Image processing method, device, electronic device, and storage medium

Abstract

The present disclosure relates to an image processing method, device, electronic device, and storage medium, the method comprising: performing at least two reduction operations based on an original image to obtain an image sequence including the original image; performing blemish removal processing on the images in the image sequence in reverse order until a blemish-removed image corresponding to the original image is obtained; wherein, for the i-th image in the image sequence, the following blemish removal processing is performed: performing blemish removal processing on the low-frequency image corresponding to the i-th image to obtain a low-frequency blemish-removed image; performing blemish removal processing on the high-frequency image corresponding to the i-th image to obtain a high-frequency blemish-removed image; fusing the low-frequency blemish-removed image and the high-frequency blemish-removed image to obtain a blemish-removed image corresponding to the i-th image; and using the blemish-removed image corresponding to the i-th image as the low-frequency image corresponding to the i-1-th image in the image sequence. The method can effectively eliminate facial blemishes, retain the texture of facial skin, and improve the blemish removal effect of the image.

Description

Technical Field

[0001] The present disclosure relates to the field of image technology, and in particular to an image processing method, device, electronic device, storage medium and computer program product. Background Art

[0002] With the development of image processing technology, beautification functions such as skin smoothing and blemish removal have emerged. For example, it can remove spots, acne, moles, wrinkles, dark circles, tear troughs, blackheads, pores, etc. on the facial skin. However, using traditional blemish removal methods to process images will make the overall picture unclear and the blemish removal effect insufficient. The skin texture of the face itself will also be lost due to excessive skin smoothing.

[0003] Therefore, the related art has the problem of poor effect of image processing for blemish removal. Summary of the invention

[0004] The present disclosure provides an image processing method, device, electronic device, storage medium and computer program product to at least solve the problem of poor image processing effect of defect removal in the related art. The technical solution of the present disclosure is as follows:

[0005] According to a first aspect of an embodiment of the present disclosure, there is provided an image processing method, including:

[0006] Perform at least two reduction operations on the original image to obtain an image sequence including the original image; the images in the image sequence are arranged in descending order according to size;

[0007] Performing defect removal processing on the images in the image sequence in reverse order until a defect-removed image corresponding to the original image is obtained;

[0008] The following defect removal process is performed for the i-th image in the image sequence:

[0009] Performing defect removal processing on the low-frequency image corresponding to the i-th image to obtain a low-frequency defect-removed image;

[0010] Performing defect removal processing on the high-frequency image corresponding to the i-th image to obtain a high-frequency defect-removed image;

[0011] Fusing the low-frequency defect-removed image and the high-frequency defect-removed image to obtain a defect-removed image corresponding to the i-th image;

[0012] The blemish-removed image corresponding to the i-th image is used as the low-frequency image corresponding to the i-1-th image in the image sequence, where i is a positive integer greater than 1 and less than N.

[0013] In a possible implementation, performing at least two reduction operations based on the original image to obtain an image sequence including the original image includes:

[0014] Acquire an original image, perform N downsampling operations on the original image, and obtain N reduced images of different sizes, where N is greater than or equal to 2; the downsampling operation is performed based on preset size reduction multiple information;

[0015] The original image is taken as the largest size, the reduced image obtained by the Nth downsampling operation is taken as the smallest size, and the N reduced images of different sizes and the original image are arranged in descending order according to their sizes to obtain the image sequence.

[0016] In a possible implementation, performing defect removal processing on the low-frequency image corresponding to the i-th image to obtain the low-frequency defect-removed image includes:

[0017] The low-frequency image corresponding to the i-th image is input into the defect removal network to obtain the low-frequency defect-removed image corresponding to the i-th image.

[0018] In a possible implementation, performing defect removal processing on the high-frequency image corresponding to the i-th image to obtain the high-frequency defect-removed image includes:

[0019] Using the i+1th image as the initial low-frequency image of the i-th image;

[0020] Obtaining a high-frequency image corresponding to the ith image according to the ith image and the initial low-frequency image;

[0021] The high-frequency image is input into a defect removal network to obtain a high-frequency defect removal image corresponding to the i-th image.

[0022] In a possible implementation manner, obtaining a high-frequency image corresponding to the i-th image according to the i-th image and the initial low-frequency image includes:

[0023] Performing an upsampling operation on the initial low-frequency image to obtain an upsampling result corresponding to the initial low-frequency image; the upsampling operation is performed based on a preset upsampling multiple;

[0024] A high-frequency image corresponding to the ith image is obtained according to upsampling results corresponding to the ith image and the initial low-frequency image.

[0025] In a possible implementation, the fusing the low-frequency defect-removed image and the high-frequency defect-removed image to obtain the defect-removed image corresponding to the i-th image includes:

[0026] Performing an upsampling operation on the low-frequency defect-removed image to obtain an upsampling result corresponding to the low-frequency defect-removed image; the upsampling operation is performed based on a preset upsampling multiple;

[0027] The upsampling result corresponding to the low-frequency de-defected image and the high-frequency de-defected image are fused to obtain a de-defected image corresponding to the i-th image.

[0028] According to a second aspect of an embodiment of the present disclosure, there is provided an image processing apparatus, including:

[0029] an image sequence obtaining unit, configured to perform at least two reduction operations based on the original image to obtain an image sequence including the original image; the images in the image sequence are arranged in descending order according to size;

[0030] a defect removal processing unit configured to perform defect removal processing on the images in the image sequence in reverse order until a defect-removed image corresponding to the original image is obtained;

[0031] The following defect removal process is performed for the i-th image in the image sequence:

[0032] Performing defect removal processing on the low-frequency image corresponding to the i-th image to obtain a low-frequency defect-removed image;

[0033] Performing defect removal processing on the high-frequency image corresponding to the i-th image to obtain a high-frequency defect-removed image;

[0034] Fusing the low-frequency defect-removed image and the high-frequency defect-removed image to obtain a defect-removed image corresponding to the i-th image;

[0035] The blemish-removed image corresponding to the i-th image is used as the low-frequency image corresponding to the i-1-th image in the image sequence, where i is a positive integer greater than 1 and less than N.

[0036] In one possible implementation, the image sequence obtaining unit is specifically configured to obtain an original image, perform N downsampling operations on the original image, and obtain N reduced images of different sizes, where N ≥ 2; the downsampling operation is performed based on preset size reduction multiple information; the original image is taken as the maximum size, the reduced image obtained by the Nth downsampling operation is taken as the minimum size, and the N reduced images of different sizes and the original image are arranged in descending order according to size to obtain the image sequence.

[0037] In a possible implementation, the defect removal processing unit is specifically configured to input the low-frequency image corresponding to the i-th image into a defect removal network to obtain a low-frequency defect removal image corresponding to the i-th image.

[0038] In a possible implementation, the blemish removal processing unit is specifically configured to execute the steps of taking the i+1th image as an initial low-frequency image of the i-th image; obtaining a high-frequency image corresponding to the i-th image according to the i-th image and the initial low-frequency image; and inputting the high-frequency image into the blemish removal network to obtain a high-frequency blemish removal image corresponding to the i-th image.

[0039] In a possible implementation, the blemish removal processing unit is specifically configured to perform an upsampling operation on the initial low-frequency image to obtain an upsampling result corresponding to the initial low-frequency image; the upsampling operation is performed based on a preset upsampling multiple; and a high-frequency image corresponding to the i-th image is obtained according to the upsampling results corresponding to the i-th image and the initial low-frequency image.

[0040] In a possible implementation, the blemish removal processing unit is further configured to perform an upsampling operation on the low-frequency blemish removal image to obtain an upsampling result corresponding to the low-frequency blemish removal image; the upsampling operation is performed based on a preset upsampling multiple; and the upsampling result corresponding to the low-frequency blemish removal image and the high-frequency blemish removal image are merged to obtain a blemish removal image corresponding to the i-th image.

[0041] According to a third aspect of an embodiment of the present disclosure, there is provided an electronic device, including:

[0042] processor;

[0043] a memory for storing instructions executable by the processor;

[0044] The processor is configured to execute the instructions to implement the image processing method as described in any one of the above items.

[0045] According to a fourth aspect of an embodiment of the present disclosure, a computer-readable storage medium is provided. When instructions in the computer-readable storage medium are executed by a processor of an electronic device, the electronic device is enabled to perform the image processing method as described in any one of the above items.

[0046] According to a fifth aspect of an embodiment of the present disclosure, a computer program product is provided, wherein the computer program product includes instructions, and when the instructions are executed by a processor of an electronic device, the electronic device is enabled to perform the image processing method as described in any one of the above items.

[0047] The technical solution provided by the embodiments of the present disclosure brings at least the following beneficial effects:

[0048] The scheme disclosed herein performs at least two reduction operations based on the original image to obtain an image sequence including the original image, wherein the images in the image sequence are arranged in descending order according to the size, and then the images in the image sequence are sequentially subjected to the defect removal process in reverse order until the defect removal image corresponding to the original image is obtained, wherein the following defect removal process is performed for the i-th image in the image sequence: the low-frequency image corresponding to the i-th image is subjected to the defect removal process to obtain the low-frequency defect removal image, the high-frequency image corresponding to the i-th image is subjected to the defect removal process to obtain the high-frequency defect removal image, the low-frequency defect removal image and the high-frequency defect removal image are fused to obtain the defect removal image corresponding to the i-th image, and the defect removal image corresponding to the i-th image is used as the low-frequency image corresponding to the i-1-th image in the image sequence, where i is a positive integer greater than 1 and less than N. In this way, the facial defects in the image can be separated into different frequencies of different sizes for processing, which is more targeted, can effectively eliminate facial defects, and retain the texture of the facial skin to a great extent, thereby improving the defect removal effect of the image.

[0049] It is to be understood that the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the present disclosure. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] The drawings herein are incorporated into and constitute a part of the specification, illustrate embodiments consistent with the present disclosure, and together with the description are used to explain the principles of the present disclosure, and do not constitute improper limitations on the present disclosure.

[0051] Figure 1 The figure is a flowchart of an image processing method according to an exemplary embodiment.

[0052] Figure 2 It is a schematic diagram of an overall framework of image processing according to an exemplary embodiment.

[0053] Figure 3a It is a schematic diagram showing a comparison diagram of image processing effects according to an exemplary embodiment.

[0054] Figure 3b It is a schematic diagram of another image processing effect comparison diagram according to an exemplary embodiment.

[0055] Figure 4a It is a schematic diagram showing a target texture filtering network structure according to an exemplary embodiment.

[0056] Figure 4b It is a schematic diagram showing a network module structure according to an exemplary embodiment.

[0057] Figure 5 is a flowchart of another image processing method according to an exemplary embodiment.

[0058] Figure 6 is a block diagram of an image processing apparatus according to an exemplary embodiment.

[0059] Figure 7 The figure is a diagram showing the internal structure of an electronic device according to an exemplary embodiment. DETAILED DESCRIPTION

[0060] In order to enable ordinary persons in the art to better understand the technical solutions of the present disclosure, the technical solutions in the embodiments of the present disclosure will be clearly and completely described below in conjunction with the accompanying drawings.

[0061] It should be noted that the terms "first", "second", etc. in the specification and claims of the present disclosure and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present disclosure described herein can be implemented in an order other than those illustrated or described herein. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the present disclosure.

[0062] It should also be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for display, data for analysis, etc.) involved in this disclosure are all information and data authorized by the user or fully authorized by all parties.

[0063] Figure 1 is a flowchart of an image processing method according to an exemplary embodiment. Figure 1 As shown, the method can be used in computer devices such as terminals, which can be but not limited to various personal computers, laptops, smart phones, tablet computers, etc., and includes the following steps.

[0064] In step S110, at least two reduction operations are performed based on the original image to obtain an image sequence including the original image; the images in the image sequence are arranged in descending order according to size;

[0065] As an example, the original image may be a face image to be blemish-free, for example, Figure 2 The input image.

[0066] Among them, the resolution corresponding to the original image can be HxW, and at least two images with different resolutions can be generated based on the original image, such as an image with 1 / 2 resolution of the original image, that is, an image with a resolution of H / 2xW / 2, and an image with 1 / 4 resolution of the original image, that is, an image with a resolution of H / 4xW / 4. It can also be other resolutions, which are not specifically limited in this embodiment.

[0067] In practical applications, the facial image to be blemish-free can be used as the original image, and then the original image can be reduced at least twice to obtain an image sequence including the original image. The images in the image sequence can be arranged in descending order according to size, and there can be a multiple relationship between adjacent sizes, so as to further perform blemish removal processing based on the image sequence, so that through the high and low frequency separation technology, targeted blemish removal processing can be performed on images of different resolutions.

[0068] In one example, if Figure 2 As shown, Figure 2 The input image can be the original image. By separating the high and low frequencies of the original image, at least two images with different resolutions can be obtained, such as Figure 2 The i-1th image, the i-th image, and the N-th image are included, wherein the resolution of the i-1th image is H / 2xW / 2, the resolution of the i-th image is H / 4xW / 4, and the resolution of the N-th image is H / 8xW / 8.

[0069] In another example, when high and low frequencies are separated from the original image, a low frequency image for the original image can be obtained by downsampling the original image, that is, an image obtained by performing at least two reduction operations on the original image, such as Figure 2 By downsampling the input image, the i-1th image, the i-th image, and the Nth image can be obtained. Other high- and low-frequency separation methods can also be used, which are not specifically limited in this embodiment.

[0070] In step S120, the images in the image sequence are subjected to a blemish removal process in reverse order until a blemish-removed image corresponding to the original image is obtained;

[0071] Among them, for the i-th image in the image sequence, the following defect removal processing is performed:

[0072] Performing defect removal processing on a low-frequency image corresponding to the i-th image to obtain a low-frequency defect-removed image; performing defect removal processing on a high-frequency image corresponding to the i-th image to obtain a high-frequency defect-removed image; fusing the low-frequency defect-removed image and the high-frequency defect-removed image to obtain a defect-removed image corresponding to the i-th image;

[0073] The blemish-removed image corresponding to the i-th image is used as the low-frequency image corresponding to the i-1-th image in the image sequence, and i is a positive integer greater than 1 and less than N.

[0074] As an example, through the defect removal process, the target texture in the low-frequency image can be filtered out to obtain a low-frequency defect-free image, and the target texture in the high-frequency image can be filtered out to obtain a high-frequency defect-free image, such as Figure 2As shown, the de-defected image corresponding to the i-th image can be used as the low-frequency image corresponding to the i-1-th image, and the low-frequency image can be processed by a convolutional neural network to obtain a low-frequency de-defected image corresponding to the i-1-th image; for the high-frequency image corresponding to the i-1-th image, the convolutional neural network can also be used to perform de-defect processing on the high-frequency image to obtain a high-frequency de-defected image corresponding to the i-1-th image.

[0075] After obtaining the image sequence, the images in the image sequence can be subjected to a reverse blemish removal process in sequence, such as by calculating from a low resolution bottom to top, until a blemish-removed image corresponding to the original image is obtained, such as by calculating from a low resolution bottom to top, such as Figure 2 The output image.

[0076] For example, spots, acne, moles, wrinkles, dark circles, tear troughs, blackheads, pores and other blemishes on the facial skin of a face image can be used as target textures. Through the blemish removal processing of the convolutional neural network, blemishes of varying degrees can be removed to achieve the skin smoothing and blemish removal function for face images.

[0077] In one example, if Figure 2 As shown, when high and low frequencies are separated for the i-1th image, a low frequency image for the i-1th image, i.e., the i-th image, can be obtained by downsampling the i-1th image, and then an upsampling result of the low frequency image, i.e., the upsampling result of the i-th image, can be obtained. Furthermore, a difference operation can be performed to subtract the upsampling result of the corresponding low frequency image from the i-1th image to obtain a high frequency image for the i-th image, i.e., the high frequency image corresponding to the i-1th image. Other high and low frequency separation methods can also be used, which are not specifically limited in this embodiment.

[0078] In another example, Figure 2 As shown, the convolutional neural network output result of the high-frequency image corresponding to the i-1th image can be used as the high-frequency defect-removing image, and then the upsampling result of the low-frequency defect-removing image corresponding to the i-1th image can be superimposed to obtain the defect-removing image corresponding to the i-1th image; or through a summation operation, the high-frequency defect-removing image corresponding to the i-1th image and the upsampling result of the low-frequency defect-removing image corresponding to the i-1th image can be summed to obtain the defect-removing image corresponding to the i-1th image.

[0079] Compared with the traditional method, the encoder-decoder network structure is used to learn the blemish removal task end-to-end, but this method does not take into account the frequency division processing of the human face skin, which will lead to the loss of high-frequency texture on the human face skin, and the final result after processing has poor texture. The technical solution of this embodiment, by combining the high-low frequency separation method with the convolutional neural network (CNN) in the field of deep learning, can perform face image blemish removal based on multi-scale frequency separation, and perform blemish removal at different frequencies at different scales (such as different resolutions), such as using high-low frequency separation at scales of 1 / 2, 1 / 4, and 1 / 8 of the original image resolution. For each scale, the low-frequency image corresponding to the scale can be input into the convolutional neural network for processing, and then the output result can be upsampled and fused with the high-frequency image corresponding to the scale, and then the final output image is obtained by bottom-up calculation from low resolution, so that the blemish removal can be cleaner, and the texture of the skin can be retained without degrading the image quality. Moreover, by separating the blemishes of different degrees for separate processing, more and more complex scenes and blemish types can be adapted. like Figure 3a As shown, it can characterize the image blemish removal effect of the technical solution of this embodiment ( Figure 3a 1 is the original image, 2 is the de-defected image), or Figure 3b As shown ( Figure 3b 1 is the original image, and 2 is the de-defected image).

[0080] In the above image processing method, by performing at least two reduction operations based on the original image, an image sequence including the original image is obtained, and the images in the image sequence are arranged in descending order according to the size, and then the images in the image sequence are sequentially subjected to the defect removal process in reverse order until the defect removal image corresponding to the original image is obtained, wherein the following defect removal process is performed for the i-th image in the image sequence: the low-frequency image corresponding to the i-th image is subjected to the defect removal process to obtain the low-frequency defect removal image, the high-frequency image corresponding to the i-th image is subjected to the defect removal process to obtain the high-frequency defect removal image, the low-frequency defect removal image and the high-frequency defect removal image are fused to obtain the defect removal image corresponding to the i-th image, and the defect removal image corresponding to the i-th image is used as the low-frequency image corresponding to the i-1-th image in the image sequence, where i is a positive integer greater than 1 and less than N. In this way, the facial defects in the image can be separated into different frequencies of different sizes for processing, which is more targeted, can effectively eliminate facial defects, and retain the texture of the facial skin to a great extent, thereby improving the defect removal effect of the image.

[0081] In an exemplary embodiment, at least two reduction operations are performed based on the original image to obtain an image sequence including the original image, including: acquiring the original image, performing N downsampling operations on the original image to obtain N reduced images of different sizes, N ≥ 2; the downsampling operation is performed based on preset size reduction multiple information; the original image is used as the maximum size, the reduced image obtained by the Nth downsampling operation is used as the minimum size, and the N reduced images of different sizes and the original image are arranged in descending order according to size to obtain an image sequence.

[0082] As an example, the preset downsizing factor information may be used to indicate a resolution of 1 / 2 of the original image resolution, a resolution of 1 / 4 of the original image resolution, and a resolution of 1 / 8 of the original image resolution.

[0083] In practical applications, a facial image to be blemish-free can be obtained as an original image, and then the original image can be downsampled N times to obtain N reduced images of different sizes, N ≥ 2, and then the original image can be used as the largest size, and the reduced image obtained by the Nth downsampling operation can be used as the smallest size. The N reduced images of different sizes and the original image are arranged in descending order according to the size to obtain an image sequence, and adjacent sizes can have a multiple relationship, so as to further perform blemish removal processing based on the image sequence, and perform blemish removal processing specifically for images of different sizes.

[0084] For example, Figure 2 As shown, the input image can be the original image. By performing N downsampling operations on the input image, N reduced images of different sizes can be obtained, such as the i-1th image, the i-th image, and the N-th image. Then, with the input image as the largest size and the N-th image as the smallest size, they can be arranged in descending order according to size to obtain an image sequence including the input image, the i-1th image, the i-th image, and the N-th image.

[0085] In an optional embodiment, by obtaining preset size reduction factor information, it can indicate the number and combination of resolutions, and then the resolution values ​​corresponding to N different resolutions, that is, N different sizes, can be determined to obtain N reduced images of different sizes, so that by presetting the specified number and combination of resolutions, the multi-scale frequency separation form can be expanded.

[0086] The technical solution of this embodiment obtains an original image, performs N downsampling operations on the original image, obtains N reduced images of different sizes, and then takes the original image as the largest size and the reduced image obtained by the Nth downsampling operation as the smallest size, arranges the N reduced images of different sizes and the original image in descending order according to the size, and obtains an image sequence, which provides data support for subsequent blemish removal processing on images of different sizes.

[0087] In an exemplary embodiment, performing blemish removal processing on a low-frequency image corresponding to the ith image to obtain a low-frequency blemish-removed image includes: inputting the low-frequency image corresponding to the ith image into a blemish removal network to obtain a low-frequency blemish-removed image corresponding to the ith image.

[0088] In a specific implementation, the low-frequency image corresponding to the i-th image can be input into the defect removal network, such as Figure 2 The convolutional neural network can then use the output of the defect removal network as the low-frequency defect removal image corresponding to the i-th image.

[0089] For example, Figure 2 As shown, the blemish-removed image corresponding to the i-th image can be used as the low-frequency image corresponding to the i-1-th image, and the low-frequency image can be blemish-removed using a convolutional neural network to obtain a low-frequency blemish-removed image corresponding to the i-1-th image.

[0090] The technical solution of this embodiment obtains a low-frequency defect-removed image corresponding to the ith image by inputting the low-frequency image corresponding to the ith image into the defect removal network. Defect removal processing can be performed on the low frequencies of different scales based on the original image, thereby providing data support for the subsequent fusion of high- and low-frequency defect-removed images.

[0091] In an exemplary embodiment, a high-frequency image corresponding to an i-th image is subjected to defect removal processing to obtain a high-frequency defect-removed image, including: using an i+1-th image as an initial low-frequency image of an i-th image; obtaining a high-frequency image corresponding to the i-th image according to the i-th image and the initial low-frequency image; and inputting the high-frequency image into a defect removal network to obtain a high-frequency defect-removed image corresponding to the i-th image.

[0092] In practical applications, such as Figure 2 As shown, when the i-th image is separated into high and low frequencies, the i-th image can be downsampled to obtain a low-frequency image for the i-th image, that is, the i+1-th image, and then the upsampling result of the low-frequency image can be obtained, that is, the upsampling result of the i+1-th image. The i-th image can be subtracted from the upsampling result of the corresponding low-frequency image by performing a difference operation to obtain a high-frequency image corresponding to the i-th image, and then the high-frequency image can be input into the defect removal network, as shown in Figure 2In the convolutional neural network, the output of the defect removal network is used as the high-frequency defect removal image corresponding to the i-th image.

[0093] The technical solution of this embodiment uses the i+1th image as the initial low-frequency image of the i-th image, and then obtains the high-frequency image corresponding to the i-th image based on the i-th image and the initial low-frequency image, and then inputs the high-frequency image into the defect removal network to obtain the high-frequency defect-removed image corresponding to the i-th image. Defect removal can be performed on the high frequencies of different scales based on the original image, providing data support for the subsequent fusion of high and low frequency defect-removed images.

[0094] In an exemplary embodiment, obtaining a high-frequency image corresponding to the ith image according to the ith image and an initial low-frequency image includes: performing an upsampling operation on the initial low-frequency image to obtain an upsampling result corresponding to the initial low-frequency image; performing the upsampling operation based on a preset upsampling multiple; and obtaining a high-frequency image corresponding to the ith image according to the upsampling results corresponding to the ith image and the initial low-frequency image.

[0095] As an example, the preset upsampling multiple may be a 2x upsampling operation.

[0096] In one example, if Figure 2 As shown, when high and low frequencies are separated for the ith image, a low-frequency image for the ith image, i.e., the i+1th image, can be obtained by downsampling the ith image, and then an upsampling result of the low-frequency image, i.e., the upsampling result of the i+1th image, can be obtained. Furthermore, a difference operation can be performed to subtract the upsampling result of the corresponding low-frequency image from the ith image to obtain a high-frequency image corresponding to the ith image.

[0097] The technical solution of this embodiment obtains an upsampling result corresponding to the initial low-frequency image by performing an upsampling operation on the initial low-frequency image, and then obtains a high-frequency image corresponding to the ith image based on the upsampling result corresponding to the ith image and the initial low-frequency image. The high-frequency image can be obtained based on high- and low-frequency separation, providing data support for subsequent high-frequency image defect removal processing.

[0098] In an exemplary embodiment, a low-frequency blemish-removed image and a high-frequency blemish-removed image are fused to obtain a blemish-removed image corresponding to the i-th image, including: performing an upsampling operation on the low-frequency blemish-removed image to obtain an upsampling result corresponding to the low-frequency blemish-removed image; performing the upsampling operation based on a preset upsampling multiple; and fusing the upsampling result corresponding to the low-frequency blemish-removed image and the high-frequency blemish-removed image to obtain the blemish-removed image corresponding to the i-th image.

[0099] As an example, the preset upsampling multiple may be a 2x upsampling operation.

[0100] In one example, an upsampling operation may be performed on the low-frequency blemish-removed image corresponding to the ith image to obtain an upsampling result corresponding to the low-frequency blemish-removed image, and then the high-frequency blemish-removed image corresponding to the ith image may be superimposed thereon to fuse the upsampling result corresponding to the low-frequency blemish-removed image and the high-frequency blemish-removed image to obtain the blemish-removed image corresponding to the ith image; or a summation operation may be performed to obtain the blemish-removed image corresponding to the ith image by summing up the high-frequency blemish-removed image corresponding to the ith image and the upsampling result of the low-frequency blemish-removed image corresponding to the ith image.

[0101] The technical solution of this embodiment obtains an upsampling result corresponding to the low-frequency blemish-removed image by upsampling the low-frequency blemish-removed image, and then fuses the upsampling result corresponding to the low-frequency blemish-removed image and the high-frequency blemish-removed image to obtain a blemish-removed image corresponding to the ith image. The final blemish-removed image can be obtained based on the result of blemish removal processing at high and low frequencies of different scales of the original image. The blemish-removing image is more targeted, can effectively eliminate facial blemishes, and greatly retains the texture of facial skin, thereby improving the blemish removal effect of the image.

[0102] In an exemplary embodiment, the method further includes: inputting the Nth image into a defect removal network to obtain a defect removal image corresponding to the Nth image as a low-frequency image corresponding to the N-1th image.

[0103] In practical applications, for the image with the smallest size in the image sequence, such as Figure 2 The Nth image in the image can be used as the initial low-frequency image corresponding to the N-1th image, and then the initial low-frequency image can be input into the defect removal network, and the result output by the defect removal network (such as Figure 2 The defect-removed image corresponding to the Nth image in the image is up-sampled to obtain a low-frequency image corresponding to the N-1th image.

[0104] The technical solution of this embodiment is to input the Nth image into the defect removal network to obtain the defect removal image corresponding to the Nth image as the low-frequency image corresponding to the N-1th image. When the initial low-frequency image is of the minimum size, the initial low-frequency image can be subjected to corresponding defect removal processing.

[0105] In an exemplary embodiment, a defect removal network (such as Figure 2The convolutional neural network (CNN) may include multiple network modules connected in series, each of which may include a separation channel processing layer and a channel result aggregation layer. The separation channel processing layer and the channel result aggregation layer are connected in series. By obtaining the image to be input of the current network module, the image to be input can be input into the separation channel processing layer to obtain an image separation processing result, and then the image separation processing result can be input into the channel result aggregation layer to obtain an image aggregation processing result, and then the image to be input and the image aggregation processing result can be fused to obtain the output image of the current network module, and the output image is used as the image to be input to the next network module of the current network module, and the next network module is used as the current network module, and the step of obtaining the image to be input to the current network module is returned until the network output result is obtained after the last network module is processed.

[0106] Among them, the image separation processing result can be obtained by filtering out the target textures in different channels of the input image; the image aggregation processing result can be obtained by aggregating the results of filtering out the target textures in different channels based on the image separation processing result.

[0107] As an example, a defect removal network (such as Figure 2 Convolutional neural networks can include multiple network modules connected in series, such as Figure 4a As shown, the network module can be a bottleneck-depthwise-conv-module. Compared with the traditional CNN module, it can achieve similar performance to the traditional CNN module with extremely low computational complexity, thereby constructing a target texture filtering network based on the bottleneck-depthwise-conv-module, which can enable the image blemish removal process to run in real time at a speed of 30fps in a mobile terminal.

[0108] In one example, the structures of the defect removal networks are the same, such as Figure 2 As shown in the figure, each convolutional neural network is composed of n bottleneck depth separable modules (i.e., network modules). Based on the bottleneck depth separable modules, the amount of calculation is minimized while achieving the same effect. The bottleneck depth separable module mainly uses separable convolution (depth-wise conv) and bottleneck structure. Its network structure is as follows Figure 4b As shown in the figure, the bottleneck depth separable module can be composed of separable convolution, ordinary convolution and short-circuit module. Among them, the separable convolution (i.e., separation channel processing layer) can greatly reduce the amount of calculation, the ordinary convolution (i.e., channel result aggregation layer) can establish connections between different channels of features obtained by the separable convolution processing of the current network module, and the short-circuit module can fuse the underlying features and high-level features. For example, the underlying features are the input features of the current network module. Figure 4bThe leftmost arrow is the input object, and the high-level features are the features obtained by ordinary convolution processing.

[0109] In practical applications, each bottleneck depth separable module (i.e., network module) can process features of a specific scale and a specific semantic level, such as processing high and low frequency features of different sizes of images. Based on each bottleneck depth separable module (i.e., current network module), the input image can be processed by separable convolution (i.e., separation channel processing layer) to remove defects in different channels at the current scale and semantic level in the image, and then the features obtained by the separable convolution processing (i.e., image aggregation processing results) can be processed by ordinary convolution kernels (i.e., channel result aggregation layer) to aggregate the results after defect processing of different channels, and then the bottom-level features (i.e., the input image) and high-level features (i.e., image aggregation processing results) can be fused through the short-section module to obtain the output image of the current network module.

[0110] In one example, since defects of different degrees are at different scales and frequencies, multi-scale frequency division processing can be used to separate the defects into different scales and frequencies for processing, which is more targeted, thereby decoupling the functions of different convolutional neural networks from each other.

[0111] In another example, since blemish removal is a computationally intensive task, the input image and output image processed by the network need to have consistent resolution. In addition, in order to ensure that the face resolution is high and has texture, the resolution of the input image input to the network is relatively large, which will result in high computational overhead for the entire network. The traditional convolutional network uses a basic-block module, which requires two 3x3 convolution stacks. The total computational cost is 2*in*in*3*3*H*W. For devices with low computing power such as mobile phones, traditional methods will cause huge delays and cannot achieve real-time processing.

[0112] Compared with the traditional method, the technical solution of this embodiment is based on a bottleneck depth separable module, which may include a 3x3 depth-wise convolution with an output of in / 4, and a 1x1 ordinary convolution with an input of in / 4. The total computational complexity is in / 4*3*3*H*W+in / 4*in*H*W, which is approximately 1 / 72 of the traditional convolution module. This can greatly reduce the computational complexity while maintaining the processing effect, improve the processing speed, and ensure that the defect removal task can run in real time on the mobile phone.

[0113] Figure 5 is a flowchart of another image processing method according to an exemplary embodiment. Figure 5 As shown, the method is used in a computer device such as a terminal, and includes the following steps.

[0114] In step S510, an original image is obtained, and N downsampling operations are performed on the original image to obtain N reduced images of different sizes, where N ≥ 2; the downsampling operation is performed based on preset size reduction multiple information. In step S520, the original image is taken as the largest size, and the reduced image obtained by the Nth downsampling operation is taken as the smallest size. The N reduced images of different sizes and the original image are arranged in descending order according to the size to obtain an image sequence. In step S530, the images in the image sequence are sequentially subjected to defect removal processing in reverse order until a defect removal image corresponding to the original image is obtained; wherein, for the i-th image in the image sequence, the following defect removal processing is performed: the low-frequency image corresponding to the i-th image is subjected to defect removal processing to obtain a low-frequency defect removal image; the high-frequency image corresponding to the i-th image is subjected to defect removal processing to obtain a high-frequency defect removal image; the low-frequency defect removal image and the high-frequency defect removal image are fused to obtain a defect removal image corresponding to the i-th image; wherein the defect removal image corresponding to the i-th image is used as the low-frequency image corresponding to the i-1-th image in the image sequence, and i is a positive integer greater than 1 and less than N. It should be noted that the specific definition of the above steps can refer to the specific definition of an image processing method above, which will not be repeated here.

[0115] It should be understood that, although the various steps in the flowcharts involved in the above-mentioned embodiments are displayed in sequence according to the indication of the arrows, these steps are not necessarily executed in sequence according to the order indicated by the arrows. Unless there is a clear explanation in this article, the execution of these steps does not have a strict order restriction, and these steps can be executed in other orders. Moreover, at least a part of the steps in the flowcharts involved in the above-mentioned embodiments can include multiple steps or multiple stages, and these steps or stages are not necessarily executed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily to be carried out in sequence, but can be executed in turn or alternately with other steps or at least a part of the steps or stages in other steps.

[0116] It can be understood that the same / similar parts between the various embodiments of the above method in this specification can refer to each other, and each embodiment focuses on the differences from other embodiments. For related points, please refer to the description of other method embodiments.

[0117] Based on the same inventive concept, the embodiment of the present disclosure also provides an image processing device for implementing the above-mentioned image processing method.

[0118] Figure 6 FIG. 1 is a block diagram of an image processing device according to an exemplary embodiment. Figure 6 , the device comprises:

[0119] The image sequence obtaining unit 601 is configured to perform at least two reduction operations based on the original image to obtain an image sequence including the original image; the images in the image sequence are arranged in descending order of size;

[0120] The blemish removal processing unit 602 is configured to perform blemish removal processing on the images in the image sequence in reverse order until a blemish removal image corresponding to the original image is obtained; wherein, for the i-th image in the image sequence, the following blemish removal processing is performed: a blemish removal processing is performed on a low-frequency image corresponding to the i-th image to obtain a low-frequency blemish removal image; a blemish removal processing is performed on a high-frequency image corresponding to the i-th image to obtain a high-frequency blemish removal image; the low-frequency blemish removal image and the high-frequency blemish removal image are merged to obtain a blemish removal image corresponding to the i-th image; wherein the blemish removal image corresponding to the i-th image is used as the low-frequency image corresponding to the i-1-th image in the image sequence, and i is a positive integer greater than 1 and less than N.

[0121] In one possible implementation, the image sequence obtaining unit 601 is specifically configured to obtain an original image, perform N downsampling operations on the original image, and obtain N reduced images of different sizes, where N ≥ 2; the downsampling operation is performed based on preset size reduction multiple information; the original image is taken as the maximum size, the reduced image obtained by the Nth downsampling operation is taken as the minimum size, and the N reduced images of different sizes and the original image are arranged in descending order according to size to obtain the image sequence.

[0122] In a possible implementation, the defect removal processing unit 602 is specifically configured to input the low-frequency image corresponding to the i-th image into a defect removal network to obtain a low-frequency defect removal image corresponding to the i-th image.

[0123] In a possible implementation, the blemish removal processing unit 602 is specifically configured to execute the steps of taking the i+1th image as an initial low-frequency image of the i-th image; obtaining a high-frequency image corresponding to the i-th image according to the i-th image and the initial low-frequency image; and inputting the high-frequency image into the blemish removal network to obtain a high-frequency blemish removal image corresponding to the i-th image.

[0124] In a possible implementation, the blemish removal processing unit 602 is further configured to perform an upsampling operation on the initial low-frequency image to obtain an upsampling result corresponding to the initial low-frequency image; the upsampling operation is performed based on a preset upsampling multiple; and a high-frequency image corresponding to the i-th image is obtained according to the upsampling results corresponding to the i-th image and the initial low-frequency image.

[0125] In a possible implementation, the blemish removal processing unit 602 is further configured to perform an upsampling operation on the low-frequency blemish removal image to obtain an upsampling result corresponding to the low-frequency blemish removal image; the upsampling operation is performed based on a preset upsampling multiple; and the upsampling result corresponding to the low-frequency blemish removal image and the high-frequency blemish removal image are merged to obtain a blemish removal image corresponding to the i-th image.

[0126] Regarding the device in the above embodiment, the specific manner in which each module performs operations has been described in detail in the embodiment of the method, and will not be elaborated here.

[0127] Each module in the above image processing device can be implemented in whole or in part by software, hardware or a combination thereof. Each module can be embedded in or independent of a processor in a computer device in the form of hardware, or can be stored in a memory in a computer device in the form of software, so that the processor can call and execute operations corresponding to each module.

[0128] Figure 7 1 is a block diagram of an electronic device 700 for implementing an image processing method according to an exemplary embodiment. For example, the electronic device 700 may be a mobile phone, a computer, a digital broadcast terminal, a messaging device, a game console, a tablet device, a medical device, a fitness device, a personal digital assistant, etc.

[0129] Reference Figure 7 , the electronic device 700 may include one or more of the following components: a processing component 702 , a memory 704 , a power component 706 , a multimedia component 708 , an audio component 710 , an input / output (I / O) interface 712 , a sensor component 714 , and a communication component 716 .

[0130] The processing component 702 generally controls the overall operation of the electronic device 700, such as operations associated with display, phone calls, data communications, camera operations, and recording operations. The processing component 702 may include one or more processors 720 to execute instructions to complete all or part of the steps of the above-mentioned method. In addition, the processing component 702 may include one or more modules to facilitate the interaction between the processing component 702 and other components. For example, the processing component 702 may include a multimedia module to facilitate the interaction between the multimedia component 708 and the processing component 702.

[0131] The memory 704 is configured to store various types of data to support operations on the electronic device 700. Examples of such data include instructions for any application or method operating on the electronic device 700, contact data, phone book data, messages, pictures, videos, etc. The memory 704 may be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as a static random access memory (SRAM), an electrically erasable programmable read-only memory (EEPROM), an erasable programmable read-only memory (EPROM), a programmable read-only memory (PROM), a read-only memory (ROM), a magnetic memory, a flash memory, a magnetic disk, an optical disk, or a graphene memory.

[0132] The power supply component 706 provides power to the various components of the electronic device 700. The power supply component 706 may include a power management system, one or more power supplies, and other components associated with generating, managing, and distributing power to the electronic device 700.

[0133] The multimedia component 708 includes a screen that provides an output interface between the electronic device 700 and the user. In some embodiments, the screen may include a liquid crystal display (LCD) and a touch panel (TP). If the screen includes a touch panel, the screen may be implemented as a touch screen to receive input signals from the user. The touch panel includes one or more touch sensors to sense touch, slide, and gestures on the touch panel. The touch sensor may not only sense the boundaries of the touch or slide action, but also detect the duration and pressure associated with the touch or slide operation. In some embodiments, the multimedia component 708 includes a front camera and / or a rear camera. When the electronic device 700 is in an operating mode, such as a shooting mode or a video mode, the front camera and / or the rear camera may receive external multimedia data. Each front camera and rear camera may be a fixed optical lens system or have a focal length and optical zoom capability.

[0134] The audio component 710 is configured to output and / or input audio signals. For example, the audio component 710 includes a microphone (MIC), and when the electronic device 700 is in an operating mode, such as a call mode, a recording mode, and a voice recognition mode, the microphone is configured to receive an external audio signal. The received audio signal can be further stored in the memory 704 or sent via the communication component 716. In some embodiments, the audio component 710 also includes a speaker for outputting an audio signal.

[0135] I / O interface 712 provides an interface between processing component 702 and peripheral interface modules, such as keyboards, click wheels, buttons, etc. These buttons may include but are not limited to: a home button, a volume button, a start button, and a lock button.

[0136] The sensor assembly 714 includes one or more sensors for providing various aspects of status assessment for the electronic device 700. For example, the sensor assembly 714 can detect the open / closed state of the electronic device 700, the relative positioning of the components, such as the display and keypad of the electronic device 700, and the sensor assembly 714 can also detect the position change of the electronic device 700 or the components of the electronic device 700, the presence or absence of contact between the user and the electronic device 700, the orientation or acceleration / deceleration of the device 700, and the temperature change of the electronic device 700. The sensor assembly 714 may include a proximity sensor configured to detect the presence of nearby objects without any physical contact. The sensor assembly 714 may also include an optical sensor, such as a CMOS or CCD image sensor, for use in imaging applications. In some embodiments, the sensor assembly 714 may also include an acceleration sensor, a gyroscope sensor, a magnetic sensor, a pressure sensor, or a temperature sensor.

[0137] The communication component 716 is configured to facilitate wired or wireless communication between the electronic device 700 and other devices. The electronic device 700 can access a wireless network based on a communication standard, such as WiFi, a carrier network (such as 2G, 3G, 4G or 5G), or a combination thereof. In an exemplary embodiment, the communication component 716 receives a broadcast signal or broadcast-related information from an external broadcast management system via a broadcast channel. In an exemplary embodiment, the communication component 716 also includes a near field communication (NFC) module to facilitate short-range communication. For example, the NFC module can be implemented based on radio frequency identification (RFID) technology, infrared data association (IrDA) technology, ultra-wideband (UWB) technology, Bluetooth (BT) technology and other technologies.

[0138] In an exemplary embodiment, the electronic device 700 may be implemented by one or more application-specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field programmable gate arrays (FPGAs), controllers, microcontrollers, microprocessors, or other electronic components to perform the above methods.

[0139] In an exemplary embodiment, a computer-readable storage medium including instructions is also provided, such as a memory 704 including instructions, and the above instructions can be executed by the processor 720 of the electronic device 700 to perform the above method. For example, the computer-readable storage medium can be a ROM, a random access memory (RAM), a CD-ROM, a magnetic tape, a floppy disk, an optical data storage device, etc.

[0140] In an exemplary embodiment, a computer program product is further provided. The computer program product includes instructions. The instructions can be executed by the processor 720 of the electronic device 700 to complete the above method.

[0141] It should be noted that the above-mentioned devices, electronic devices, computer-readable storage media, computer program products, etc. may also include other implementation methods according to the description of the method embodiments. The specific implementation methods can refer to the description of the relevant method embodiments, which will not be described one by one here.

[0142] Those skilled in the art will readily appreciate other embodiments of the present disclosure after considering the specification and practicing the invention disclosed herein. The present disclosure is intended to cover any variations, uses or adaptations of the present disclosure that follow the general principles of the present disclosure and include common knowledge or customary techniques in the art that are not disclosed in the present disclosure. The description and examples are to be considered exemplary only, and the true scope and spirit of the present disclosure are indicated by the claims.

[0143] It should be understood that the present disclosure is not limited to the exact structures that have been described above and shown in the drawings, and that various modifications and changes may be made without departing from the scope thereof. The scope of the present disclosure is limited only by the appended claims.

Claims

1. An image processing method, characterized in that: The method comprises: Perform at least two reduction operations on the original image to obtain an image sequence including the original image; the images in the image sequence are arranged in descending order according to size; Performing defect removal processing on the images in the image sequence in reverse order until a defect-removed image corresponding to the original image is obtained; The following defect removal process is performed for the i-th image in the image sequence: Performing defect removal processing on the low-frequency image corresponding to the i-th image to obtain a low-frequency defect-removed image; Performing defect removal processing on the high-frequency image corresponding to the ith image to obtain a high-frequency defect-removed image; including: using the i+1th image as the initial low-frequency image of the ith image; obtaining a high-frequency image corresponding to the ith image according to the ith image and the initial low-frequency image; inputting the high-frequency image into a defect removal network to obtain a high-frequency defect-removed image corresponding to the ith image; Fusion of the low-frequency defect-removed image and the high-frequency defect-removed image to obtain a defect-removed image corresponding to the i-th image; The blemish-removed image corresponding to the i-th image is used as the low-frequency image corresponding to the i-1-th image in the image sequence, where i is a positive integer greater than 1 and less than N.

2. The method according to claim 1, characterized in that The performing at least two reduction operations based on the original image to obtain an image sequence including the original image comprises: Acquire an original image, perform N downsampling operations on the original image, and obtain N reduced images of different sizes, where N is greater than or equal to 2; the downsampling operation is performed based on preset size reduction multiple information; The original image is taken as the largest size, the reduced image obtained by the Nth downsampling operation is taken as the smallest size, and the N reduced images of different sizes and the original image are arranged in descending order according to their sizes to obtain the image sequence.

3. The method according to claim 1, characterized in that: The step of performing defect removal processing on the low-frequency image corresponding to the i-th image to obtain a low-frequency defect-removed image includes: The low-frequency image corresponding to the i-th image is input into the defect removal network to obtain the low-frequency defect-removed image corresponding to the i-th image.

4. The method according to claim 1, characterized in that The obtaining, according to the i-th image and the initial low-frequency image, a high-frequency image corresponding to the i-th image comprises: Performing an upsampling operation on the initial low-frequency image to obtain an upsampling result corresponding to the initial low-frequency image; the upsampling operation is performed based on a preset upsampling multiple; A high-frequency image corresponding to the ith image is obtained according to upsampling results corresponding to the ith image and the initial low-frequency image.

5. The method according to claim 1, characterized in that The fusing the low-frequency defect-removed image and the high-frequency defect-removed image to obtain the defect-removed image corresponding to the i-th image includes: Performing an upsampling operation on the low-frequency defect-removed image to obtain an upsampling result corresponding to the low-frequency defect-removed image; the upsampling operation is performed based on a preset upsampling multiple; The upsampling result corresponding to the low-frequency de-defected image and the high-frequency de-defected image are fused to obtain a de-defected image corresponding to the i-th image.

6. An image processing device, characterized in that: The device comprises: an image sequence obtaining unit, configured to perform at least two reduction operations based on the original image to obtain an image sequence including the original image; the images in the image sequence are arranged in descending order according to size; a defect removal processing unit configured to perform defect removal processing on the images in the image sequence in reverse order until a defect-removed image corresponding to the original image is obtained; The following defect removal process is performed for the i-th image in the image sequence: Performing defect removal processing on the low-frequency image corresponding to the i-th image to obtain a low-frequency defect-removed image; Performing defect removal processing on the high-frequency image corresponding to the i-th image to obtain a high-frequency defect-removed image; Fusion of the low-frequency defect-removed image and the high-frequency defect-removed image to obtain a defect-removed image corresponding to the i-th image; The defect-removed image corresponding to the i-th image is used as the low-frequency image corresponding to the i-1-th image in the image sequence, where i is a positive integer greater than 1 and less than N; The defect removal processing unit is specifically configured to execute the steps of taking the i+1th image as an initial low-frequency image of the i-th image; obtaining a high-frequency image corresponding to the i-th image according to the i-th image and the initial low-frequency image; and inputting the high-frequency image into the defect removal network to obtain a high-frequency defect removal image corresponding to the i-th image.

7. The device according to claim 6, characterized in that The image sequence obtaining unit is specifically configured to obtain an original image, perform N downsampling operations on the original image, and obtain N reduced images of different sizes, N≥2; the downsampling operation is performed based on preset size reduction multiple information; the original image is taken as the maximum size, the reduced image obtained by the Nth downsampling operation is taken as the minimum size, and the N reduced images of different sizes and the original image are arranged in descending order according to size to obtain the image sequence.

8. The device according to claim 6, characterized in that The defect removal processing unit is specifically configured to input the low-frequency image corresponding to the i-th image into the defect removal network to obtain the low-frequency defect removal image corresponding to the i-th image.

9. The device according to claim 6, characterized in that The blemish removal processing unit is specifically configured to perform an upsampling operation on the initial low-frequency image to obtain an upsampling result corresponding to the initial low-frequency image; the upsampling operation is performed based on a preset upsampling multiple; and a high-frequency image corresponding to the i-th image is obtained according to the upsampling results corresponding to the i-th image and the initial low-frequency image.

10. The device according to claim 6, characterized in that The defect removal processing unit is further configured to perform an upsampling operation on the low-frequency defect removal image to obtain an upsampling result corresponding to the low-frequency defect removal image; the upsampling operation is performed based on a preset upsampling multiple; The upsampling result corresponding to the low-frequency de-defected image and the high-frequency de-defected image are fused to obtain a de-defected image corresponding to the i-th image.

11. An electronic device, characterized in that: include: processor; a memory for storing instructions executable by the processor; The processor is configured to execute the instructions to implement the image processing method according to any one of claims 1 to 5.

12. A computer-readable storage medium, characterized in that: When the instructions in the computer-readable storage medium are executed by a processor of an electronic device, the electronic device is enabled to perform the image processing method as claimed in any one of claims 1 to 5.

13. A computer program product, comprising instructions, characterized in that: When the instruction is executed by a processor of an electronic device, the electronic device is enabled to execute the image processing method as described in any one of claims 1 to 5.

Images