Noise reduction method, apparatus, electronic device, and computer-readable medium

Abstract

The present invention provides a noise reduction method, apparatus, electronic device, and computer-readable medium. The method includes the steps of: filtering a residual image of an image to generate a filtered image; enhancing weak edge regions in the filtered image and weakening strong edge regions in the filtered image to generate a sharpened image; and obtaining a target image based on a reconstructed image and the sharpened image, wherein the residual image is the difference value of the image at different resolutions, at least two residual images correspond to the image, there is a predetermined layer order relationship between the at least two residual images, and the reconstructed image is a superposition of all residual images whose layer order precedes the residual image. During the reconstruction superposition process, sharpening is performed to increase local contrast, enhance edges and details, eliminate the need for sharpening after image reconstruction, and reduce noise in strong edge regions and reduce image burrs by weakening strong edge regions.

Description

Technical Field

[0001] This application belongs to the field of image processing, and specifically relates to a noise reduction method, apparatus, electronic device, and computer-readable medium.

Background Art

[0002] In application scenarios of image sensors such as mobile phones and video surveillance, for example, CMOS Sensor (Complementary Metal-Oxide-Semiconductor Sensor), CCD Sensor (Charge-Coupled Device Sensor), etc., generally, it is necessary to perform spatial noise reduction on an image to increase the signal-to-noise ratio of the output image. However, after noise reduction, since the image is blurred and the resolution is relatively low, generally, it is necessary to increase the resolution of the image by a sharpening algorithm. Also, since common noise reduction algorithms cannot hold edges, the intensity of noise reduction in the strong edge part of the image is insufficient, and more noise remains in the strong edge part of the image, resulting in "aliasing" of the edges.

Summary of the Invention

[0003] In view of this, an object of this application is to provide a noise reduction method, apparatus, electronic device, and computer-readable medium that can solve the drawback that the resolution of a noise-reduced image generated by a conventional noise reduction method is inferior.

[0004] In the first aspect, the present application provides a noise reduction method. The method includes the steps of: filtering a residual image of an image to generate a filtered image; enhancing weak edge regions in the filtered image and weakening strong edge regions in the filtered image to generate a sharpened image; and obtaining a target image based on a reconstructed image and the sharpened image, wherein the residual image is the difference value of the image at different resolutions, at least two residual images correspond to the image, there is a predetermined layer order relationship between the at least two residual images, and the reconstructed image is a superposition of all residual images whose layer order precedes the residual image.

[0005] In the above embodiment, sharpening is performed during the reconstruction process of the image to be noise-reduced. This increases the local contrast of the image, emphasizing edges and details in the image. This eliminates the need for sharpening after image reconstruction, making the implementation process simpler and more convenient. Furthermore, by weakening strong edge regions, noise in strong edge regions can be reduced, and image burrs can be reduced.

[0006] In one selectable embodiment of the present application, the step of obtaining a target image based on a reconstructed image and the sharpened image includes the steps of establishing a guide image and a noise image based on the reconstructed image and the sharpened image, performing noise reduction on the noise image based on the guide image to obtain a noise-reduced image, and performing upsampling on the noise-reduced image to obtain the target image.

[0007] In the above embodiment, guide images and noise images can be established quickly and easily. Using guide images allows more weight to be assigned to strong edge regions, increasing the intensity of noise reduction for strong edges in the image. Since the noise image is a superposition of the sharpened residual image, the local contrast of the image is increased, edges and details in the image are emphasized, and the generated image becomes sharper and more vivid.

[0008] In one selectable embodiment of the present application, the method further includes the steps of: assigning values ​​to the filtered image based on a predetermined threshold range to obtain the weak edge region; and assigning values ​​to the filtered image based on a predetermined threshold to obtain the strong edge region. In the above embodiment, by assigning a predetermined threshold, strong edge regions and weak edge regions of the image can be obtained quickly and easily.

[0009] In one selectable embodiment of the present invention, the sharpened image includes a weak edge region image, and the step of enhancing the weak edge region in the filtered image includes enhancing the weak edge region using a predetermined first scalar amount, while leaving the rest of the filtered image other than the weak edge region to generate the weak edge region image.

[0010] In the above embodiment, the emphasis on weak edge regions can be quickly achieved by using a scalar quantity, while leaving the rest of the image excluding the weak edge regions intact. As a result, the generated image of the weak edge region has more image detail.

[0011] In one selectable embodiment of the present invention, the sharpened image includes a strong edge region image, and the step of weakening the strong edge region in the filtering image includes weakening the strong edge region using a predetermined second scalar amount, while leaving the rest of the filtering image other than the strong edge region to generate the strong edge region image.

[0012] In the above embodiment, the weakening of strong edge regions can be quickly achieved by using a scalar amount, while retaining the other parts outside the strong edge regions, thereby resulting in the generated image of the strong edge region having more image detail.

[0013] In one selectable embodiment of the present application, the sharpened image includes a weak edge region image and a strong edge region image, and the steps of establishing a guide image and a noise image based on the reconstructed image and the sharpened image include the steps of obtaining the noise image by adding the weak edge region image and the reconstructed image, and obtaining the guide image by adding the strong edge region image and the reconstructed image.

[0014] In one selectable embodiment of the present application, the step of generating a noise-reduced image by performing noise reduction on the noise image based on the guide image includes the steps of calculating bilateral filtering weight parameters based on the guide image, and generating the noise-reduced image by performing noise reduction on the noise image based on the weight parameters.

[0015] In the above embodiment, by calculating the weights of bilateral filtering using a guide image, noise reduction is performed on the noisy image, the intensity of noise reduction is increased for strong edges in the image, the burrs on strong edges of the image after noise reduction by the edge-preserving noise reduction algorithm are reduced, and sharpening is performed simultaneously with noise reduction, thereby emphasizing the details of the image.

[0016] In a second aspect, the present application further provides a noise reduction method. The method includes the steps of: acquiring a plurality of residual images of an image; generating a noise reduction image of the image by superimposing the plurality of residual images multiple times in a layer order, wherein each residual image is a difference value of the image at different resolutions, there is a predetermined layer order between the plurality of residual images, and at least one of the multiple superpositions is performed to weaken strong edge regions and strengthen weak edge regions in a filtering image corresponding to the residual image of the current layer to generate a sharpened image; and obtaining a reconstructed image of the current layer based on a reconstructed image of the layer above and the sharpened image.

[0017] In the above embodiment, during the process of superimposing the residual images of each layer of the image to be noise-reduced, sharpening is performed on the superimposed image. This increases the local contrast of the image, emphasizes edges and details in the image, eliminates the need for sharpening after image reconstruction, makes the implementation process simpler and more convenient, and reduces noise in strong edge regions by weakening strong edge regions, thereby reducing image burrs.

[0018] In one selectable embodiment of the present application, the step of generating a noise reduction image by superimposing multiple residual images in a layer order includes starting from a target residual image among the multiple residual images, performing multiple iterations in a layer order, iterating down to the residual image with the lowest layer rank, and making the reconstructed image of the final iteration the noise reduction image, wherein the target residual image refers to the residual image with the third layer rank, and each iteration includes the steps of: filtering the residual image of the current iteration to generate a filtered image of the current iteration; weakening strong edge regions and strengthening weak edge regions in the filtered image of the current iteration to generate a sharpened image of the current iteration; and obtaining a reconstructed image of the current iteration based on the reconstructed image of the previous iteration and the sharpened image of the current iteration. In the above embodiment, reconstruction can be rapidly performed on multiple residual images of the image using an iterative method, thereby obtaining an image with reduced noise.

[0019] In one selectable embodiment of the present application, a method for generating a reconstructed image of an iteration immediately preceding the first iteration of a plurality of iterations includes the steps of: performing a first filtering on a first residual image to generate a first filtered image; performing a second filtering on a second residual image to generate a second filtered image; weakening strong edge regions and enhancing weak edge regions in the second filtered image to generate a corresponding second sharpened image; and obtaining a reconstructed image of an iteration immediately preceding the first iteration of a plurality of iterations based on the first filtered image and the second sharpened image, wherein the first residual image is the residual image with the highest layer rank, the first filtering and the second filtering are different, and the second residual image is an image of the layer one layer below the first residual image.

[0020] In the third aspect, embodiments of the present application provide a noise reduction device. The device comprises a filtering module configured to filter a residual image of an image to generate a filtered image; a sharpening module configured to enhance weak edge regions in the filtered image and weaken strong edge regions in the filtered image to generate a sharpened image; and a generation module configured to obtain a target image based on a reconstructed image and the sharpened image, wherein the residual image is the difference value of the image at different resolutions, at least two residual images correspond to the image, there is a predetermined layer order relationship between the at least two residual images, and the reconstructed image is a superposition of all residual images whose layer order precedes the residual image.

[0021] In the fourth phase, an embodiment of the present application provides an electronic device. The electronic device comprises a memory and a processor, the processor and the memory being connected, the memory being configured to store a program, and the processor being configured to read the program stored in the memory and execute any one of the methods of the first phase.

[0022] In the fifth phase, an embodiment of the present application provides a computer-readable medium. A computer program is stored in the computer-readable medium, and when the computer program is executed by a processor, one of the methods of the first phase is executed.

[0023] Other features and advantages of this application will be described in the following sections. The purpose and other advantages of this application can be realized and obtained by the structure specifically shown in the specification and drawings. [Brief explanation of the drawing]

[0024] [Figure 1] This is a flowchart of the noise reduction method according to the embodiment of this application. [Figure 2]It is a flowchart of another noise reduction method according to an embodiment of the present application. [Figure 3] It is a schematic block diagram of a noise reduction device according to an embodiment of the present application. [Figure 4] It is a schematic block diagram of an electronic device according to an embodiment of the present application.

Modes for Carrying Out the Invention

[0025] Hereinafter, embodiments of the present application will be described with reference to the drawings. The same elements in different drawings are denoted by the same reference numerals. In the following description, the specific arrangements and detailed descriptions of the components are merely for a complete understanding of the embodiments of the present application. Therefore, as will be understood by those skilled in the art, various changes and modifications can be made to the embodiments described herein without departing from the scope of the present application. Also, for clarity and simplicity, descriptions of well-known functions and structures are omitted. The terms described below are defined based on the functions in the present application and may vary depending on the user, the user's intention, or custom, etc. Therefore, the definitions of the terms are determined based on the content of this specification. The present application may have various modifications and various embodiments. In the present application, hereinafter, embodiments will be specifically described with reference to the drawings. Note that the present application is not limited to the embodiments, and includes all changes, equivalents, and alternatives within the scope of the present application. The terms used in the present application are merely for explaining various embodiments of the present application and do not limit the present application. Unless otherwise specified, the singular form can also refer to the plural. In the present application, the term "comprising" or "having" represents that there is a feature, number, step, operation, component, member, or a combination thereof, but does not mean excluding the existence or addition of one or more other features, numbers, steps, operations, components, members, or a combination thereof. Unless otherwise specified, all terms used herein have the same meaning as those understood by those skilled in the art. In this application, unless otherwise specified, terms (e.g., terms defined in general dictionaries) have the same meaning as they appear in context within the relevant field, and not just their ideal or formal meaning.

[0026] An electronic device according to one embodiment may be one of each type of electronic device. Examples of electronic devices include portable communication devices (e.g., smartphones), computers, portable multimedia devices, portable medical devices, cameras, wearable devices, or consumer electronics. In the disclosed embodiment, the electronic device is not limited to the above examples.

[0027] The terminology used herein is not limiting to this application and includes various modifications, equivalents, or substitutes for applicable embodiments thereof. In the description of the drawings, similar reference numerals represent similar or related elements. Unless otherwise specified, the singular form of a noun may represent one or more. Each of the phrases used herein, for example, “A or B,” “A and at least one of B,” “A or at least one of B,” “A, B or C,” “A, B and at least one of C,” and “A, B or at least one of C,” may include all combinations of the listed items. The terms “first,” “second,” “primary,” and “secondary” used herein are for the sole purpose of distinguishing parts and do not limit parts in any other respect (e.g., importance or order). Where the terms “operable” or “communicable” are explained or not explained, when an element (e.g., a first element) is described as being “coupled” or “coupled to” another element (e.g., a second element), or being “connected” or “connected to” another element, it means that the element is coupled to the other element directly (e.g., wired), wirelessly, or via a third element.

[0028] As used herein, the term “module” may include units implemented by hardware, software, or firmware, and may be used interchangeably with other terms (e.g., “logic,” “logic block,” “component,” and “circuit”). A module may be a single integrated component or its smallest unit or component that performs one or more functions. For example, in one embodiment, a module may be implemented in the form of an application-specific integrated circuit (ASIC).

[0029] Figure 1 is a flowchart of a noise reduction method according to an embodiment of this application. As shown in Figure 1, the method includes steps S110 to S130.

[0030] S110: Filtering is performed on the residual image of the image to generate a filtered image, wherein the residual image is the difference value of the image at different resolutions, at least two residual images correspond to the image, and there is a predetermined layer order relationship between the at least two residual images.

[0031] A residual image is obtained from one of the images, and this residual image is any one residual image in a set of residual images corresponding to the image, and each residual image in the set of residual images is a difference value at a different resolution of the image, and there is a predetermined layer order relationship between the residual images in the set of residual images. For distinction, in this embodiment, the obtained residual image is called the first residual image.

[0032] For example, multiple residual images corresponding to the image to be noise-reduced can be generated using the Laplacian pyramid method.

[0033] The method for generating residual image sets using the Laplacian pyramid method includes the following:

[0034] 1. Construct a Gaussian pyramid. Assume that the original resolution of the image to be denoised is 64x64, and that this image will be the lowest layer (G-Level 0) of the Gaussian pyramid. Filtering is performed on this image, and as an example, Gaussian filtering is performed on the image. Gaussian filtering is a method of blurring an image, which can reduce high-frequency information in the image and is mainly used for denoising and smoothing images. For example, Gaussian filtering is achieved by performing convolution on the image to be denoised using a Gaussian kernel. Then, downsampling is performed on the Gaussian filtered image, and as an example, downsampling is performed using the nearest neighbor interpolation method, acquiring one pixel every other pixel in the image to be used as the pixel point of the new image, thereby reducing the original image to 32x32. This new image is constructed as the image of the layer one level above the lowest layer image in the Gaussian pyramid, i.e., the second level image (G-Level 1). Filtering and downsampling are performed on the second-level image (G-Level 1) to obtain a new layer image (G-Level 2), and this process is repeated until the image to be denoised is reduced to a specified size or until a predetermined number of layers are obtained. After this process is stopped, a Gaussian pyramid of the image to be denoised is obtained, and this Gaussian pyramid contains multiple resolution images arranged from bottom to top in descending order of resolution. In the embodiments of this application, the filtering method is not limited to Gaussian filtering and may be other filtering methods. Similarly, the method of downsampling in the Gaussian pyramid may be bilinear downsampling, area sampling, max pooling, etc., and the downsampling method is not limited in this application.

[0035] 2. Construct a Laplacian pyramid based on the obtained Gaussian pyramid. Each layer in the Laplacian pyramid is constructed by subtracting the upsampled Gaussian blurred image of the layer above from one layer of the Gaussian pyramid. The method for constructing the Laplacian pyramid starts with the second level image (G-Level1) of the Gaussian pyramid. Upsampling is performed on the second level image to obtain a 64x64 image. Gaussian blurring is performed on the upsampled image to obtain a blurred image. Subtracting this blurred image from the first level image (G-Level0) of the Gaussian pyramid yields the image of the lowest layer (L-Level0) of the Laplacian pyramid, which is the residual image. Upsampling and Gaussian blurring are performed on the third level image (G-Level2) of the Gaussian pyramid to generate a blurred image. Subtracting this blurred image from the second level image of the Gaussian pyramid yields the image of the second layer (L-Level1) of the Laplacian pyramid. Repeat the above process until you reach the last layer of the Gaussian pyramid, and when this process stops, you will obtain the Laplacian pyramid in the image.

[0036] Each image in the Laplacian pyramid is a residual image, and each residual image is a difference value obtained by subtracting noise reduction target images of different resolutions. All residual images in the Laplacian pyramid correspond to each noise reduction target image, and all residual images in the Laplacian pyramid have a predetermined layer order. Image reconstruction is achieved using the Laplacian pyramid. Starting with the smallest scale Laplacian pyramid image, this image is upsampled and added to the Laplacian pyramid image of the layer below it. The superimposed image is then upsampled and added to the Laplacian pyramid image of the layer below it, and this process is repeated in descending order of layer order until the lowest layer of the Laplacian pyramid is reached. After all residual images have been superimposed, reconstruction of the noise reduction target image is achieved. During the reconstruction process, noise reduction is performed on the image; for example, Gaussian filtering is performed before upsampling.

[0037] As an example, a Laplacian pyramid corresponding to the image to be noise-reduced is obtained, and an image of any one layer in the Laplacian pyramid is used as the first residual image, for example, the image of the third layer in the Laplacian pyramid is used as the first residual image.

[0038] Filtering is performed on the first residual image to generate a corresponding first filtered image. The filtering method may be bilateral filtering, Gaussian filtering, or the like, and the application is not limited to this method. In one embodiment of this application, Gaussian filtering is used, and filtering is performed on the first residual image using a 3*3 Gaussian kernel with a bandwidth of σ1. The purpose of convolution is to resist noise, and it is possible to prevent strong and weak edges in the residual image from being affected by noise.

[0039] S120: A sharpened image is generated by enhancing weak edge regions in the filtered image and weakening strong edge regions in the filtered image.

[0040] In image processing, edges are areas in an image where brightness changes are pronounced, and are typically the boundaries of objects or features. Edge strength generally refers to the magnitude of the change in pixel values ​​in this region of the image. Strong edges correspond to areas with large brightness changes and are generally clear, prominent boundaries or contours, representing the main contours of objects or features in the image. Weak edges correspond to areas with small brightness changes and are blurred or indistinct boundaries.

[0041] For example, strong edge regions and weak edge regions can be distinguished by calculating the gradient of pixel values, for instance using the Sobel operator or a simple difference.

[0042] Enhancement is performed on weak edge regions in the first filtered image, for example, by using a high-pass filter or a Laplacian filter to enhance high-frequency information in the image and thereby enhance weak edges. Enhancement may also be performed by contrast adjustment or non-local means, and the specific means of enhancement are not limited in the embodiments of this application.

[0043] Then, strong edge regions in the first filtered image are weakened, for example, by smoothing the strong edges using a bilateral filtering method. For example, in this application, strong edge regions may be weakened by methods such as Gaussian blurring or waveform compression, and the specific means of weakening are not limited in the embodiments of this application.

[0044] By emphasizing the weak edge regions, the high-frequency part of the image is emphasized, thereby increasing the local contrast of the image, enhancing the edges and details in the image, and making the image appear clearer and more vivid. And by weakening the strong edge regions, the noise in the strong edge regions can be reduced and the variance of the image can be reduced.

[0045] In one embodiment of the present application, the method further includes the steps of assigning values to the filtered image based on a predetermined threshold range to obtain the weak edge regions, and assigning values to the filtered image based on a predetermined threshold to obtain the strong edge regions.

[0046] The method of obtaining strong edge regions by threshold and obtaining strong and weak edge regions includes presetting edge detection thresholds th1 and th2. Exemplarily, those with pixel values < th1 are determined as flat regions, those with th1 ≤ pixel values < th2 are determined as weak edge regions, and those with pixel values ≥ th2 are determined as strong edge regions.

[0047] Furthermore, the absolute value is obtained for the filtered result of the first residual image and denoted as ResIntensity1. For example, for the first residual image, convolution is performed with a predefined kernel function, and then the absolute value is obtained to get ResIntensity1, which represents the intensity of the residual layer. By obtaining the absolute value, the plus and minus signs of the values in the residual image can be removed. According to a single value, it cannot be represented which of the strong edge region, weak edge region, and flat region the current pixel is in.

[0048] Filter Pyr[3] using a 3*3 Gaussian kernel with a bandwidth of, and the filtering can be represented by Equation 1.

[0049]

Equation

[0050] The kernel represents a Gaussian kernel, Pyr[3] represents the first residual image described above, and since the image of the third layer in the Laplacian pyramid is used in this embodiment, it is represented by Pyr[3].

[0051] Based on ResIntensity1 and the edge detection threshold, it is possible to distinguish to which of the flat region, weak edge region, and strong edge region each pixel in the current residual layer belongs.

[0052] To obtain the weak edge region, a method of assigning values to ResIntensity1 is used. For example, 0 is assigned to all pixels that satisfy th1 ≤ pixel value < th2. 1 is assigned to other pixels, and in this case, it can be expressed by Equation 2.

[0053]

Equation

[0054] However, (x, y) represents the coordinates of the pixel.

[0055] By assigning values to ResIntensity1, one weak edge region image can be generated.

[0056] To obtain the strong edge region, a method of assigning values to ResIntensity1 can also be used. For example, 1 is assigned to all pixels that satisfy pixel value ≥ th2. 0 is assigned to other pixels, and in this case, it can be expressed by Equation 3.

[0057]

Equation

[0058] (x, y) represents the coordinates of the pixel.

[0059] <00​​In the above embodiment, by assigning a predetermined threshold, strong edge regions and weak edge regions of the image can be obtained quickly and easily.

[0060] In one embodiment of the present application, the sharpened image includes a weak edge region image, and the step of enhancing the weak edge region in the filtering image includes enhancing the weak edge region using a predetermined first scalar amount, while leaving the other parts of the first filtering image other than the weak edge region, thereby generating the weak edge region image.

[0061] To enhance the residual layer in weak edge regions, the following equation 4 can be used.

[0062]

number

[0063] Gain1 is a predetermined scalar quantity satisfying Gain1 > 1. Pyr[3] represents the first residual image described above, and in this embodiment, the image of the third layer in the Laplacian pyramid is used, hence it is represented by Pyr[3]. WeakEdge represents the weak edge region, for example, the weak edge region image obtained by assigning values.

[0064] In Equation 4, by multiplying each pixel value in the first residual image Pyr[3] and the weak edge region WeakEdge by a scalar amount (i.e., Gain1), an image of the same size as the original image is obtained, and linear scaling is performed on the weak edge region in the first residual image using the scalar amount in this image. Furthermore, by adding Pyr[3]*(1-WeakEdge) to the linearly scaled image, a weak edge region image is generated. Through this addition, in the final output image, the parts other than the weak edge region remain, and only the weak edge region is scaled. If the weak edge region is an independent weak edge region image, the enhanced weak edge region image becomes part of the sharpened image.

[0065] In the above embodiment, it is possible to quickly realize the enhancement of the weak edge region by a scalar quantity, while leaving other parts other than the weak edge region, whereby the generated weak edge region image has more image details.

[0066] In one embodiment of the present application, the sharpened image includes a strong edge region image, and the step of weakening the strong edge region in the filtered image includes weakening the strong edge region using a predetermined second scalar quantity, while leaving other parts other than the strong edge region in the filtered image, and generating the strong edge region image.

[0067] The enhancement of the residual layer of the strong edge region can utilize the following Equation 5.

[0068]

Equation

[0069] Gain2 is a predetermined scalar quantity, satisfying 0 < Gain2 < 1. Pyr[3] represents the above first residual image, and in this embodiment, since the image of the third layer in the Laplacian pyramid is used, it is represented by Pyr[3]. StrongEdge represents a strong edge region, for example, a strong edge region image obtained by value assignment.

[0070] In Equation 5, by multiplying each pixel value in the first residual image Pyr[3] and the strong edge region StrongEdge by a scalar amount (i.e., Gain2), an image of the same size as the original image is obtained, and linear reduction is performed on the strong edge region in the first residual image using the scalar amount in this image. Furthermore, by adding Pyr[3]*(1-StrongEdge) to the linearly reduced image, a strong edge region image is generated. Through this addition, the remaining parts of the output image, excluding the strong edge region, are left, and only the strong edge region is reduced. If the strong edge region is an independent strong edge region image, the weakened strong edge region image becomes part of the sharpened image.

[0071] In the above embodiment, the weakening of strong edge regions can be quickly achieved by using a scalar amount, while retaining the other parts outside the strong edge regions, thereby resulting in the generated image of the strong edge region having more image detail.

[0072] S130: A target image is obtained based on the reconstructed image and the sharpened image, the reconstructed image being a superposition of all residual images whose layer order precedes the residual image.

[0073] A reconstructed image is obtained, and the reconstructed image is the result of reconstructing residual images whose layer order precedes the layer order of the first residual image. For example, if the Laplacian pyramid has 6 layers, and the first residual image is the image of the 3rd layer of the Laplacian pyramid, then the reconstructed image is the result of reconstructing the image of the 6th layer, the 5th layer, and the 4th layer of the Laplacian pyramid. The above reconstruction may include image sharpening, superposition, noise reduction, etc. For example, the reconstruction may involve superimposing images onto adjacent residual images in the layer order using the noise reduction method according to this embodiment.

[0074] Based on the reconstructed and sharpened images, reconstruction of the image to be denoised can be performed to obtain the target image. For example, the reconstructed image and the sharpened image can be superimposed to obtain the target image. The target image may also be an intermediate image in the process of denoising the image. For example, the target image can be further superimposed with the image below the third layer image in the Laplacian pyramid to obtain the denoised image. Alternatively, the target image may be the denoised image after denoising. For example, if the Laplacian pyramid has only three layers, the resulting image is the denoised image after denoising.

[0075] During the reconstruction process, sharpening is applied to the superimposed images, and weak edge regions of the superimposed images are enhanced. This enhances the high-frequency components of the image, increasing local contrast, emphasizing edges and details, and making the image appear sharper and more vivid. Furthermore, by weakening strong edge regions, noise in those areas is reduced, and image burrs can be minimized.

[0076] In the above embodiment, by performing sharpening during the reconstruction process of the image to be noise-reduced, the local contrast of the image is increased, edges and details in the image are emphasized, and it becomes unnecessary to perform sharpening after the image reconstruction, making the implementation process simpler and more convenient. Furthermore, by weakening strong edge regions, noise in strong edge regions can be reduced, and image burrs can be reduced.

[0077] In one embodiment of this application, the step of obtaining a target image based on a reconstructed image and the sharpened image includes the steps of establishing a guide image and a noise image based on the reconstructed image and the sharpened image, performing noise reduction on the noise image based on the guide image to obtain a noise-reduced image, and performing upsampling on the noise-reduced image to obtain the target image.

[0078] A guide image and a noise image are established based on the reconstructed image and the sharpened image, respectively. The guide image is used to perform noise reduction on the noise image, and the noise image is a superposition of part or all of the reconstructed image and the sharpened image. Noise reduction is performed on the noise image based on the guide image to obtain a noise-reduced image. Upsampling is performed on the noise-reduced image to obtain the target image described above.

[0079] Furthermore, the sharpened image includes a weak edge region image and a strong edge region image, and the steps of establishing a guide image and a noise image based on the reconstructed image and the sharpened image include the steps of obtaining the noise image by adding the weak edge region image and the reconstructed image, and obtaining the guide image by adding the strong edge region image and the reconstructed image.

[0080] The sharpened image includes images of weak and strong edge regions, the noisy image is represented by NoisyImage, and the guide image is represented by GuideImage. The process of establishing both can be expressed by Equation 6.

[0081]

number

[0082] NewRes1 represents the image of the weak edge region, NewRes2 represents the image of the strong edge region, and Up1 represents the reconstructed image, which generally requires upsampling. For example, if the first residual image above is the image of the third layer in the Laplacian pyramid, the reconstructed image is the result of upsampling the reconstruction of the images of the sixth, fifth, and fourth layers in the Laplacian pyramid. The degree of upsampling is determined by the size of the residual image, meaning that the image generated by upsampling is the same size as the residual image, thus contributing to the superposition of the reconstructed image and the sharpened image.

[0083] In the above embodiment, guide images and noise images can be established quickly and easily. Using guide images allows more weight to be assigned to strong edge regions, increasing the intensity of noise reduction for strong edges in the image. Since the noise image is a superposition of the sharpened residual image, the local contrast of the image is increased, edges and details in the image are emphasized, and the generated image becomes sharper and more vivid.

[0084] In one embodiment of this application, the step of generating a noise-reduced image by performing noise reduction on the noise image based on the guide image includes the step of calculating bilateral filtering weight parameters based on the guide image, and the step of generating the noise-reduced image by performing noise reduction on the noise image based on the weight parameters.

[0085] The weights for bilateral filtering are calculated using a guide image, and noise reduction is performed on the noisy image to generate a noise-reduced image. Bilateral filtering calculates both the pixel difference Img(x,y)-Img(x+k1,y+k2) between the pixel value of the current point (x,y) and the pixel value of another point (x+k1,y+k2) in the adjacent region of the current point, and the position difference of the pixel point. These two values ​​are used to determine the weights for the filtering process of points in the adjacent region of the pixel point (x,y).

[0086] If the noise reduction result is represented by NR2, the noise reduction process can be expressed by Equation 7.

[0087]

number

[0088] k1 represents the offset value of the row coordinate of the pixel point, with a range of [-r~r], k2 represents the offset value of the column coordinate of the pixel point, with a range of [-r~r], r represents the radius of the value, and x+k1,y+k2 represents points within the adjacent region with radius r centered on the current point (x,y). NoisyImage represents a noisy image, e represents the exponential function, σ1 and σ2 represent the bandwidth of the Gaussian function, respectively, and are used to control the intensity of noise reduction, with σ1 being the Gaussian function bandwidth with respect to the pixel value and σ2 being the Gaussian function bandwidth with respect to the pixel coordinate distance.

[0089] NR2(x,y) represents the noise-reduced image and is a weight calculated by bilateral filtering. The weighted average of points selected from the adjacent region of pixel point (x,y) is used to determine the noise reduction result for point (x,y).

[0090] In the above embodiment, by calculating the weights of bilateral filtering using a guide image, noise reduction is performed on the noisy image, the intensity of noise reduction is increased for strong edges in the image, the burrs on strong edges of the image after noise reduction by the edge-preserving noise reduction algorithm are reduced, and sharpening is performed simultaneously with noise reduction, thereby emphasizing the details of the image.

[0091] The embodiment shown in Figure 1 is suitable for processing adjacent residual images in the layer order during the image reconstruction process. If other methods are used for other residual images of the image, the noise-reduced image obtained in the reconstruction may not be optimal. In view of this, this application further provides a second embodiment shown in Figure 2. In this embodiment, the noise reduction method shown in Figure 1 is used for each of the adjacent residual images in the layer order, resulting in a better-responding image after reconstruction.

[0092] Figure 2 is a flowchart of another noise reduction method according to an embodiment of the present application. As shown in Figure 2, the method includes steps S210 to S220. S210: Multiple residual images of an image are acquired, each residual image being a difference value at different resolutions of the image, and there is a predetermined layer order between the multiple residual images.

[0093] All residual images corresponding to the image to be noise-reduced are obtained, each residual image being a difference value at a different resolution of the image, and there is a predetermined layer order relationship between the residual images. For example, a Laplacian pyramid of the image to be noise-reduced is obtained, the Laplacian pyramid consists of multiple residual images, and there is a predetermined layer order relationship between the residual images in the Laplacian pyramid.

[0094] S220: Multiple superpositions are performed on the multiple residual images in accordance with the layer order to generate a noise-reduced image of the image. At least one of the multiple superpositions includes the steps of: weakening strong edge regions and strengthening weak edge regions in the filtering image corresponding to the residual image of the current layer to generate a sharpened image; and obtaining a reconstructed image of the current layer based on the reconstructed image of the layer above and the sharpened image.

[0095] A noise-reduced image can be generated by superimposing all residual images corresponding to the image to be noise-reduced, in the order of the layers. For example, if there are a total of three residual images, superimposing the third-level image and the second-level image yields a superposition result, and superimposing this superposition result with the first-level image generates the noise-reduced image to be noise-reduced.

[0096] In the process of superimposing some images where layers are adjacent, a filtered image of the current layer is obtained, and this filtered image is obtained by filtering the residual image of the current layer. For example, if there are six residual images, and the superimposition is performed from the top layer, when it proceeds to the third layer, the residual image of the current layer, i.e., the third layer, is obtained, and filtering is performed on this image to generate the filtered image of the current layer. Strong edge regions in the filtered image of the current layer are weakened, and weak edge regions are strengthened to generate a sharpened image of the current layer. A reconstructed image of the layer one layer above the current layer is obtained, and the reconstructed image is a superimposition of all images above the current layer. For example, if the current layer is the residual image of the third layer, the reconstructed image of the layer one layer above it, i.e., the fourth layer, is obtained, and the reconstructed image of the fourth layer is an image generated by processing and superimposing the residual image of the sixth layer, the residual image of the fifth layer, and the residual image of the fourth layer. Reconstruction is performed based on the reconstructed image of the layer one layer above and the sharpened image of the current layer to generate the reconstructed image of the current layer.

[0097] In the above embodiment, during the process of superimposing the residual images of each layer of the image to be noise-reduced, sharpening is performed on the superimposed image. This increases the local contrast of the image, emphasizes edges and details in the image, eliminates the need for sharpening after image reconstruction, makes the implementation process simpler and more convenient, and reduces noise in strong edge regions by weakening strong edge regions, thereby reducing image burrs.

[0098] In one embodiment of this application, the step of generating a noise-reduced image by superimposing a plurality of residual images multiple times in accordance with the layer order is performed starting from a target residual image among the plurality of residual images, performing multiple iterations in accordance with the layer order, and continuing until the residual image with the lowest layer rank is the noise-reduced image. The target residual image refers to the residual image with the third layer rank. Each iteration includes the steps of: filtering the residual image of the current iteration to generate a filtered image of the current iteration; weakening strong edge regions and strengthening weak edge regions in the filtered image of the current iteration to generate a sharpened image of the current iteration; and obtaining a reconstructed image of the current iteration based on the reconstructed image of the previous iteration and the sharpened image of the current iteration.

[0099] Processing and overlaying of residual images can be performed using an iterative method. The iteration method includes the following:

[0100] The process begins with a target residual image among multiple residual images and performs multiple iterations in descending order of layer rank. The target residual image is the image from which the iteration begins, and is typically the third residual image layer in the residual image sequence. In other embodiments of this application, it may be any other layer among the residual images. The third residual image in the layer sequence is designated as the target residual image, and the iteration begins from the target residual image and continues down to the residual image with the lowest layer rank. In this case, the reconstructed image generated during the final iteration is the noise-reduced image. Exemplarily, the residual images take the form of a Laplacian pyramid, with a total of six layers in the Laplacian pyramid, and the layer rank numbers of the residual images are sequentially 5 to 0 from the top layer to the bottom layer, and the iteration begins from the third residual image (i.e., the residual image with number 3) and continues down to the bottom residual image (i.e., the residual image with number 0).

[0101] An iteration includes the following: obtaining the residual image of the current iteration, assuming, for illustrative purposes, that the current iteration is the first iteration, and that the residual image of the current iteration is the third residual image with the third layer order, obtaining this residual image, filtering it, and generating a filtered image of the current iteration.

[0102] Currently, strong edge regions in the filtered image of the current iteration are softened, and weak edge regions are enhanced to generate a sharpened image of the current iteration.

[0103] The reconstructed image from the previous iteration is obtained, and each reconstructed image from each iteration is the result of processing and overlaying all residual images preceding the residual image of the current iteration. If the residual image of the current iteration is the third residual image, the reconstructed image from the previous iteration is the image generated by processing and overlaying the fourth and fifth residual images. Furthermore, the reconstructed image from the previous iteration is upsampled, and the size of the upsampled image is the same as the size of the filtered image of the current iteration.

[0104] The reconstructed image from the previous iteration and the sharpened image from the current iteration are processed and superimposed to generate the reconstructed image for the current iteration.

[0105] In the above embodiment, reconstruction can be rapidly performed on multiple residual images of the image using an iterative method, thereby obtaining an image with reduced noise.

[0106] In one embodiment of the present application, a method for generating a reconstructed image of an iteration immediately preceding the first iteration of a plurality of iterations includes the steps of: generating a first filtered image by performing a first filtering on a first residual image; generating a second filtered image by performing a second filtering on a second residual image; generating a corresponding second sharpened image by weakening strong edge regions and enhancing weak edge regions in the second filtered image; and obtaining a reconstructed image of an iteration immediately preceding the first iteration of a plurality of iterations based on the first filtered image and the second sharpened image, wherein the first residual image is the residual image with the highest layer rank, the first filtering and the second filtering are different, and the second residual image is an image of the layer one layer below the first residual image.

[0107] In the process of iteration, if there is no reconstructed image for the first iteration, the method for generating the reconstructed image from the previous iteration to be used for the first iteration includes the following steps.

[0108] The residual image with the highest layer rank among multiple residual images is obtained. For example, if a 6-layer Laplacian pyramid is used in the above iteration, the residual image of the 6th layer is obtained, and a first filtering is performed on the residual image of that layer to obtain a first filtered image. In this application, the specific means of the first filtering are not limited, and the first filtering may be bilateral filtering, Gaussian filtering, etc. The image of the layer one level below the residual image with the highest layer rank is obtained. For example, the image of the layer one level below the 6th layer image, i.e., the 5th layer image, is obtained, and a second filtering is performed on this image to obtain a second filtered image. The second filtering may be Gaussian filtering, bilateral filtering. If the methods of the first filtering and the second filtering are different, this application can be implemented with relatively good effectiveness. Exemplarily, in one embodiment of this application, the first filtering is bilateral filtering and the second filtering is Gaussian filtering.

[0109] Strong edge regions in the second filtered image are softened, and weak edge regions in the second filtered image are enhanced, generating a corresponding second sharpened image.

[0110] Reconstruction is performed based on the second sharpened image and the first filtered image to generate a reconstructed image from the previous iteration to be used in the first iteration. Furthermore, reconstruction requires upsampling of the first filtered image, and the degree of upsampling is determined by the size of the second residual image; that is, the image generated by upsampling the first filtered image should be the same size as the second filtered image. For example, if the second filtered image is twice the size of the first filtered image, the first filtered image is upsampled by 2x using bilinear interpolation to obtain the first enlarged image.

[0111] The above reconstruction includes image sharpening, overlay, and noise reduction. For example, the absolute value is calculated for the second overlay image to obtain the second absolute value, and a value is assigned to the second absolute value based on a predetermined edge detection threshold to generate strong edge layers and weak edge layers, respectively. The weak edge layer is enhanced to generate a weak edge image, and the strong edge layer is weakened to generate a strong edge image. The weak edge image and the first enlarged image are added to obtain a noise image, the strong edge image and the first enlarged image are added to obtain a guide image, the guide image is used to calculate bilateral filtering weights, and based on the obtained weights, bilateral filtering is performed on the noise image to generate the reconstructed image from the previous iteration to be used in the first iteration.

[0112] Figure 3 is a block diagram showing the configuration of a noise reduction module 300 according to an embodiment of the present application. As shown in Figure 3, the noise reduction module 300 comprises a filtering module 310, a sharpening module 320, and a generation module 330.

[0113] The filtering module 310 is configured to perform filtering on residual images of an image to generate a filtered image, wherein the residual images are difference values ​​of the image at different resolutions, at least two residual images correspond to the image, and there is a predetermined layer order relationship between the at least two residual images.

[0114] The sharpening module 320 is configured to enhance weak edge regions in the filtered image and weaken strong edge regions in the filtered image to generate a sharpened image. The generation module 330 is configured to obtain a target image based on the first filtered image and the sharpened image.

[0115] In one selectable embodiment of the present application, the generation module 330 comprises: an image generation submodule configured to establish a guide image and a noise image based on the reconstructed image and the sharpened image; a noise reduction submodule configured to perform noise reduction on the noise image based on the guide image to obtain a noise reduction image; and an upsampling submodule configured to perform upsampling on the noise reduction image to obtain a target image.

[0116] In one selectable embodiment of the present application, the noise reduction module 300 further comprises a first assignment module configured to acquire the weak edge regions by assigning values ​​to the filtered image based on a predetermined threshold range, and a second assignment module configured to acquire the strong edge regions by assigning values ​​to the filtered image based on a predetermined threshold.

[0117] In one selectable embodiment of the present application, the sharpened image includes a weak edge region image, and the sharpening module 320 is configured to specifically enhance the weak edge region using a predetermined first scalar amount, while leaving the rest of the filtered image other than the weak edge region to generate the weak edge region image.

[0118] In one selectable embodiment of the present application, the sharpened image includes a strong edge region image, and the sharpening module 320 is configured to specifically weaken the strong edge region using a predetermined second scalar amount, while leaving the rest of the filtered image other than the strong edge region, thereby generating the strong edge region image.

[0119] In one selectable embodiment of the present application, the sharpened image includes a weak edge region image and a strong edge region image, and the image generation submodule is configured to obtain the noise image by adding the weak edge region image and the reconstructed image, and to obtain the guide image by adding the strong edge region image and the reconstructed image.

[0120] In one selectable embodiment of the present application, the noise reduction submodule is configured to calculate bilateral filtering weight parameters based on the guide image and to perform noise reduction on the noise image based on the weight parameters to generate the noise reduction image.

[0121] Figure 4 is a block diagram showing the configuration of an electronic device 400 according to an embodiment of this application. As shown in Figure 4, the electronic device 400 comprises a processor 410 and a memory 420.

[0122] The processor 410, the memory 420, and each element are electrically connected directly or indirectly to enable data transmission or exchange. For example, these elements are electrically connected by one or more communication buses or signal lines. The memory 420 is configured to store computer programs, and for example, the software function module shown in Figure 3, namely the noise reduction module 300, is stored in the memory 420. The noise reduction module 300 comprises at least one software function module stored in the memory 420 in the form of software or firmware, or incorporated into the operating system (OS) of the electronic device 400. The processor 410 is configured to execute executable modules stored in the memory 420, such as the software function modules or computer programs of the noise reduction module 300.

[0123] Memory 420 may be, but is not limited to, Random Access Memory (RAM), Read Only Memory (ROM), Programmable Read-Only Memory (PROM), Erasable Programmable Read-Only Memory (EPROM), or Electric Erasable Programmable Read-Only Memory (EEPROM).

[0124] The processor 410 may be an integrated circuit chip having signal processing capabilities. The processor may be a general-purpose processor including a Central Processing Unit (CPU), a Network Processor (NP), a microprocessor, a Digital Signal Processor (DSP), an Application Specific Integrated Circuit (ASIC), a Field Programmable Gate Array (FPGA), or other programmable logic device, discrete gate or transistor logic device, or discrete hardware component, capable of realizing or executing each method, step, and logic block disclosed in the embodiments of this application. Alternatively, the processor 410 may be any conventional processor.

[0125] Embodiments of this application further provide a computer-readable non-volatile storage medium (hereinafter abbreviated as "storage medium"). A computer program is stored in the storage medium, and when the computer program is executed by a computer, for example, the electronic device 400 described above, the noise reduction method described above is performed.

[0126] Each example in this specification is described progressively, with emphasis on describing the differences between each example and the similarities between them, which can be discussed through cross-referencing.

[0127] In some embodiments of this application, the described apparatus and methods can be implemented in other ways. The embodiments of the apparatus described above are illustrative only. For example, the flowcharts and block diagrams in the drawings illustrate implementable architectures, functions, and operations based on the apparatus, methods, and computer program products according to some embodiments of this application, where each block in the flowchart or block diagram can represent a module, program segment, or part of code containing one or more executable commands capable of implementing a predetermined logical function. In some interchangeable implementations, the implementation of the functions described in the blocks may differ from the order shown in the drawings. For example, two consecutive blocks may actually be executed almost in parallel, or in some cases in reverse order. This is determined by the required functions. Each block in the block diagram and / or flowchart, and combinations of blocks in the block diagram and / or flowchart, may be implemented by a dedicated, hardware-based system that performs a predetermined function or operation, or by a combination of dedicated hardware and computer commands.

[0128] Furthermore, each functional module in each embodiment of this application may be formed by integrating them into a single independent part, each module may exist independently, or it may be formed by integrating two or more modules into a single independent part.

[0129] The aforementioned functions can be implemented in the form of software function modules, which, when sold or used as independent products, can be stored on a single computer-readable medium. Under this understanding, the technical proposal of this application, or any part thereof that can contribute to the prior art, or any part thereof, can be implemented in the form of a software product. This computer software product is stored on a computer-readable medium and includes a number of commands for a computer device (which may be a personal computer, laptop computer, server, or electronic device) to perform all or part of the steps of the methods according to each embodiment of this application. The computer-readable medium includes various media capable of storing program code, such as USB disks, portable hard disks, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0130] The above are merely specific embodiments of the present application, and the scope of protection of this application is not limited thereto. A person skilled in the art will know that any changes or substitutions made within the scope of the art disclosed in this application are also within the scope of protection of this application. Therefore, the scope of protection of this application is equivalent to the claims.

Claims

1. The steps include: applying filtering to the residual image of the image to generate a filtered image; The steps include: enhancing weak edge regions in the filtered image and weakening strong edge regions in the filtered image to generate a sharpened image; The steps include obtaining a target image based on the reconstructed image and the sharpened image, The residual image is the difference value of the image at different resolutions, at least two residual images correspond to the image, there is a predetermined layer order relationship between the at least two residual images, and the reconstructed image is a superposition of all residual images whose layer order precedes the residual image. A noise reduction method characterized by the following.

2. The step of obtaining a target image based on the reconstructed image and the sharpened image is: The steps include establishing a guide image and a noise image based on the reconstructed image and the sharpened image, The steps include: performing noise reduction on the noise image based on the guide image to obtain a noise-reduced image; The step includes performing upsampling on the noise-reduced image to obtain the target image. The noise reduction method according to claim 1.

3. A step of assigning values ​​to the filtered image based on a predetermined threshold range to obtain the weak edge region, The step of assigning values ​​to the filtered image based on a predetermined threshold to obtain the strong edge region further includes The noise reduction method according to claim 1.

4. The step of enhancing the weak edge regions in the filtered image, provided that the sharpened image includes images of weak edge regions, is as follows: This includes enhancing the weak edge region using a predetermined first scalar quantity, while retaining the other parts of the filtered image other than the weak edge region, and generating the weak edge region image. The noise reduction method according to claim 1.

5. The sharpened image includes an image of strong edge regions, and the step of weakening the strong edge regions in the filtered image is: This includes weakening the strong edge region using a predetermined second scalar quantity, while leaving the other parts of the filtered image that are not the strong edge region, and generating the strong edge region image. The noise reduction method according to claim 1.

6. The step of establishing a guide image and a noise image based on the reconstructed image and the sharpened image, wherein the sharpened image includes a weak edge region image and a strong edge region image, The steps include obtaining the noise image by adding the weak edge region image and the reconstructed image, The step of obtaining the guide image by adding the strong edge region image and the reconstructed image is included. The noise reduction method according to claim 2.

7. The step of generating a noise-reduced image by performing noise reduction on the noise image based on the guide image is: The steps include: calculating bilateral filtering weight parameters based on the aforementioned guide image; The step includes generating a noise-reduced image by performing noise reduction on the noise image based on the weight parameters. The noise reduction method according to claim 2.

8. The steps include obtaining multiple residual images of the image, The step includes generating a noise-reduced image of the aforementioned plurality of residual images by performing multiple superpositions in accordance with the layer order, Each residual image is the difference value at different resolutions of the aforementioned image, and there is a predetermined layer order between the plurality of residual images. At least one of the multiple superposition steps includes the steps of: weakening strong edge regions and strengthening weak edge regions in the filtering image corresponding to the residual image of the current layer to generate a sharpened image; and obtaining a reconstructed image of the current layer based on the reconstructed image of the layer above and the sharpened image. A noise reduction method characterized by the following.

9. The step of generating a noise-reduced image by performing multiple superpositions of the aforementioned multiple residual images in accordance with the layer order is: The process includes starting with the target residual image among the multiple residual images, performing multiple iterations in accordance with the layer order, iterating down to the residual image with the lowest layer rank, and using the reconstructed image from the final iteration as the noise reduction image, wherein the target residual image refers to the residual image with the third layer rank. Each iteration includes the steps of: filtering the residual image of the current iteration to generate a filtered image of the current iteration; weakening strong edge regions and strengthening weak edge regions in the filtered image of the current iteration to generate a sharpened image of the current iteration; and obtaining a reconstructed image of the current iteration based on the reconstructed image of the previous iteration and the sharpened image of the current iteration. The noise reduction method according to claim 8.

10. The method for generating the reconstructed image of the iteration immediately preceding the first iteration among the multiple iterations is as follows: The steps include: performing a first filtering on the first residual image to generate a first filtered image; The steps include: performing a second filtering on the second residual image to generate a second filtered image; The steps include: weakening strong edge regions and enhancing weak edge regions in the second filtered image to generate a corresponding second sharpened image; The process includes the step of obtaining a reconstructed image of the iteration immediately preceding the first iteration among the multiple iterations, based on the first filtered image and the second sharpened image, The first residual image is the residual image with the highest layer rank, the first filtering and the second filtering are different, and the second residual image is the image of the layer one layer below the first residual image. The noise reduction method according to claim 9.

11. A filtering module configured to perform filtering on the residual image of an image to generate a filtered image, A sharpening module is configured to generate a sharpened image by enhancing weak edge regions in the filtered image and weakening strong edge regions in the filtered image. The system comprises a generation module configured to obtain a target image based on the first filtered image and the sharpened image, The residual image is the difference value of the image at different resolutions, and at least two residual images correspond to the image, and there is a predetermined layer order relationship between the at least two residual images. A noise reduction device characterized by the following features.

12. The noise reduction method according to any one of claims 1 to 10 is performed, comprising a processor and memory, wherein the processor and memory communicate via a bus, and program commands executable by the processor are stored in the memory, and when the program commands are read by the processor, the noise reduction method according to any one of claims 1 to 10 is performed. An electronic device characterized by the following features.

13. A computer program command is stored in the computer-readable medium, and when the computer program command is read and executed by a computer, the noise reduction method according to any one of claims 1 to 10 is performed. A computer-readable medium characterized by the following:

Images