Method, device and equipment capable of improving image quality twice

By combining greedy algorithm and POCS algorithm, local variance selection filters are used to perform image noise reduction processing, the problem of noise amplification in image super-resolution reconstruction of convex projection algorithm is solved, and the image quality is improved twice, ensuring the edge and texture preservation and image constraint satisfaction.

CN119090763APending Publication Date: 2024-12-06QUANZHOU LIGHT IND VOCATIONAL COLLEGE

Patent Information

Application Number
CN202411081415.6
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-08
Publication Date
2024-12-06

AI Technical Summary

Technical Problem

The convex set projection algorithm may amplify noise during image super-resolution reconstruction, especially when there are many iterations, and the interpolation method is not accurate enough, which affects the noise processing effect.

Method used

Combining the greedy algorithm and the POCS algorithm, the image is denoised by selecting appropriate filters through local variance, ensuring that important features of the image, such as edges and textures are preserved while removing noise.

Benefits of technology

Effectively improve image quality, balance noise reduction and detail retention, avoid image smoothing or brightness abnormalities caused by noise amplification, and ensure that the final output image meets preset constraints.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119090763A_ABST
    Figure CN119090763A_ABST
Patent Text Reader

Abstract

The invention discloses a method, a device and equipment capable of improving image quality twice. The method comprises the following steps: acquiring an original image, and performing super-resolution reconstruction on the original image by using a convex set projection algorithm to obtain a first image; based on the first image, traversing each local pixel region of the image, calculating a local variance of each local pixel region by using a greedy algorithm, and selecting a corresponding filter according to the local variance; and performing noise reduction processing on each pixel region of the first image according to the selected filter, and outputting a second image. According to the method, the greedy algorithm and the POCS algorithm are combined for use, so that the important features of the image can be effectively retained while the image noise is removed, and particularly, the important features such as edges and textures of the image are retained while the noise is removed under the condition that balanced noise reduction and detail retention are needed.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the field of image technology, and in particular to a method, a device and a equipment capable of improving image quality twice. Background Art

[0002] As a spatial high-resolution image reconstruction algorithm, the convex set projection algorithm can process the information in multiple frames of low-resolution images to obtain a high-resolution image that can be used for recording. At the same time, it can iteratively correct the image by simulating errors. The introduction of the concept of boundary amplitude makes the obtained image space have diverse characteristics. However, the convex set projection algorithm may amplify the noise in the image, especially when the number of iterations is large. This is because the algorithm corrects the current high-resolution estimate in each iteration. If the correction is not appropriate, noise may be introduced or amplified. In addition, if the interpolation method used in the algorithm is not accurate enough, or if the influence of noise is not fully considered when simulating the image acquisition process, the effect of noise processing may also be affected. Summary of the invention

[0003] In view of this, the purpose of the present invention is to propose a method, a device and a equipment that can improve the image quality twice. This case introduces a greedy algorithm to refine the restored and reconstructed image to improve the image quality.

[0004] According to one aspect of the present invention, a method for improving image quality twice is provided, comprising:

[0005] Acquire an original image, and use a convex set projection algorithm to perform super-resolution reconstruction on the original image to obtain a first image;

[0006] Based on the first image, traverse each local pixel area of ​​the image, calculate the local variance of each local pixel area by using a greedy algorithm and select a corresponding filter according to the local variance;

[0007] A noise reduction process is performed on each pixel area of ​​the first image according to the selected filter, and a second image is output.

[0008] In the above technical scheme, the present invention combines the greedy algorithm and the POCS algorithm to effectively retain the important features of the image while removing image noise, especially when it is necessary to balance noise reduction and detail retention, that is, to remove noise while maintaining important features of the image, such as edges and textures.

[0009] In some embodiments, an original image is acquired, and a convex set projection algorithm is used to perform super-resolution reconstruction on the original image to obtain a first image. Specifically:

[0010] When using the convex set projection algorithm, bilinear interpolation is used for upsampling.

[0011] In the above technical solution, bilinear interpolation is a relatively simple and computationally efficient upsampling method. It calculates the value of the new pixel by weighted averaging the four nearest neighbor pixels. This process is less computationally intensive and suitable for use in iterative algorithms. Bilinear interpolation can provide a smoother image and reduce the aliasing and sawtooth effects that may be introduced during the upsampling process, thereby maintaining the visual quality of the image. Compared with nearest neighbor interpolation or the simplest form of linear interpolation, bilinear interpolation can better reduce the blurring effect during the upsampling process.

[0012] In some embodiments, the range of the local pixel area is i*n; wherein i is 1 / 5-1 / 10 of the number of pixels in the length direction of the original image, n is 1 / 5-1 / 10 of the number of pixels in the width direction of the original image, and i and n are rounded up.

[0013] In the above technical solution, the purpose of the above setting is to balance the noise reduction effect and iteration speed. If the subdivision is too small, the noise reduction effect will be improved, but the application of the filter and the subsequent iteration time will be greatly increased. If the subdivision is too large, the effect cannot be guaranteed. After many common sense, the selection of local pixel areas limited to the above range can well balance the relationship between the two.

[0014] In some embodiments, a greedy algorithm is used to calculate the local variance of each local pixel region and a corresponding filter is selected according to the local variance. Specifically:

[0015] According to the filter characteristics, it is divided into smoothing filter group and edge detail filter group;

[0016] If the local variance is greater than the threshold, the filters in the edge detail filter group are applied one by one, and each filter is evaluated using a greedy algorithm and one is selected for application;

[0017] Otherwise, the filters in the smoothing filter bank are applied one by one, each filter is evaluated using a greedy algorithm, and one is selected for application.

[0018] In the above technical solution, the greedy algorithm selects a filter based on the local variance. In low variance areas, a smoother filter (such as a Gaussian filter) may be selected, while in high variance areas (which may contain edges), a more conservative filter (such as a mean filter or a median filter) is selected to preserve the edges.

[0019] In some embodiments, the filter comprises a mean filter, a Gaussian filter, or a median filter.

[0020] In the above technical solution, different filters can process different areas. For example, the Gaussian filter is more suitable for smooth areas, while the other two are more suitable for processing edge areas.

[0021] In some embodiments, noise reduction processing is performed on each pixel area of ​​the first image according to the selected filter to output the second image, specifically:

[0022] Set several constraint sets;

[0023] Project the pixel values ​​in the second image into the constraint set. If the constraint set conditions are not met, then

[0024] Based on the second image, traverse each local pixel area of ​​the image, calculate the local variance of each local pixel area by using a greedy algorithm and select a corresponding filter according to the local variance;

[0025] Perform noise reduction processing on each pixel area of ​​the second image according to the selected filter, output a third image, and repeat the above steps until the constraint set conditions are met or the maximum number of iterations is reached, and then output a final image.

[0026] In the above technical solution, although this case adopts a greedy algorithm to solve the problem of regional noise reduction, there may still be problems in which the pixel values ​​near the injected edge in the image after noise reduction are too smooth or the pixel values ​​are negative or the image brightness is too high. Based on the above situation, this case chooses to further use the convex set projection algorithm to solve this problem. Specifically, the image after greedy filtering is projected into a set that satisfies all constraints to ensure that each pixel value of the image meets specific requirements. This method makes it possible for the greedy algorithm to not consider the global constraints in the selection of the local optimal solution, and the POCS algorithm can also ensure that the final result satisfies all preset constraints. This method prevents the above-mentioned problems from occurring in the final output image.

[0027] In some embodiments, the plurality of constraint sets include: non-negativity constraints, range constraints, sum constraints, statistical constraints, texture and detail preservation constraints, and edge preservation constraints.

[0028] In the above technical solution, the non-negativity constraint means that if the pixel value needs to be non-negative, for any pixel value less than 0, it is set to 0. The range constraint means that for an 8-bit grayscale image, all pixel values ​​are ensured to be within the range of 0 to 255. Pixel values ​​outside this range need to be clipped to the nearest boundary value. The sum constraint means calculating the total brightness of the original image and the filtered image. If the two are not equal, the total brightness can be adjusted by scaling all pixel values ​​of the filtered image. Statistical constraints mean that if specific statistical properties need to be maintained, these statistical constraints can be met by adjusting each pixel value. Texture and detail preservation constraints refer to using specific algorithms (such as bilateral filters) to maintain the texture and details of the image while satisfying other constraints. Edge preservation constraints refer to applying an edge detection algorithm and then using an edge enhancement or preservation algorithm to ensure that the edge information of the image is not over-smoothed. It should be stated that the above constraints need to be determined according to the specific usage scenario, and no further elaboration is given here.

[0029] According to another aspect of the present invention, a device capable of improving image quality twice is provided, based on the above method; comprising:

[0030] The first processing module is used to obtain an original image, and use a convex set projection algorithm to perform super-resolution reconstruction on the original image to obtain a first image;

[0031] A selection module: used for traversing each local pixel area of ​​the image based on the first image, calculating the local variance of each local pixel area by using a greedy algorithm, and selecting a corresponding filter according to the local variance;

[0032] The second processing module is used to perform noise reduction processing on each pixel area of ​​the first image according to the selected filter, and output a second image.

[0033] In the above technical solution, in order to better use the above method, the present application proposes a device that can improve the image quality twice. Each module corresponds to each step of the above method. The specific principles have been described above and will not be repeated here.

[0034] According to another aspect of the present invention, there is provided a device capable of improving image quality twice, comprising:

[0035] at least one processor and a memory communicatively coupled to the at least one processor;

[0036] The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can perform the above method.

[0037] In the above technical solution, in order to better run and process the method, the above method is stored in a memory, and the stored method is executed by a processor. It should be noted that the principle and effect of each step have been described above and will not be described in detail here.

[0038] According to another aspect of the present invention, a computer-readable storage medium is provided, storing a computer program, wherein the computer program implements the above method when executed by a processor.

[0039] In the above technical solution, in order to better run and use the method, the above method is stored in a computer-readable storage medium, and the above method is implemented by a processor. It should be noted that the principle and effect of each step have been described above and will not be described in detail here. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0041] Figure 1 It is a flow chart of an embodiment of a method for improving image quality twice according to the present invention;

[0042] Figure 2 It is a structural schematic diagram of an embodiment of a device capable of improving image quality twice according to the present invention. DETAILED DESCRIPTION

[0043] The present invention will be further described in detail below in conjunction with the accompanying drawings and examples. It is particularly noted that the following examples are only used to illustrate the present invention, but are not intended to limit the scope of the present invention. Similarly, the following examples are only partial embodiments of the present invention rather than all embodiments, and all other embodiments obtained by those of ordinary skill in the art without creative work are within the scope of protection of the present invention.

[0044] The present invention provides a method for improving image quality twice. In this case, a greedy algorithm is introduced to refine the restored and reconstructed image to improve image quality.

[0045] Embodiment 1

[0046] See also Figure 1 , a method that can improve image quality twice, including:

[0047] S1, obtaining an original image, and performing super-resolution reconstruction on the original image using a convex set projection algorithm to obtain a first image;

[0048] In this embodiment, the specific steps of super-resolution reconstruction using the convex set projection algorithm can be referred to the prior art, and will not be described in detail in this case. There are many ways to obtain the original image, which will not be described in detail here.

[0049] In this embodiment, when using the convex set projection algorithm, bilinear interpolation is used for upsampling. Bilinear interpolation is a relatively simple and computationally efficient upsampling method. It calculates the value of a new pixel by weighted averaging the four nearest neighbor pixels. This process has a small amount of calculation and is suitable for use in an iterative algorithm. Bilinear interpolation can provide a smoother image, reduce the aliasing effect and sawtooth effect that may be introduced during the upsampling process, and thus maintain the visual quality of the image. Compared with nearest neighbor interpolation or the simplest form of linear interpolation, bilinear interpolation can better reduce the blurring effect during the upsampling process.

[0050] S2, based on the first image, traverse each local pixel area of ​​the image, calculate the local variance of each local pixel area by using a greedy algorithm and select a corresponding filter according to the local variance;

[0051] In this embodiment, the range of the local pixel area is i*n; wherein, i is 1 / 5-1 / 10 of the number of pixels in the length direction of the original image, and n is 1 / 5-1 / 10 of the number of pixels in the width direction of the original image, and i and n are rounded up. The purpose of the above setting is to balance the noise reduction effect and the iteration speed. If the subdivision is too small, the noise reduction effect will be improved, but the application of the filter and the subsequent iteration time will be greatly increased. If the subdivision is too large, the effect cannot be guaranteed. After many times of common sense, the selection of the local pixel area is limited to the above range to balance the relationship between the two well, and the optimal values ​​can be selected as 1 / 8 and 1 / 5.

[0052] In this embodiment, the filter includes a mean filter, a Gaussian filter, and a median filter. Different filters can process different areas. For example, the Gaussian filter is more suitable for smooth areas, while the other two are more suitable for processing edge areas.

[0053] In this embodiment, as a specific implementation scheme, S2, using a greedy algorithm to calculate the local variance of each local pixel area and selecting a corresponding filter according to the local variance, specifically:

[0054] S21, dividing into a smoothing filter group and an edge detail filter group according to filter characteristics;

[0055] S22, if the local variance is greater than the threshold, applying filters in the edge detail filter group one by one, using a greedy algorithm to evaluate each filter, and selecting one to apply;

[0056] S23. Otherwise, apply the filters in the smoothing filter bank one by one, evaluate each filter using a greedy algorithm, and select one to apply.

[0057] According to the local variance, the greedy algorithm selects a filter. For each local area, the variance of its pixel values ​​is calculated. Areas with high variance may contain more details or edges, while areas with low variance may mainly be smooth areas or noise. In low variance areas, a smoother filter (such as a Gaussian filter) may be selected, while in high variance areas (which may contain edges), a more conservative filter (such as a mean filter or a median filter) is selected to retain the edges. It should be noted that the threshold selection needs to be set according to different processing items, and this embodiment does not limit this. Those skilled in the art can limit it according to actual needs.

[0058] In this embodiment, the evaluation criteria of the greedy algorithm for evaluating the filter effect may be: quantization error (e.g., mean square error MSE), signal-to-noise ratio (SNR), structural similarity (SSIM), and edge preservation capability. Based on the evaluation results, specific selections may be made in actual applications, which will not be described in detail here.

[0059] In this embodiment, the greedy algorithm can also be used to enhance the details of the image during the noise reduction process. For example, important edges and textures in the image can be identified after preliminary noise reduction, and specific operations can be applied to enhance these features. This will not be described in detail here, and the prior art can be referred to to further improve the rules for selecting the filter to enhance the edge and texture enhancement.

[0060] In this embodiment, the local variance can be calculated as follows:

[0061] Select window size: Determine an appropriate pixel area size, such as 7x7 pixels.

[0062] Calculate the local average: For each local pixel region, calculate the average of the pixel values. This can be calculated by summing all the pixel values ​​in the pixel region and dividing by the total number of pixels in the pixel region.

[0063] Calculate the local variance: Use the local mean, calculate the sum of the squares of the differences between each pixel value and the mean in the pixel area, and finally divide it by the total number of pixels in the pixel area (or subtract 1 to get an unbiased estimate).

[0064] The estimation of local variance can provide useful information for image denoising algorithms. For example, in areas with low variance, it may indicate the presence of more noise, so a stronger denoising filter can be applied; while in areas with high variance, it may contain important image details, so a more conservative denoising strategy should be used.

[0065] S3. Perform noise reduction processing on each pixel area of ​​the first image according to the selected filter, and output a second image.

[0066] In this embodiment, by combining the greedy algorithm with the POCS algorithm, the present invention can effectively retain important features of the image while removing image noise, especially when it is necessary to balance noise reduction and detail retention, that is, remove noise while maintaining important features of the image, such as edges and textures.

[0067] It should be noted that the selected filter is applied to each pixel region to greedily reduce noise. This process may involve the allocation of weights, where pixels near the edge have higher weights. For specific adjustment of filter weights and other parameters, please refer to the existing technology, which will not be repeated in this embodiment.

[0068] In this embodiment, S3, performing noise reduction processing on each pixel area of ​​the first image according to the selected filter, and outputting the second image, specifically:

[0069] S31, setting a number of constraint sets;

[0070] S32, projecting the pixel values ​​in the second image into the constraint set, if the constraint set conditions are not met, then

[0071] S33, based on the second image, traverse each local pixel area of ​​the image, calculate the local variance of each local pixel area by using a greedy algorithm and select a corresponding filter according to the local variance;

[0072] S34, performing noise reduction processing on each pixel area of ​​the second image according to the selected filter, outputting a third image, and repeating the above steps until the constraint set conditions are met or the maximum number of iterations is reached, and then outputting a final image.

[0073] In this embodiment, the several constraint sets include: non-negativity constraints, range constraints, sum constraints, statistical constraints, texture and detail preservation constraints, and edge preservation constraints. During the denoising process, some physical or statistical constraints may need to be met, such as the non-negativity of the image, the total brightness remaining unchanged, etc. The convex set projection algorithm can project the current denoising result into a set that satisfies these constraints in each iteration. If there are multiple constraints, such as image smoothness, edge preservation, etc., the convex set projection algorithm can ensure that these constraints are considered simultaneously in each iteration, thereby obtaining a more balanced denoising effect. The greedy algorithm provides a fast local optimization method, while the convex set projection algorithm ensures the global feasibility and constraint satisfaction of the solution. The combination of the two provides an effective denoising strategy. The following is an explanation of the constraint set:

[0074] 1) Non-negativity constraint means that if the pixel value needs to be non-negative, for any pixel value less than 0, it is set to 0. Range constraint means that for 8-bit grayscale images, all pixel values ​​are ensured to be in the range of 0 to 255. Pixel values ​​outside this range need to be clipped to the nearest boundary value.

[0075] 2) The sum constraint refers to calculating the total brightness of the original image and the filtered image. If the two are not equal, the total brightness can be adjusted by scaling all pixel values ​​of the filtered image.

[0076] 3) Statistical constraints refer to the need to maintain specific statistical properties, which can be satisfied by adjusting each pixel value.

[0077] 4) Texture and detail preservation constraints refer to the use of specific algorithms (such as bilateral filters) to preserve the texture and details of the image while satisfying other constraints.

[0078] 5) Edge preservation constraint refers to applying an edge detection algorithm and then using an edge enhancement or preservation algorithm to ensure that the edge information of the image is not over-smoothed.

[0079] It should be noted that the above constraints need to be determined based on the specific usage scenario, and will not be elaborated here.

[0080] In this embodiment, although the greedy algorithm is used to solve the problem of regional noise reduction, there may still be problems in which the pixel values ​​near the injected edge in the image after noise reduction are too smooth or the pixel values ​​are negative or the image brightness is too high. Based on the above situation, this case chooses to further use the convex set projection algorithm to solve this problem. Specifically, the image after greedy filtering is projected into a set that satisfies all constraints to ensure that each pixel value of the image meets specific requirements. This method may prevent the greedy algorithm from considering the global constraints in the selection of the local optimal solution, and the POCS algorithm can also ensure that the final result meets all preset constraints. This method prevents the above-mentioned problems from occurring in the final output image.

[0081] In this embodiment, in addition to outputting the final image after the constraint set conditions are satisfied or the maximum number of iterations is reached, the final image may also be output after the noise reduction effect is no longer significantly improved.

[0082] In this embodiment, some post-processing steps may be used to further improve the image quality after noise reduction, such as using a sharpening filter to enhance details, etc., which will not be described in detail here.

[0083] Embodiment 2

[0084] See also Figure 2 , a device capable of improving image quality twice, based on the method described in one of the embodiments; comprising:

[0085] The first processing module is used to obtain an original image, and use a convex set projection algorithm to perform super-resolution reconstruction on the original image to obtain a first image;

[0086] A selection module: used for traversing each local pixel area of ​​the image based on the first image, calculating the local variance of each local pixel area by using a greedy algorithm, and selecting a corresponding filter according to the local variance;

[0087] The second processing module is used to perform noise reduction processing on each pixel area of ​​the first image according to the selected filter, and output a second image.

[0088] In this embodiment, in order to better use the method described in one of the embodiments, the present application proposes a device that can improve image quality twice, and each module corresponds to each step of the above method. The specific principles have been described above and will not be repeated here.

[0089] Embodiment 3

[0090] A device for improving image quality twice, comprising:

[0091] at least one processor and a memory communicatively coupled to the at least one processor;

[0092] The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the method described in one of the embodiments.

[0093] In this embodiment, in order to better run and process the method described in one of the embodiments, the above method is stored in a memory, and a processor is used to execute the stored method. It should be noted that the principle and effect of each step have been described above and will not be described in detail here.

[0094] Embodiment 4

[0095] A computer-readable storage medium stores a computer program, wherein when the computer program is executed by a processor, the method described in one of the embodiments is implemented.

[0096] In this embodiment, in order to better run and use the method described in one of the embodiments, the method is stored in a computer-readable storage medium and implemented by a processor. It should be noted that the principle and effect of each step have been described above and will not be described in detail here.

[0097] The above descriptions are only some embodiments of the present invention, and are not intended to limit the protection scope of the present invention. Any equivalent device or equivalent process transformation made using the contents of the present invention specification and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present invention.

Claims

1. A method for improving image quality twice, characterized in that: include: Acquire an original image, and use a convex set projection algorithm to perform super-resolution reconstruction on the original image to obtain a first image; Based on the first image, traverse each local pixel area of ​​the image, calculate the local variance of each local pixel area by using a greedy algorithm and select a corresponding filter according to the local variance; A noise reduction process is performed on each pixel area of ​​the first image according to the selected filter, and a second image is output.

2. A method for improving image quality twice as claimed in claim 1, characterized in that: The original image is obtained, and a convex set projection algorithm is used to perform super-resolution reconstruction on the original image to obtain a first image. Specifically: When using the convex set projection algorithm, bilinear interpolation is used for upsampling.

3. The method for improving image quality twice as claimed in claim 1, characterized in that: The range of the local pixel area is i*n; Among them, i is 1 / 5-1 / 10 of the number of pixels in the length direction of the original image, n is 1 / 5-1 / 10 of the number of pixels in the width direction of the original image, and i and n are rounded up.

4. The method for improving image quality twice as claimed in claim 1, characterized in that: The greedy algorithm is used to calculate the local variance of each local pixel area and select the corresponding filter according to the local variance. Specifically: According to the filter characteristics, it is divided into smoothing filter group and edge detail filter group; If the local variance is greater than the threshold, the filters in the edge detail filter group are applied one by one, and each filter is evaluated using a greedy algorithm and one is selected for application; Otherwise, the filters in the smoothing filter bank are applied one by one, each filter is evaluated using a greedy algorithm, and one is selected for application.

5. A method for improving image quality twice as claimed in claim 1 or 4, characterized in that: The filters include a mean filter, a Gaussian filter, and a median filter.

6. The method for improving image quality twice as claimed in claim 1, characterized in that: Perform noise reduction processing on each pixel area of ​​the first image according to the selected filter, and output a second image, specifically: Set several constraint sets; Project the pixel values ​​in the second image into the constraint set. If the constraint set conditions are not met, then Based on the second image, traverse each local pixel area of ​​the image, calculate the local variance of each local pixel area by using a greedy algorithm and select a corresponding filter according to the local variance; Perform noise reduction processing on each pixel area of ​​the second image according to the selected filter, output a third image, and repeat the above steps until the constraint set conditions are met or the maximum number of iterations is reached, and then output a final image.

7. The method for improving image quality twice as claimed in claim 6, characterized in that: The plurality of constraint sets include: non-negativity constraint, range constraint, sum constraint, statistical constraint, texture and detail preservation constraint, and edge preservation constraint.

8. A device capable of improving image quality twice, characterized in that: A method according to any one of claims 1 to 7; comprising: The first processing module is used to obtain an original image, and use a convex set projection algorithm to perform super-resolution reconstruction on the original image to obtain a first image; A selection module: used for traversing each local pixel area of ​​the image based on the first image, calculating the local variance of each local pixel area by using a greedy algorithm, and selecting a corresponding filter according to the local variance; The second processing module is used to perform noise reduction processing on each pixel area of ​​the first image according to the selected filter, and output a second image.

9. A device capable of improving image quality twice, characterized in that: include: at least one processor and a memory communicatively coupled to the at least one processor; The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the method as claimed in any one of claims 1 to 7.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the method according to any one of claims 1 to 7 is implemented.

Citation Information

Patent Citations

  • Image stripe noise removal method based on compressed sensing

    CN110298803A

  • Image division method and device based on guide filter and electronic device

    CN110766634A

Cited By

  • Video noise reduction method based on hybrid filtering

    CN120219228A