JP2015108960A - Image processor, control method thereof and program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To generate a suitable noise reduction image by synthesizing different noise with image data subjected to noise reduction processing.SOLUTION: An image processor includes smoothing processing means for performing smoothing processing of each pixel of input image data, derivation means for deriving the edge intensity of each pixel of the smoothed image data, and synthesis means for synthesizing noise data corresponding to the edge intensity with the smoothed image data to generate output image data.

Description

  The present invention relates to an image processing apparatus that smoothes image data and a control method thereof.

  Digital imaging devices such as digital still cameras and digital video cameras are widely spread and are generally used. These digital imaging devices generate digital image data by converting light received by a photoelectric charge conversion device (imaging device) such as a CCD or CMOS sensor into a digital signal.

  In the process of generating digital image data, dark current noise, thermal noise, shot noise, and the like are generated due to the characteristics of the imaging device and circuit, and the noise is mixed into the digital image data. With recent miniaturization of image sensors and higher pixel counts, the pixel pitch has been minimized, making noise more noticeable. In particular, when the photographing sensitivity is increased, noise is noticeably generated in the image data, which is a major cause of image quality deterioration. Therefore, in order to obtain a high-quality image, it is necessary to reduce mixed noise.

  2. Description of the Related Art Conventionally, a technique for reducing noise by applying a low-pass filter that passes a signal component below a noise frequency and smoothing is known. However, this method has a problem that the edge portion in the image data may be blurred. On the other hand, it is known that an image including noise is perceived to have a higher resolution than an image including no noise. Patent Document 1 proposes a method for visually enhancing resolution by synthesizing noise with image data.

JP 2008-187260 A

  When smoothing processing is applied to a noisy image such as an image taken with high sensitivity, the edge portion and the texture region are blurred, resulting in a decrease in resolution. There is also a case where a lump of noise remains (so-called residual noise). In particular, residual noise is likely to be perceived in flat areas and gradation areas. Therefore, the method disclosed in Patent Document 1 can enhance the resolution of the edge region, but cannot solve the problem of residual noise.

  Therefore, an object of the present invention is to reduce the blur of the edge portion due to the smoothing process by appropriately combining noise with the smoothed image data.

  In order to solve the above-mentioned problems, the present invention provides a smoothing processing means for smoothing each pixel of input image data, a derivation means for deriving an edge strength of each pixel of the smoothed image data, Combining means for generating output image data by combining noise data corresponding to the edge intensity with the smoothed image data.

  According to the present invention, it is possible to reduce blur of the edge portion due to the smoothing process by appropriately combining noise with the smoothed image data.

Block diagram showing hardware configuration of image processing apparatus 1 is a block diagram showing a logical configuration of an image processing apparatus according to a first embodiment. Flowchart diagram showing the flow of image processing The flowchart figure which shows the flow of the noise data derivation method of 1st Embodiment. The flowchart figure which shows the flow of the noise synthetic | combination process of 1st Embodiment. The block diagram which shows the logic structure of the image processing apparatus of 2nd Embodiment. The flowchart figure which shows the flow of the noise synthesis process of 2nd Embodiment. The block diagram which shows the logic structure of the image processing apparatus of 3rd Embodiment. The flowchart figure which shows the flow of the noise synthesis process of 3rd Embodiment.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings. In addition, the following embodiment shows an example of this invention and this invention is not limited to the following embodiment. In addition, all the features described in the embodiments are not necessarily essential to the present invention.

<First Embodiment>
FIG. 1 is a block diagram illustrating a hardware configuration of an image processing apparatus applicable to the present embodiment. Here, a general personal computer (PC) will be described as an example of the image processing apparatus, but the present invention is not limited to this. It may be a portable information processing device such as a tablet device or a smart phone, or may be configured such that the imaging device has built-in hardware capable of executing the image processing of this embodiment, such as a digital camera or a camera-equipped mobile phone. . The image processing apparatus according to the present embodiment includes a CPU 101, a RAM 102, an HDD 103, a general-purpose interface (I / F) 104, a monitor 108, and a main bus 109. The general-purpose I / F 104 is an interface for connecting the image processing apparatus and an external apparatus. The external devices are an imaging device 105 such as a camera, an input device 106 such as a mouse and a keyboard, and an external memory 107 such as a memory card. The general-purpose I / F 104 connects these external devices to the main bus 109. The general-purpose I / F 104 may be configured to exchange data with a communication device using infrared communication, wireless LAN, or the like.

  The CPU 101 controls each unit of the image processing apparatus in accordance with an input signal and a program described later. The HDD 103 stores computer programs for the CPU 101 to execute various software. The CPU 101 activates software (image processing application) stored in the HDD 103 to realize various image processing in the image processing apparatus.

  The RAM 102 is used as an area where the image data to be processed is temporarily stored or the CPU 101 develops work. The monitor 108 displays image data to be processed and a user interface (UI) for executing image processing. As described above, various data stored in the HDD 103 and the external memory 107, images captured by the imaging device 105, instructions from the input device 106, and the like are transferred to the RAM 102. Furthermore, according to the processing in the image processing application, the data stored in the RAM 102 performs various calculations based on commands from the CPU 101. The calculation result is displayed on the monitor 108 or stored in the HDD 103 or the external memory 107. Note that image data stored in the HDD 103 or the external memory 107 may be transferred to the RAM 102. Further, image data transmitted from the server via a network (not shown) may be transferred to the RAM 102.

  In the following, various processes realized by the CPU 101 operating various software (computer programs) stored in the HDD 103 will be described. In the above configuration, details of processing for generating and outputting noise-reduced image data by inputting image data to the image processing application based on a command from the CPU 101 will be described.

(Logical configuration of image processing device)
FIG. 2 is a block diagram illustrating a logical configuration of an image processing apparatus applicable to the present embodiment. The image processing apparatus according to the present embodiment includes a smoothing processing unit 201, an image-dependent noise generation unit 202, an edge strength derivation unit 203, and a noise synthesis unit 204. The smoothing processing unit 201 acquires input image data based on an instruction from the CPU 101, applies noise reduction processing by smoothing to the input image data, and generates smoothed image data. A conventional smoothing filter may be used for the smoothing process. For example, a desired method such as an average value filter, an ε filter, or a bilateral filter can be used. Input image data is input from the imaging device 105, HDD 103, or external memory 107. Of course, the image data obtained by photographing with the imaging device 105 may be temporarily stored in a storage device such as the HDD 103 and then input. The generated smoothed image data is stored in the RAM 102 or HDD 103.

  The image-dependent noise generation unit 202 acquires input image data and smoothed image data based on an instruction from the CPU 101, and generates image-dependent noise data from the input image data. Input image data and smoothed image data may be acquired from the RAM 102 or the HDD 103. The image dependent noise data means noise data generated based on input image data. Here, the high frequency component of the input image data is generated as image dependent data. It can be said that the smoothed image data obtained from the smoothing processing unit 201 is a low-frequency component of the input image data. Therefore, the image-dependent noise generation unit 202 generates a high-frequency component of the input image data by subtracting the smoothed image data for each pixel from the input image data. The generated image dependent noise data is stored in the RAM 102 or the HDD 103.

  The edge strength deriving unit 203 acquires smoothed image data based on an instruction from the CPU 101, and derives the edge strength for each pixel of the smoothed image data. A conventionally known method may be used for the edge strength. For example, a general process such as a Roberts filter, a Sobel filter, a Prewitt filter, or a Laplacian filter can be used. In this embodiment, a Laplacian filter is used. The edge strength derived for each pixel indicates that the greater the value, the stronger the edge degree. Here, a pixel having an edge strength larger than the threshold Eth can be determined as an edge portion.

  The noise synthesis unit 204 acquires smoothed image data, edge strength, image-dependent noise data, and image-independent noise data (image-independent noise data) based on an instruction from the CPU 101. The image-independent noise data is input from the HDD 103 or the external memory 107. Alternatively, the image may be input from the input device 106 such as a keyboard from the UI of the image processing application displayed on the monitor 103. The noise synthesis unit 204 synthesizes noise for each pixel with respect to the smoothed image data according to the edge strength. Specifically, when it is determined that the target pixel is an edge portion, the pixel value at the target pixel position in the image-dependent noise data is added. When it is determined that the target pixel is a non-edge portion, pixel values at the same pixel position as the target pixel in the image-independent noise data are added. The noise synthesizing unit 204 adds the image-dependent noise or the image-independent noise for each pixel according to the edge strength, and generates output image data. The output image data after the noise synthesis is output to the monitor 108, the HDD 103, etc. based on an instruction from the CPU 101. The output destination is not limited to this. For example, the output destination may be output to the external memory 107 connected to the general-purpose I / F 104, or may be output by connecting a printer or the like.

(Image processing flow)
In the following, the flow of image processing in which image data is input and noise reduction processing is applied in the logical configuration of the image processing apparatus described in FIG. 2 will be described. FIG. 3 shows a flowchart of processing in the image processing apparatus executed by the CPU 101.

  In step S301, the image processing apparatus inputs input image data to be processed.

  In step S302, the smoothing processing unit 201 applies a smoothing process to the input image data input in step S301, and generates smoothed image data. Here, an average value filter is used.

  In step S303, a plurality of types of noise data to be combined with the smoothed image data generated in step S302 are derived. The noise data derived here is at least two different noise data. In the present embodiment, description will be made assuming that the image-dependent noise data generated by the image-dependent noise generation unit 202 and Gaussian noise data as image-independent noise data prepared in advance. The detailed flow of the noise data derivation process will be described later.

  In step S304, the edge strength deriving unit 203 calculates the edge strength of each pixel in the smoothed image data. As described above, the Laplacian filter is used here.

  In step S305, the noise synthesis unit 204 synthesizes, for each pixel, the noise data derived in step S304 according to the edge strength with respect to the smoothed image generated in step S302. A detailed flow of the synthesis process will be described later.

  In step S306, the noise synthesis unit 204 outputs the synthesized image data generated in step S305.

(Noise data derivation process)
Hereinafter, the details of the noise data deriving process in step S303 described in FIG. 3 will be described with reference to the flowchart in FIG.

  In step S401, image-independent noise data is acquired by inputting Gaussian noise parameters. Parameters can be set arbitrarily. For example, an average A and variance V of Gaussian noise are set. Note that the noise data input here is not limited to Gaussian noise as long as it is noise data that can be determined without depending on the input image, and may be noise with a uniform distribution, for example. Further, since the amount of noise changes according to the luminance, the variance may be set to change according to the average value.

  In step S402, the image data input in step S301 is acquired.

  In step S403, the smoothed image data generated in step S302 is acquired.

  In step S404, the image-dependent noise generation unit 202 generates image-dependent noise data from the input image data and the smoothed image data. Here, the image dependent noise data is a high frequency component of the input image data. As described above, the image-dependent noise generation unit 202 obtains a high-frequency component of the input image data by taking the difference between the input image data and the smoothed image data. When the above processing is completed, the derivation of the image dependent noise data and the image independent noise data is completed.

(Detailed flow of noise synthesis processing)
Hereinafter, details of the noise synthesis processing in step S305 described with reference to FIG. 3 will be described with reference to the flowchart of FIG.

  In step S501, the noise synthesis unit 204 acquires the smoothed image data generated by the smoothing processing unit 201.

  In step S502, a plurality of different noise data derived in step S303 is acquired. Specifically, image-dependent noise data and image-independent noise data.

  In step S503, a pixel to be processed (hereinafter referred to as a target pixel) is selected for the smoothed image data acquired in step S501.

  In step S504, the edge strength E at the target pixel selected in step S503 is acquired.

  In step S505, the edge intensity E of the target pixel calculated in step S504 is compared with a predetermined threshold Eth. When the edge intensity E is smaller than the threshold Eth, that is, when the target pixel is a non-edge portion and is a pixel constituting a flat region, the process proceeds to step S506 to synthesize image-independent noise data. On the other hand, when the edge strength E is larger than the threshold Eth, that is, when the target pixel is a pixel in the edge portion constituting the edge region or the texture region, the process proceeds to step S507 to synthesize image-dependent noise data.

  In step S506, the image-independent noise input in step S502, that is, Gaussian noise, is combined with the target pixel selected in step S503. Here, the pixel value of the target pixel is x0, the average is 0, Gaussian noise with variance V is g (0, v), and the coefficient for controlling the magnitude of the synthesized noise is k (0.0 <= k <= 1. 0), the pixel value x1 after noise synthesis is expressed by the following equation.

  If the pixel of interest is a pixel constituting a flat region, residual noise tends to be noticeable as a result of the noise reduction processing by smoothing in step S302. Therefore, by combining image-independent noise, residual noise that is image-dependent noise can be made inconspicuous.

In step S507, the image-dependent noise input in step S502, that is, the high-frequency component of the input image data is combined with the target pixel selected in step S503. Here, the pixel value of the pixel of interest is x ₀ , the high-frequency component of the pixel of the input image data corresponding to the pixel of interest is h, and the coefficient for controlling the magnitude of the noise to be combined is k (0.0 <= k <= 1). When .0), the pixel value x ₂ after the noise synthesis becomes the following equation.

  When the pixel of interest is a pixel constituting an edge region or a texture region, a decrease in resolution tends to be a problem as a result of the noise reduction processing by smoothing in step S302. Therefore, by combining image-dependent noise including components of the edge region and the texture region, it is possible to obtain an effect of improving the resolution.

  In step S508, it is determined whether the noise synthesis process has been completed for all the pixels of the smoothed image input in step S501. If completed, step S305 is terminated, and if not completed, the process proceeds to step S303 and the processing is continued.

  As a result of the above processing, image-dependent noise is combined with an image having a high edge strength and image-independent noise is combined with a pixel having a low edge strength in the image data subjected to noise reduction processing by smoothing. As a result, it is possible to suppress a problem caused by a general noise reduction process and generate a suitable noise reduced image.

  In the present embodiment, the high-frequency component of the input image data that becomes the image-dependent image data is calculated from the difference between the input image data and the smoothed image data. However, other modifications can be applied depending on the purpose. For example, a result obtained by applying a high-pass filter or a differential filter to input image data may be used. It is also possible to derive by applying Fourier transform to the input image data and applying inverse Fourier transform to the high frequency component in the frequency space.

Second Embodiment
In the first embodiment described above, the method of synthesizing the image dependent noise data or the image independent noise data based on the edge intensity of each pixel of the smoothed image data has been described. In this method, two different types of noise data are switched. In the second embodiment, a method of synthesizing noise in which noise data is continuously changed, instead of switching noise data to be synthesized, using edge strength that is a continuous value will be described.

(Logical configuration of image processing device)
FIG. 6 is a schematic diagram showing a logical configuration of an image processing apparatus applicable to the second embodiment. The image processing apparatus according to the second embodiment includes a smoothing processing unit 201, an edge strength derivation unit 203, and a noise synthesis unit 601. Since the smoothing processing unit 201 and the edge strength deriving unit 203 have the same configuration as in the first embodiment, description thereof is omitted.

The noise synthesis unit 601 acquires smoothed image data, edge strength, high amplitude noise data, and low amplitude noise data based on an instruction from the CPU 101. The noise synthesis unit 601 synthesizes noise for each pixel according to the edge strength with respect to the smoothed image data. High amplitude noise data and low amplitude noise data are input from the HDD 103 or the external memory 107. Alternatively, the image may be input from the input device 106 such as a keyboard from the UI of the image processing application displayed on the monitor 103. The edge synthesis unit 601 synthesizes noise having different amplitudes with the smoothed image data. Specifically, when the edge strength is a pixel, that is, when the target pixel is a pixel constituting an edge region or a texture region, high-frequency noise is synthesized. This is to improve the resolution of edge regions and texture regions that are easily blurred by noise reduction processing. On the other hand, if the pixel with low edge strength, that is, the pixel of interest is a pixel that forms a flat region, by combining low-amplitude noise, noise unevenness with the pixel combined with high-amplitude noise can be reduced while retaining the noise reduction effect. It is for suppressing. Here, the pixel value of the pixel of interest is x ₀ , the maximum edge strength is E _max , the minimum edge strength is E _min , and the coefficient that controls the magnitude of the synthesized noise is k (0.0 <= k < = 1.0) and when the pixel value _{x 3} after noise synthesis becomes equation (3).

  The image data after the noise synthesis is output to the monitor 108, the HDD 103, or the like based on an instruction from the CPU 101. The output destination is not limited to this. For example, the output destination may be output to the external memory 107 connected to the general-purpose I / F 104, or may be output by connecting a printer or the like.

(Image processing flow)
In the following, details of image processing in which image data is input and noise reduction processing is applied in the logical configuration of the image processing apparatus described in FIG. 6 will be described with reference to the flowchart in FIG. 3. Note that description of the same processing as in the first embodiment is omitted.

In step S303, the noise synthesis unit 601 acquires high amplitude noise data and low amplitude noise data. The noise data acquired here may be any noise data as long as the amplitude data can be continuously changed according to the edge strength of the image data. For example, Gaussian noise or uniform distribution noise can be used. In the present embodiment, description will be made assuming that Gaussian noise data having an average value 0 and a variance v having different amplitudes is acquired. Here, constants for controlling the magnitude of noise are A ₁ and A ₂ (0.0 <= A ₁ <= 1.0, 0.0 <= A ₂ <= 1.0, A ₁ > A ₂ ). Then, the high amplitude noise N1 and the low amplitude noise N2 are respectively expressed by the following equations.

In step S305, the noise synthesis unit 601 synthesizes the noise data derived in step S303 with the smoothed image generated in step S302. (Detailed flow of composition processing)
Hereinafter, details of the noise synthesis processing in step S305 described with reference to FIG. 3 will be described with reference to the flowchart of FIG. Note that the description of the same processing as in the first embodiment is omitted.

  Step S701 is the same as step S501. In step S702, the high amplitude noise data and the low amplitude noise data derived in step S303 are acquired.

  Step S703 is the same as step S503. Step S704 is the same as step S504. In step S705, the noise synthesizing unit 601 synthesizes noise having different amplitudes according to the edge strength for each pixel with respect to the smoothed image data using Expression (3).

  As in the first embodiment, noise whose amplitude is switched by a threshold value may be synthesized. In this embodiment, an example in which noise having different amplitudes is combined with respect to image-independent noise has been described. However, noise having different amplitudes may be combined with respect to image-dependent noise. Furthermore, using both image-dependent noise and image-independent noise, the greater the edge strength, the more image-dependent noise, the less image-independent noise, and the smaller the edge strength, the less image-dependent noise and the image-independent noise. You may synthesize | combine so that it may increase.

  Step S706 is the same as step S507.

  With the above processing, it is possible to synthesize high-amplitude noise for pixels with strong edge strength and low-amplitude noise for pixels with low edge strength in image data that has undergone noise reduction processing by smoothing. . In this way, by continuously changing and synthesizing noise with different noise characteristics such as amplitude according to the edge strength, it is possible to suppress the problem of reduced resolution due to general noise reduction processing. It is possible to generate a suitable noise-reduced image that suppresses the noise unevenness.

<Third Embodiment>
In the above-described embodiment, the method of synthesizing different noise data based on the edge intensity of each pixel of the smoothed image data has been described. In the present embodiment, a method of dividing image data in advance and synthesizing different noise for each region will be described. Detailed description of the same configuration and processing as those in the above-described embodiment will be omitted.

(Logical Configuration of Image Processing Apparatus of Embodiment 3)
FIG. 8 is a schematic diagram illustrating a logical configuration of the image processing apparatus according to the third embodiment. The image processing apparatus according to the present embodiment includes a smoothing processing unit 201, an image-dependent noise generation unit 202, a region division unit 801, and a noise synthesis unit 802. Since the smoothing processing unit 201 and the image-dependent noise generation unit 202 are the same as those in the first embodiment, description thereof is omitted.

  The area dividing unit 801 applies area dividing processing to the smoothed image data obtained from the smoothing processing unit 201 based on an instruction from the CPU 101 to generate an area divided image. There may be at least two areas to be divided, and the number of areas to be divided may be three or more. In the present embodiment, description will be made assuming that the image is divided into two areas, a flat area and a texture area. A method for dividing the area will be described later.

  The noise synthesis unit 802 acquires smoothed image data, region-divided images, image-dependent noise data, and image-independent noise data based on an instruction from the CPU 101, and adds noise to the smoothed image data. Synthesize. The image dependent noise data is acquired from the image dependent noise generation unit 202, and the image independent noise data is input from the HDD 103 or the external memory 107. Alternatively, the image may be input from the input device 106 such as a keyboard from the UI of the image processing application displayed on the monitor 103. Details of the noise synthesis method will be described later. The image data after the noise synthesis is output to the monitor 108, the HDD 103, or the like based on an instruction from the CPU 101. The output destination is not limited to this. For example, the output destination may be output to the external memory 107 connected to the general-purpose I / F 104, or may be output by connecting a printer or the like.

(Composition process flow)
Hereinafter, details of the noise synthesis processing in step S305 described with reference to FIG. 3 will be described with reference to the flowchart of FIG.

  Step S901 is the same as step S501.

  In step S902, the area dividing unit 801 divides the smoothed image data acquired in step S501 into areas and generates area divided image data. It should be noted that a desired method can be used to create the region-divided image data. For example, after obtaining the edge strength for each pixel by the same method as in step S504, the area is divided into a flat area and a texture area with a threshold, and the area to which the target pixel belongs is determined based on the majority of neighboring pixel areas including the target pixel. Update. By such a process, it is possible to suppress the occurrence of noise unevenness due to the presence of a texture region of only a few pixels in the flat region and the locally synthesized noise being changed. Of course, the region dividing method is not limited to this, and a desired method such as a region growing method or a division rule method may be used.

  In step S903, the noise data derived in step S303 is acquired.

  Step S904 is the same as step S504.

  In step S905, the region divided image generated in step S902 is referenced to determine the region to which the target pixel selected in step S904 belongs. If the target pixel belongs to the flat area, the process proceeds to step S906 to synthesize image-independent noise. If the target pixel belongs to the texture area, the process proceeds to step S907 to synthesize the image-dependent noise.

  Step S906 is the same as step S506.

  Step S907 is the same as step S507.

  Step S908 is the same as step S508.

  Through the above processing, image-dependent noise is synthesized in the flat area of the smoothed image data, and image-independent noise is synthesized in the texture area. Accordingly, it is possible to suppress a problem due to general noise reduction processing and generate a suitable noise-reduced image in which noise non-uniformity between pixels that may be generated by combining noises is suppressed.

<Other embodiments>
The present invention can also be realized by supplying a storage medium storing a program code of software for realizing the functions of the above-described embodiments to a system or apparatus. In this case, the function of the above-described embodiment is realized by the computer (or CPU or MPU) of the system or apparatus reading and executing the program code stored in the computer-readable storage medium.

Claims (10)

Smoothing processing means for smoothing each pixel of the input image data;
Derivation means for deriving an edge intensity of each pixel of the smoothed image data;
An image processing apparatus comprising: combining means for generating output image data by combining noise data corresponding to the edge intensity with the smoothed image data.   The noise data to be synthesized includes image-dependent noise data generated based on the input image data and image-independent noise data generated regardless of the input image data. The image processing apparatus according to 1.   The image processing apparatus according to claim 2, wherein the image-dependent noise data is a high-frequency component of the input image data.   The image processing apparatus according to claim 2, wherein the image-independent noise data is Gaussian noise.   When the edge intensity of the target pixel in the smoothed image data is larger than a predetermined threshold, the synthesizing unit converts the pixel value at the target pixel position in the image-dependent noise data in the smoothed image data. When the edge strength of the target pixel in the smoothed image data is smaller than the predetermined threshold value, the pixel value at the target pixel position in the image-independent noise data is added to the pixel value of the target pixel. 5. The image processing apparatus according to claim 2, wherein the image processing apparatus adds the pixel value of the target pixel in the converted image data.   The image processing apparatus according to claim 1, wherein the synthesizing unit acquires noise data having different amplitudes.   The image processing apparatus according to claim 6, wherein the synthesizing unit synthesizes corresponding pixel values in noise data with higher amplitude as the edge strength of the pixel of interest in the smoothed image data is higher. The synthesizing unit includes an area dividing unit that divides the smoothed image data into a plurality of areas,
If the area is a texture area, the corresponding pixel values in the image-dependent noise data are synthesized, and if the area is a flat area, the corresponding pixel values in the image-independent noise data are synthesized. The image processing apparatus according to claim 1, wherein the image processing apparatus is an image processing apparatus.   A program for causing a computer to operate as the image processing apparatus according to any one of claims 1 to 8. A control method for an image processing apparatus that includes a smoothing processing unit, a derivation unit, and a synthesis unit, and performs noise reduction processing on input image data,
The smoothing processing means smoothes each pixel of the input image data,
The derivation means derives an edge strength of each pixel of the smoothed image data;
The control method, wherein the synthesizing unit synthesizes noise data corresponding to the edge strength with the smoothed image data.