executed by the image forming apparatus shown in FIG. It is a figure for demonstrating the relationship between the content of the expansion | extension image data, and the amount of edges. It is a figure which shows an example of the content of the gamma table for noise reduction pixels used in 2nd Embodiment. It is a figure which shows the other example. FIG. 8 is a flowchart corresponding to FIG. 7, showing a procedure of mosquito noise reduction processing executed by the image forming apparatus of the second embodiment of the present invention. It is a figure which shows an example of the content of the gamma table used by 3rd Embodiment. It is a figure which shows the other example. It is a figure which shows another example. It is a figure which shows another example. FIG. 11 is a flowchart corresponding to FIG. 10 showing a procedure of mosquito noise reduction processing executed by the image forming apparatus of the third embodiment of the present invention. It is a figure for demonstrating distribution of the edge amount in an edge detection area | region. FIG. 13 is a flowchart corresponding to FIG. 12 showing a procedure of mosquito noise reduction processing executed by the image forming apparatus of the fourth embodiment of the present invention. It is a figure corresponding to FIG. 3 which shows the function structure of the mosquito noise reduction process part in the image forming apparatus of the 1st modification of this invention. It is a figure for demonstrating arrangement | positioning of the edge detection area | region in the image forming apparatus of the 1st modification of this invention. It is a block diagram corresponding to FIG. 1 which shows the hardware constitutions of the portable terminal device which is the 2nd modification of the image processing apparatus of this invention. It is a block diagram corresponding to FIG. 2 which shows the function structure of the portable terminal device which is the 2nd modification of the image processing apparatus of this invention.

Hereinafter, embodiments for carrying out the present invention will be described with reference to the drawings.
[First Embodiment: FIGS. 1 to 7]
First, an image processing apparatus according to a first embodiment of the present invention will be described. FIG. 1 is a block diagram illustrating a hardware configuration of the image forming apparatus.
An image forming apparatus 10 illustrated in FIG. 1 includes a controller 101, an engine control unit 110, an image reading unit 111, a plotter unit 112, and an image processing unit 113. The controller 101 includes a CPU 102, a RAM 103, a ROM 104, an HDD (hard disk drive) 105, a communication I / F (interface) 106, and an operation unit I / F 107, and the operation unit I / F 107 is connected to the operation unit 108. These are connected by a system bus 109.

Of the above configuration, the CPU 102 controls the entire image forming apparatus 10 by executing a program stored in the ROM 104 or the HDD 105 using the RAM 103 as a work area, and realizes various functions to be described later.
The ROM 104 and the HDD 105 are non-volatile storage media (storage means) and store various programs executed by the CPU 102 and various data described later.
The communication I / F 106 is an interface for connecting the image forming apparatus 10 to a network.

The operation unit I / F 107 is an interface for connecting the operation unit 108 to the system bus 109 and enabling control from the CPU 102.
The operation unit 108 is a user interface including an operation unit such as a key, a button, and a touch sensor for receiving an operation from a user, and a display unit such as a display for presenting information to the user.

The image reading unit 111 is an image reading unit having a function of reading an image of a document and acquiring the image data. The plotter unit 112 is an image forming unit having a function of forming an image on a sheet by printing based on input image data. The image processing unit 113 is an image processing unit that performs various image processes including a mosquito noise reduction process described later on the image data read by the image reading unit 111 and the image data used for image formation in the plotter unit 112.
The engine control unit 110 is a control unit that controls the image reading unit 111, the plotter unit 112, and the image processing unit 113 in accordance with commands supplied from the CPU 102 via the system bus 109.

The image forming apparatus 10 described above can be configured as, for example, a printer, a facsimile communication apparatus, a digital multifunction peripheral (MFP), or the like.
One characteristic point of the image forming apparatus 10 is mosquito noise reduction processing performed when decompressing compressed image data. Hereinafter, this point will be described.

First, mosquito noise will be briefly described. Generally, when irreversibly compressing image data, the amount of data is reduced by thinning out high-frequency components during compression. However, even if such image data is expanded, the thinned high-frequency component cannot be restored at the time of expansion, and noise is generated. This is so-called mosquito noise. It is known that this mosquito noise does not occur so much in a natural image with gradation, and is likely to occur in a boundary portion (a portion where the color is abruptly different from the background) such as characters and lines.
Therefore, in the image processing apparatus of this embodiment, the background removal processing method is switched between a place where mosquito noise is likely to occur and a place where it is not, and a background removal parameter suitable for the pixel is determined for each pixel. By executing background processing for each pixel using the background removal parameter, mosquito noise is reduced while maintaining image quality.

Next, functions related to the mosquito noise reduction process provided in the image forming apparatus 10 will be described with reference to FIGS. 2 and 3.
First, FIG. 2 shows a schematic configuration of functions for forming an image on a sheet based on compressed image data.
As shown in FIG. 2, the image forming apparatus 10 includes a compressed image acquisition unit 201, a decoder unit 202, a mosquito noise reduction processing unit 203, another processing unit 204, and a print execution unit 205. The functions of these units, the functions of the compressed image acquisition unit 201 and the decoder unit 202 are controlled by the hardware of the controller 101, and the functions of the mosquito noise reduction processing unit 203 and the other processing unit 204 are controlled by the controller 101. Thus, the function of the print execution unit 205 is realized by the controller 101 controlling the plotter unit 112, respectively.

  The compressed image acquisition unit 201 has a function of acquiring compressed image data and passing it to the decoder unit 202. This acquisition is performed by acquiring communication data such as a network from an external device such as a personal computer (PC) (Personal Computer) or a smartphone, or an external device such as a scanner. And acquiring compressed image data.

The decoder unit 202 performs decompression decoding processing (hereinafter referred to as “decompression processing”) such as inverse quantization or inverse DCT on the image acquired by the compressed image acquisition unit 201 in accordance with a compression encoding method (hereinafter referred to as “compression method”). )) To generate decompressed image data. When the compression is lossy compression, a function of passing the decompressed image data to the mosquito noise reduction processing unit 203 is provided.
Further, the decoder unit 202 has a function of bypassing the decompressed image data to the other processing unit 204 by bypassing the mosquito noise reduction processing unit 203 when the compression is lossless compression. Since the lossless compression is a compression method in which the image data before compression and the image data after decompression are completely equal, the mosquito noise is not generated by the compression and decompression, and the image is corrected. It is not necessary to consider.

The mosquito noise reduction processing unit 203 has a function of executing processing for reducing mosquito noise that occurs when decompressing irreversibly compressed image data and passing the processed image data to the other processing unit 204. The function of the mosquito noise reduction processing unit 203 will be described in detail with reference to FIG.
The other processing unit 204 has a function of performing correction processing for processing the expanded image data passed from the decoder unit 202 or the mosquito noise reduction processing unit 203 into image data suitable for printing, such as gamma correction processing. In addition, a function of passing the decompressed image data subjected to the correction process to the print execution unit 205 is also provided.
The print execution unit 205 has a function of executing printing based on the image data received from the other processing unit 204 and forming an image on a sheet.

Next, FIG. 3 shows in more detail the functional configuration of the mosquito noise reduction processing unit 203 shown in FIG.
As shown in FIG. 3, the mosquito noise reduction processing unit 203 includes an expanded image reception unit 301, a white background detection unit 302, an edge detection unit 303, a background removal parameter determination unit 304, and a background removal processing unit 305.
Among these, the expanded image receiving unit 301 has a function of receiving expanded image data supplied from the decoder unit 202.

  The white background detection unit 302 has a function of a white background detection unit that detects a white background image area in the image data of the expanded image data received by the expanded image reception unit 301. The detection of the white background image region may be appropriately performed using a known method. For example, as described in Japanese Patent Application Laid-Open No. 2003-46772, when the surrounding area of the target pixel is referred to, white pixels are determined by threshold comparison for the pixels in the surrounding area, and the pixels determined as white pixels are arranged in a certain pattern It is conceivable to determine the target pixel as a white background. In this case, even if the target pixel is not a white pixel, the white background detection unit 302 determines that the target pixel is a white background image area if it is known from the surrounding pixel information that the target pixel is white.

The edge detection unit 303 has a function of an edge detection unit that detects pixels constituting the edge portion in the expanded image data received by the expanded image reception unit 301. The edge portion is a portion where the pixel value of the image changes abruptly at the boundary between two regions composed of substantially constant pixel values, such as a boundary between a character or a line and a background portion.
In this embodiment, the edge detection unit 303 performs a calculation for each pixel in the decompressed image data to obtain a feature amount (edge amount) indicating the steepness of the variation degree of the pixel value around the pixel. Then, based on the edge amount, it is determined whether each pixel is a pixel constituting the edge portion. The method for calculating the edge amount will be described in detail later.

  The background removal parameter determination unit 304 is a function of a determination unit that determines the value of the background removal parameter to be applied to each region of the expanded image data based on the detection results of the white background detection unit 302 and the edge detection unit 303. Is provided. In this embodiment, the size of the region is one pixel, and the background removal parameter determination unit 304 determines the value of the background removal parameter for each pixel. However, the present invention is not limited to this, and the background removal parameter value may be determined for each region composed of a plurality of pixels.

The background removal processing unit 305 has a function of performing background removal processing on the decompressed image data received by the decompressed image receiving unit 301 using the value of the background removal parameter determined by the background removal parameter determining unit 304. Here, gamma processing is performed as background removal processing, and the background removal parameter used for the processing is a gamma correction table that defines the correspondence between input and output of pixel values, but is not limited thereto.
The background removal processing unit 305 also has a function of outputting the image data after the above background removal processing to the other processing unit 204 in FIG.
The above is the function of the image forming apparatus 10.

Next, a specific example of mosquito noise reduction processing executed by each of the above units will be described with reference to FIGS. 4 to 6B.
FIG. 4 shows an example of an image indicated by the expanded image data. FIG. 4 shows only a region of 10 pixels × 10 pixels in the image. A high density area indicated by W is a character area (area where pixels constituting a character or a line are gathered), and pixels of mosquito noise generated around the area due to expansion are indicated by hatching. Yes.
Here, since the mosquito noise reduction processing is performed in units of one pixel, the pixel to be processed is indicated by P1, and the edge detection region of M × M pixels including P1 is indicated by Q1. The size of the edge detection region Q1 is 5 pixels × 5 pixels here, but the size is not limited to this. Further, it is not essential that the pixel P1 to be processed is in the center of the edge detection region Q1.

In the mosquito noise reduction processing unit 203 of the image forming apparatus 10, the white background detection unit 302 detects whether or not the pixel P1 to be processed in the decompressed image data is within the white background image region.
The edge detection unit 303 detects whether or not there is a pixel constituting the edge portion in the edge detection region Q1 based on the edge amount obtained for each pixel in the edge detection region Q1 including the pixel P1 to be processed.

Here, the edge amount for each pixel is calculated by convolving a filter matrix as shown in FIG. 5 with each pixel value in a 5 × 5 pixel region centered on the target pixel (each region in the region). A value obtained by multiplying the pixel value of the pixel by the filter value at the corresponding position is added to all the pixels in the region).
The filter in FIG. 5A is a filter for detecting a vertical edge, and when a vertical edge exists in a 5 × 5 region, a large absolute value can be obtained by a convolution operation. . Similarly, in FIG. 5B, a large absolute value is obtained when a lateral edge portion exists.

Therefore, in this embodiment, a value obtained by adding the absolute value of the result of the convolution operation using the filter of FIG. 5A and the absolute value of the result of the convolution operation using the filter of FIG. The edge amount is obtained. Therefore, this edge amount becomes a large value when the edge portion in the vertical direction or the horizontal direction exists in the vicinity of the target pixel, and becomes a small value when the edge portion in any direction does not exist.
The edge detection unit 303 detects pixels in which the above edge amount is equal to or greater than a predetermined value as pixels constituting the edge part.
In this embodiment, the size of the filter matrix is 5 × 5 pixels, which is the same as the size of the edge detection region Q1, but it is not necessary to be the same, and the size itself is not limited to 5 × 5.

Based on the detection results of the white background detection unit 302 and the edge detection unit 303, the background removal parameter determination unit 304 determines the background removal parameter to be applied to the pixel P1 to be processed as shown in Table 1.

  6A and 6B show the contents of the noise reduction pixel gamma table and the non-noise reduction pixel gamma table, respectively. 6A and 6B show the correspondence between the pixel value of the input image data and the pixel value of the output image data at the time of gamma processing. In all embodiments including those described later, the pixel value “0” is the minimum density (in this case, white), and “255” is the maximum density (here, black). Color).

Among these, in the noise reduction pixel gamma table shown in FIG. 6A, the output value is kept at 0 when the input value is from 0 to an appropriate threshold value, and the input value is 255 when the input value is larger than the threshold value. The output value is linearly increased as the input value increases so that the output value becomes 255.
That is, the noise reduction pixel gamma table stipulates that a pixel whose gradation value is on the white side of the threshold value is changed so that the gradation value becomes a white value by gamma processing. . As a result, it is possible to delete dots having low gradation values that are considered to be formed as mosquito noise while leaving dots having large gradation values (high density).

In the example of FIG. 6A, the output value is set to 0 when the input value is between 0 and an appropriate threshold. However, even if it is not necessarily set to 0, if the output value is made lower than the input value, the effect of reducing mosquito noise can be obtained. It is done.
In the gamma table for non-noise reduction pixels shown in FIG. 6B, the input value and the output value are the same for all input values. That is, when this gamma table is used, the contents of the image are not changed.

Here, when decompressing irreversibly compressed image data, the high-frequency component signal is dropped at the time of compression, so the high-frequency component cannot be restored at the time of decoding. In addition, mosquito noise is generated. On the other hand, it does not occur much in natural images such as gradation.
In this embodiment, the mosquito noise is reduced without degrading the image quality of the highlight portion of the gradation by switching the background removal processing method between the place where the mosquito noise such as characters and lines is likely to occur and the place where the mosquito noise is not likely to occur. I have to.

Specifically, as shown in Table 1, mosquito noise is likely to occur in pixels that are in the white area and in the vicinity of the edge (the pixels that form the edge portion in the edge detection area). Therefore, the noise reduction pixel gamma table is applied to such pixels. In addition, the non-noise reduction pixel gamma table is applied to other pixels.
As a result, mosquito noise can be selectively reduced with respect to pixels that are likely to generate mosquito noise by simple processing, and mosquito noise can be reduced while maintaining the image quality of the expanded image data. Further, this process can be performed based only on information included in the decompressed image data and information stored in advance in the image forming apparatus 10, and it is not necessary to refer to parameters used in the compression process.

  The reason for considering not only the edge but also the white background region is to prevent the image density from being changed by the background removal process. As can be seen from FIG. 6B, when the gamma correction process using the noise reduction pixel gamma table is performed, the pixel value fluctuates particularly in a low pixel value range. However, if the input pixel value is “0”, the output is also “0” and there is no fluctuation. For this reason, even if the noise reduction pixel gamma table is applied within the white background region, the influence on the image is small. In consideration of this, in this embodiment, the range for providing the noise reduction pixel gamma table is limited to pixels in the white background region.

Next, FIG. 7 shows the procedure of the mosquito noise reduction process described above. This process is executed by a processor included in the image processing unit 113. However, the execution order does not necessarily have to be as shown in FIG. 7, for example, a plurality of parts are executed in parallel. 7 is processing according to the embodiment of the image processing method of the present invention, and corresponds to the function of the mosquito noise reduction processing unit 203.
In the process of FIG. 7, first, the image processing unit 113 sets one pixel that has not yet been processed among the pixels of the decompressed image data as a processing target (S101).
Next, the image processing unit 113 performs white background detection processing for detecting whether or not the processing target pixel is in the white background pixel region (S102). This process is a white background detection procedure and corresponds to the function of the white background detection unit 302.

  If the determination result is in the white background image area (Yes in S103), the image processing unit 113 still calculates the edge amount among the pixels existing in the edge detection area of a predetermined size including the pixel to be processed. The edge amount is calculated for pixels that have not been processed (S104). This process is an edge detection procedure process and corresponds to the function of the edge detection unit 303. The amount of edge for each pixel in the edge detection area is obtained by the processing in step S104.

  Next, the image processing unit 113 determines whether there is a pixel whose edge amount is equal to or larger than a predetermined value in the edge detection region (S105). If Yes here, the image processing unit 113 determines that there is a pixel constituting the edge portion in the vicinity of the pixel to be processed, and uses the noise reduction pixel gamma table as a background removal parameter for the pixel to be processed. It is determined to apply (S106).

On the other hand, in the case of No in either step S103 or S105, the image processing unit 113 determines to apply the non-noise-reduced pixel gamma table as the background removal parameter for the pixel to be processed (S107).
The processes in steps S104 to S107 described above are determination procedure processes and correspond to the function of the background removal parameter determination unit 304.

  Further, after step S106 or S107, the image processing unit 113 performs gamma correction as background removal processing on the pixel to be processed using the gamma table determined in these steps (S108). This processing is processing of the processing procedure, and corresponds to the function of the background removal processing unit 305.

Thereafter, if there is a pixel that is not yet a processing target (Yes in S109), the image processing unit 113 returns to step S101 and repeats the processing using the next pixel as a processing target. When the processes in steps S101 to S108 are completed for all the pixels, the determination in step S109 is No and the process ends.
By the above process, the mosquito noise reduction process described with reference to FIGS. 4 to 6B can be executed.

  The size of the edge detection area Q1 may be 1 × 1 pixel. In this case, the range of the edge detection region Q1 including the pixel P1 to be processed is only the pixel P1 itself. Then, the determination in step S105 is performed according to the edge amount obtained for the pixel P1. Even in this case, since the presence / absence of an edge can be detected for pixels in a range corresponding to the size of the filter matrix, mosquito noise reduction processing can be performed with a certain degree of accuracy.

[Second Embodiment: FIGS. 8 to 10]
Next, a second embodiment of the image processing apparatus of the present invention will be described.
The basic configuration of the second embodiment is the same as that of the first embodiment, except that the background removal parameter value (content of the gamma table) applied to each pixel is determined based on the calculated edge amount. Different. Therefore, the description regarding the common part is omitted, and only matters related to the differences will be described. Moreover, the same code | symbol is used about the structure which is common or respond | corresponds with 1st Embodiment.

First, the basic concept regarding the determination of the value of the background removal parameter in the second embodiment will be described.
Here, consider two types of images 400 and 410 shown in FIG. The image 400 includes a white region 400a having a pixel value “0” and a black region 400b having a pixel value “255”. When pixel values are plotted for each horizontal position in the figure, a graph 401 is obtained. become. The boundary between the white area 400a and the black area 400b is an edge portion. It is assumed that the pixels having the same horizontal position have the same pixel value (the boundary line between the white area 400a and the black area 400b is the vertical direction in the figure).
On the other hand, the image 410 is composed of a white area 410a having a pixel value of “0” and a gray area 410b having a pixel value of “127”, and the boundary is the same position as in the case of the image 400. When the pixel value of each pixel is plotted for each position in the direction, a graph 411 is obtained.

For each of these images, the edge amount for each pixel for each horizontal position is obtained in the same manner as in the first embodiment and plotted in the same manner as in the graphs 401 and 411, so that graphs 402 and 412 are obtained. That is, when there is a boundary line between the white area 400a and the black area 400b within the 5 × 5 pixel area where the convolution calculation is performed with the filter matrix, the edge amount becomes a large value, and in other cases, the edge detection area The edge amount is 0 because there is no change in the pixel value.
However, since the pixel values of the pixels constituting the edge element are different, the value of the edge amount is different between the image 400 and the image 410, and the value of the edge amount is larger in the case of the image 400 having a larger difference in pixel value. growing.

Here, for the pixels in the vicinity of the edge, ignoring the edge amount due to the difference in density of characters and lines, and performing gamma correction uniformly using the noise reduction pixel gamma table shown in FIG. 6A, As a result, the pixel value of the light-colored character portion may fluctuate and the color of the character portion may be lost. However, if a gamma table to be used as a background removal parameter is determined using an edge amount that reflects the difference in shade in consideration of the difference in shade of color, such a situation can be prevented. .
Therefore, in the second embodiment, the edge amount E of the pixel having the maximum edge amount among the pixels in the edge detection region Q1 including the pixel P1 to be processed is set as the steepness of the edge portion in the edge detection region Q1. The background removal level is calculated using the feature amount indicating the. Further, the noise reduction pixel gamma table to be applied to the pixel P1 is determined using the background removal level.

The background removal level th_s is calculated according to the following formula A.
th_s = (edge amount E ÷ maximum value that edge amount can take) × max_th (A)
The “maximum value that the edge amount can take” is the theoretical maximum value of the edge amount obtained by using the filter matrix used for calculating the edge amount. max_th is a background removal level set when the edge amount is the maximum value.

Using the above background removal level th_s, the gamma table for the noise reduction pixel keeps the output value 0 when the input value is from 0 to th_s, and the input value is 255 when the input value is larger than th_s. The output value may be generated so as to increase linearly as the input value increases so that the value becomes 255. That is, the value of th_s is the threshold described with reference to FIG. 6A.
Further, since the value of th_s increases as the edge amount E increases, the threshold value can be set to a value farther from the white side.

For example, the noise reduction pixel gamma table for th_s = 20 has the contents shown in FIG. 9A, and the noise reduction pixel gamma table for th_s = 10 has the contents shown in FIG. 9B.
In addition, if th_s = 20 in the case of the edge amount of the pixel near the boundary between the white region 400a and the black region 400b of the image 400, the edge amount is about half at the corresponding portion of the image 410, so th_s = 10. It becomes. Therefore, according to the above concept, the strength of the background removal process can be adjusted according to the density of the characters and lines, and the light color characters and lines can be prevented from disappearing by the background removal process.
The noise reduction pixel gamma table may be calculated for each pixel, or stored in advance in an appropriate storage unit for each value of the background removal level th_s, and a table corresponding to the value of the background removal level th_s. May be read out.

Next, FIG. 10 shows the procedure of the mosquito noise reduction process of the second embodiment described above.
This procedure corresponds to the procedure of FIG. 7 described in the first embodiment, and is different from the process of FIG. 7 in that steps S121 and S122 are executed instead of step S106.
That is, in the case of Yes in step S105, the image processing unit 113 generates a noise reduction pixel gamma table having a background removal level th_s corresponding to the edge amount E of the pixel having the maximum edge amount in the edge detection region ( S121). The detailed generation procedure is as described above.

Then, the image processing unit 113 determines to apply the noise reduction pixel gamma table generated in step S121 as the background removal parameter for the pixel to be processed (S122), and proceeds to step S108.
With the above processing, the mosquito noise reduction processing described with reference to FIGS. 8 to 9B can be executed. In the case of No in step S103 or S105, the non-noise reduction pixel gamma table is applied as in the case of the first embodiment.

[Third Embodiment: FIGS. 11A to 12]
Next, a third embodiment of the image processing apparatus of the present invention will be described.
This third embodiment has the same basic configuration as that of the second embodiment, but differs in a method for determining the value of the background removal parameter (content of the gamma table) applied to each pixel based on the calculated edge amount. . Only matters related to this difference will be described below. Moreover, the same code | symbol is used about the structure which is common or respond | corresponds with 2nd Embodiment.

Assume that the background removal parameter is continuously changed according to the value of the edge amount E of the pixel having the maximum edge amount among the pixels in the edge detection region Q1 including the pixel P1 to be processed as in the second embodiment. Although fine adjustment is possible, a large number of bits are required for data transmission, which increases the cost of the circuit.
Therefore, when the influence of the mosquito noise on the image quality is such that the adjustment as precise as the example of the second embodiment is not required, in order to reduce the number of bits, the edge determination result of the edge amount E is, for example, about 2 bits. It is conceivable that the data is output after being converted to. In the third embodiment, this configuration is adopted, and a gamma table corresponding to each edge determination result is stored in advance in an appropriate storage unit, and the output gamma table corresponding to the edge determination result is subjected to background removal processing. To apply to.

Table 2 shows an example of the correspondence between the edge amount E and the edge determination result.

In addition, FIGS. 11A to 11D show examples of the contents of the gamma table corresponding to each edge determination result.
11A is used when the edge determination result is “11”, and corresponds to the noise reduction pixel gamma table having the background removal level th_s of 40 described in the second embodiment.

FIG. 11B and FIG. 11C are used when the edge determination results are “10” and “01”, respectively, and correspond to the noise reduction pixel gamma table with the background removal level th_s of 30 and 20, respectively.
FIG. 11D is used when the edge determination result is “00” and corresponds to the noise reduction pixel gamma table with the background removal level th_s being 0. This is the case in the first and second embodiments. This is the same as the non-noise-reducing pixel gamma table described in the above. That is, the edge determination result “00” corresponds to the case of No in step S105 of FIG.

Next, FIG. 12 shows the procedure of the mosquito noise reduction process of the third embodiment described above.
This procedure corresponds to the procedure of FIG. 10 described in the second embodiment, and is different from the process of FIG. 10 in that steps S131 and S132 are executed instead of steps S105 to S122.

That is, after step S104, the image processing unit 113 uses the method described with reference to Table 2 according to the edge amount E of the pixel having the maximum edge amount in the edge detection region, as illustrated in FIGS. 11A to 11D. One of the gamma tables is selected (S131). Then, the image processing unit 113 determines to apply the gamma table selected in step S131 as the background removal parameter for the pixel to be processed (S132), and proceeds to step S108.
With the above processing, the mosquito noise reduction processing described with reference to Table 2 and FIGS. 11A to 11D can be executed. In the case of No in step S103, the non-noise reduction pixel gamma table is applied as in the case of the second embodiment.

[Fourth Embodiment: FIGS. 13 and 14]
Next, a description will be given of a fourth embodiment of the image processing apparatus according to the present invention.
In the fourth embodiment, the basic configuration is the same as that of the third embodiment. When determining the background removal parameter, the pixels constituting the edge element are specified from among the pixels in the edge detection region, The difference is that the background removal parameter is determined based on the median edge amount obtained for the pixel. Only matters related to the differences will be described below. Moreover, the same code | symbol is used about the structure which is common or respond | corresponds with 3rd Embodiment.

  In the first to third embodiments so far, the background removal parameter applied to each pixel is based on the edge amount E of the pixel having the maximum edge amount among the pixels in the edge detection region Q1 including the pixel P1 to be processed. It was decided. However, in the case of an image acquired by a camera, a scanner, or the like, noise may occur during imaging or reading. For this reason, due to the influence of the noise, even in the case of a plurality of pixels constituting the same character, the pixel values of the pixels are different, and accordingly, the pixels in the vicinity of the same edge portion are also calculated. Edge amount may be different.

As an example, FIG. 13 shows an example of the edge amount obtained for each pixel in the edge detection region Q2 including the pixel P2 to be processed.
In the example of FIG. 13, it is assumed that the range for performing the convolution operation with the filter matrix is set across the white background portion and the character portion in the hatched region on the left side of the broken line, and the calculation result of the edge amount exceeds the threshold value. . Such a pixel whose edge amount exceeds a threshold value is referred to as an edge pixel.

  However, even within the pixels in the edge portion, the edge amount varies depending on the position, and the pixel P2 ′ picks up noise and has an extremely large edge amount value than the pixels in the same column. In this state, when the edge amount E of the pixel having the maximum edge amount is obtained, “230” of the pixel pixel P2 ′ is obtained. This is an appropriate feature amount indicating the sharpness of the edge portion in the edge detection region Q2. Not so.

Therefore, if the median value of each pixel in the edge portion is taken, the edge amount that is extremely large due to noise is eliminated, and the feature amount indicating the sharpness of the edge portion in the edge detection region Q2 is obtained. Appropriate values can be obtained. In the example of FIG. 13, “132.5” that is an average value of the fifth “85” and the sixth “180” from the one with the larger edge amount value among the ten pixels of the edge portion may be employed.
The reason why the median value is obtained only at the edge portion is that the steepness of the edge portion is hardly reflected in the calculation result when the value of the portion without the edge is taken into consideration.
In the fourth embodiment, by determining the background removal parameter to be applied to each pixel based on the median value as described above, it is possible to remove mosquito noise without being affected by noise generated during image shooting or reading. It can be carried out.

Next, FIG. 14 shows the procedure of the mosquito noise reduction process of the fourth embodiment described above.
This procedure corresponds to the procedure of FIG. 12 described in the third embodiment, and is different from the process of FIG. 12 in that steps S141 to S143 are executed instead of steps S131 and S132.

That is, after step S104, the image processing unit 113 identifies a pixel whose edge amount exceeds the reference value among the pixels in the edge detection region as a pixel constituting the edge element (S141). Next, the image processing unit 113 uses the method described with reference to Table 2 in the third embodiment according to the median value of the edge amount of each pixel constituting the edge element, and the gamma shown in FIGS. 11A to 11D. One of the tables is selected (S142).
Then, the image processing unit 113 determines to apply the gamma table selected in step S142 as the background removal parameter for the pixel to be processed (S143), and proceeds to step S108.
With the above processing, the mosquito noise reduction processing described with reference to FIG. 13 can be executed. In the case of No in step S103, the non-noise reduction pixel gamma table is applied as in the case of the third embodiment.
In step S142, as in the case of the second embodiment, a noise reduction pixel gamma table having a background removal level th_s using the median edge amount may be generated. However, in this case, if there is no pixel whose edge amount exceeds the reference value, the process proceeds to step S107 to determine to apply the non-noise reduction pixel gamma table.

[First Modification: FIGS. 15 and 16]
Next, a first modification of each embodiment described above will be described.
The first modification can be applied to any of the first to fourth embodiments described above, and the size of the edge detection region is changed according to the compression method used for compressing the compressed image data to be processed. It is what I did. Only the matters related to this change will be described below. Moreover, the same code | symbol is used about the structure which is common or respond | corresponds with 1st Embodiment.

  In general, compression processing of image data is performed in units of blocks including a predetermined number of pixels. Therefore, mosquito noise is also generated by being affected by the edge elements of the pixels in the block. Therefore, by using the compression processing block as an edge detection region in accordance with the compression method used for lossy compression, mosquito noise can be reduced more efficiently.

The block size varies depending on the compression method as shown in Table 3. Since the compression method is known when the decompression process is performed, in the first modification, the decoder unit 202 transmits the information to the mosquito noise reduction processing unit 203.

Here, FIG. 15 shows a functional configuration of the mosquito noise reduction processing unit 203 in the first modification.
In the first modification, the mosquito noise reduction processing unit 203 includes an edge detection region setting unit 306 in addition to the configuration of FIG. The edge detection area setting unit 306 has a function of acquiring the size of the edge detection area to be set in the decompressed image data to be processed from the compression method information supplied from the decoder unit 202 and Table 3. A function of determining the position of the edge detection area provided in the decompressed image data to be processed according to the acquired size and supplying the position to the background removal parameter determination unit 304 is also provided.
The background removal parameter determination unit 304 executes the processing after step S104 in FIG. 7 and the like according to the information on the position of the edge detection region supplied from the edge detection region setting unit 306.

FIG. 16 shows an arrangement example of the edge detection areas in the first modification.
FIG. 16 shows an example in which image data compressed by the JPEG (Joint Photographic Experts Group) method is expanded. In this case, as shown in Table 3, since the block unit of the compression process is 8 × 8 pixels, the size of the edge detection area is also set to 8 × 8 pixels accordingly. Therefore, four edge detection areas B1 to B4 are arranged in the range of 16 × 16 pixels shown in FIG. Arrangement is performed in the same manner as the arrangement of blocks in the compression process.
For example, when the pixel P3 is a processing target, the edge amount analysis in step S105 of FIG. 7 is performed on the edge detection region B4 including the pixel P3. Even if the pixel of interest moves, analysis is performed using the same edge detection region B4 while it is in the edge detection region B4. The pixel of interest does not need to be in the center of the edge detection area.

  In the first modified example described above, the edge detection area setting unit 306 determines the size of the edge detection area based on the compression method of the compressed image data, and thereby the range of the image that causes noise in each pixel. The value of the parameter used for the background removal process can be determined according to the edge amount obtained for. As a result, mosquito noise can be accurately reduced.

[Second Modification: FIGS. 17 and 18]
Next, a second modification of each embodiment described above will be described.
The first modification can be applied to any of the first to fourth embodiments described above, and the image processing apparatus is configured as a portable terminal device, and image data after mosquito noise reduction processing is formed on a sheet. Rather than being used for image display on the screen. Only the matters related to this change will be described below. Moreover, the same code | symbol is used about the structure which is common or respond | corresponds with 1st Embodiment.

FIG. 17 shows a hardware configuration of the mobile terminal device in the second modification.
As illustrated in FIG. 17, the mobile terminal device 50 includes a display unit I / F 121 in addition to the configuration having the same meaning as the CPU 102 to the system bus 109 illustrated in FIG. 1 (specific performance and structure may be different). The display unit 122 is connected to the system bus 109 via the interface. The display unit 122 is a display unit that displays an image on a screen based on image data, such as a liquid crystal display.

  Such a portable terminal device 50 can be configured as, for example, a smartphone, a tablet computer, or a laptop computer. Further, in addition to the configuration shown in FIG. 17, it can be configured as an arbitrary apparatus having an image processing function, such as providing an imaging unit and configuring as a digital camera.

FIG. 18 shows a schematic configuration of functions provided in the mobile terminal device 50 for displaying an image on a screen based on compressed image data.
Among the functions shown in FIG. 18, the functions of the compressed image acquisition unit 201 to the other processing unit 204 correspond to the functions of the same name shown in FIG. 2. However, since the mobile terminal device 50 does not include the image processing unit 113, the functions of the compressed image acquisition unit 201 to the other processing unit 204 including the functions of the mosquito noise reduction processing unit 203 and the other processing unit 204 are mainly performed by the CPU 102. Is realized by executing a required program.

  The screen display unit 206 has a function of displaying an image on the display unit 122 based on the image data received from the other processing unit 204. In response to this, the other processing unit 204 performs a correction process for processing the decompressed image data passed from the decoder unit 202 or the mosquito noise reduction processing unit 203 into image data suitable for display, and performs a screen display unit. The function to pass to 206 is provided.

According to the above mobile terminal device 50, even when image data subjected to irreversible compression is decompressed and used for display, mosquito noise is effectively reduced as in the case of image formation in each of the embodiments described above. A high-quality image can be displayed.
Of course, it is not necessary that the device for displaying an image is portable, and it may be a device capable of performing image formation and display selectively or in parallel. The present invention can also be applied to a stationary PC.

[Other variations]
This is the end of the description of the embodiment of the present invention. In this invention, the specific configuration of the apparatus, the specific processing procedure, the calculation formula used for the processing, the size of each area, the compression method, etc. It is not limited to the one described in.
For example, an example using a gamma table as a background removal parameter has been described. However, the present invention is not limited to this, and a threshold value is set and the threshold value is used as a parameter. In this case, it may be clipped with 0.
The background removal process is not limited to the gamma correction process.

  In addition, the functions of the image forming apparatus 10 of each of the above-described embodiments (hereinafter, the same applies to the portable terminal device 50 unless otherwise specified) are distributed to a plurality of apparatuses, and image formation is performed in cooperation with these apparatuses. You may make it implement | achieve the function similar to the apparatus 10. FIG. In particular, the function of the mosquito noise reduction processing unit 203 is provided in an image processing apparatus external to the image forming apparatus 10, and the decompressed image data is transmitted from the image forming apparatus 10 to the image processing apparatus as necessary, so that the mosquito is reduced. Noise reduction processing may be performed. Further, the image forming apparatus 10 may receive image data after mosquito noise reduction processing from the image processing apparatus and use it for image formation. Such an external image processing apparatus is also an embodiment of the image processing apparatus of the present invention.

  Further, the program of the present invention is a program for causing a computer to control required hardware to realize the functions provided in the image forming apparatus 10 described above, particularly the functions shown in FIG. Such a program may be stored not only in the ROM 104 but also in an external storage medium. Further, it can be provided by being recorded on an arbitrary nonvolatile recording medium such as a memory card, CD, DVD, or Blu-ray disc. Each procedure described above can be executed by installing the program recorded in the recording medium in another image forming apparatus and executing the program.

It is also possible to download the program from an external device connected to a network and provided with a recording medium that records the program or from an external device that stores the program in a storage means, and install and execute the program on a computer.
Moreover, it is needless to say that the configurations of the embodiment and the modified examples described above can be arbitrarily combined and implemented as long as they do not contradict each other.

10: image forming apparatus, 50: portable terminal device, 101: controller, 102: CPU, 103: RAM, 104: ROM, 105: HDD, 106: communication I / F, 107: operation unit I / F, 108: operation 109: System bus 110: Engine control unit 111: Image reading unit 112: Plotter unit 113: Image processing unit 121: Display unit I / F 122: Display unit 201: Compressed image acquisition unit 202: Decoder unit, 203: Mosquito noise reduction processing unit, 204: Other processing unit, 205: Print execution unit, 206: Screen display unit, 301: Expanded image reception unit, 302: White background detection unit, 303: Edge detection unit, 304: Background removal parameter determination unit, 305: Background removal processing unit, 306: Edge detection region setting unit

JP 2006-352279 A

Claims (9)

An image processing apparatus that processes decompressed image data obtained by decompressing compressed image data,
A white background detecting means for detecting a white background image area in the decompressed image data;
Edge detection means for detecting an edge portion in the decompressed image data;
A determination unit that determines a value of a background removal parameter to be applied to each region of the decompressed image data based on the detection result of the white background detection unit and the edge detection unit;
An image processing apparatus comprising: processing means for performing background removal processing on the decompressed image data using the value of the background removal parameter determined by the determination means. The image processing apparatus according to claim 1,
The white background detecting means detects whether the pixel is in a white background image area for each pixel,
The edge detection means detects whether or not there is a pixel constituting an edge portion in an edge detection region composed of a plurality of pixels,
The determining unit determines a value of a background removal parameter to be applied to each pixel of the decompressed image data, and determines whether or not there is a pixel constituting an edge portion in an edge detection region including a certain pixel. An image processing apparatus that determines a value of a background removal parameter to be applied to the image processing apparatus. The image processing apparatus according to claim 2,
The background removal process is a process of changing the gradation value of each pixel according to the value of the background removal parameter,
When there is a pixel constituting an edge portion in an edge detection region including the certain pixel, the determining unit determines a background removal parameter value to be applied to the pixel from a predetermined threshold value in a gradation value in the decompressed image data. An image processing apparatus characterized in that, for a pixel on the white side, the gradation value is determined to be a value that defines a change to a white side value. The image processing apparatus according to claim 3,
The edge detection means includes means for obtaining a feature amount indicating the steepness of the edge portion in the edge detection region,
The image processing apparatus according to claim 1, wherein the determination unit determines the threshold based on the feature amount obtained for an edge detection region including the certain pixel. The image processing apparatus according to claim 4,
The image processing apparatus according to claim 1, wherein the determination unit sets the threshold value to a value farther from the white side as the feature amount obtained for the edge detection region including the certain pixel is larger. The image processing apparatus according to claim 3,
The edge detection means includes means for obtaining a feature amount indicating a steepness of a variation degree of a pixel value around each pixel in an edge detection area including the certain image,
The determination means determines the threshold based on a median value of feature values exceeding a predetermined reference value among the feature values of each pixel in an edge detection region including the certain image. . The image processing apparatus according to any one of claims 2 to 6,
An image processing apparatus comprising: means for determining a size of the edge detection area based on a compression method of the compressed image data. An image processing method for processing decompressed image data obtained by decompressing compressed image data,
A white background detection procedure for detecting a white background image region in the decompressed image data;
An edge detection procedure for detecting pixels constituting an edge portion in the decompressed image data;
A determination procedure for determining a value of a background removal parameter to be applied to each region of the decompressed image data based on the detection result of the white background detection procedure and the edge detection procedure;
A processing procedure for performing a background removal process on the decompressed image data using the value of the background removal parameter determined by the determination procedure. A program for causing a computer to realize a function of processing decompressed image data obtained by decompressing compressed image data,
The computer,
A white background detecting means for detecting a white background image area in the decompressed image data;
Edge detection means for detecting pixels constituting an edge portion in the decompressed image data;
A determination unit that determines a value of a background removal parameter to be applied to each region of the decompressed image data based on the detection result of the white background detection unit and the edge detection unit;
A program for functioning as processing means for performing background removal processing on the decompressed image data using the value of the background removal parameter determined by the determination means.