Image super-resolution processing method, device, equipment and storage medium

Abstract

This application discloses an image super-resolution processing method, apparatus, device, and storage medium. The technical solution provided by this application involves inputting a first-resolution image with an alpha channel into an image reconstruction model. The image reconstruction model then performs super-resolution reconstruction on the image to be processed, obtaining a second-resolution target image with an alpha channel, and outputting the target image. This automatically generates a super-resolution image of the image to be processed with an alpha channel, effectively improving the generation efficiency of super-resolution images.

Description

Technical Field

[0001] This application relates to the field of image processing technology, and in particular to an image super-resolution processing method, apparatus, device, and storage medium. Background Technology

[0002] In the field of image processing technology, super-resolution reconstruction is an important image processing technique for improving the resolution of images and videos. Image super-resolution technology refers to the technique of reconstructing a corresponding second-resolution image from a first-resolution image. Using image super-resolution technology, a second-resolution image can be obtained from a first-resolution image, or blurry parts in a blurry image can be removed to generate a super-resolution image with better visual effects, thereby improving the quality and utilization value of the image.

[0003] Currently, super-resolution reconstruction of images is performed on RGB images. For ARGB images that include alpha channels, obtaining a super-resolution image generally requires re-mapping, resulting in low efficiency in generating super-resolution images. Summary of the Invention

[0004] This application provides an image super-resolution processing method, apparatus, device, and storage medium to solve the technical problem in the prior art that super-resolution images need to be obtained by re-mapping, resulting in low generation efficiency of super-resolution images. It automatically generates super-resolution images, effectively improving the generation efficiency of super-resolution images.

[0005] In a first aspect, embodiments of this application provide an image super-resolution processing method, comprising:

[0006] Acquire an image to be processed, wherein the image to be processed is a first-resolution image with an alpha channel;

[0007] The image to be processed is added to a trained image reconstruction model, which performs super-resolution reconstruction on the image to be processed and generates a target image. The target image is a second-resolution image with an alpha channel, wherein the resolution of the second-resolution image is higher than that of the first-resolution image.

[0008] Output the target image generated by the image reconstruction model.

[0009] In a second aspect, embodiments of this application provide an image super-resolution processing apparatus, including an image acquisition module, an image reconstruction module, and an image output module, wherein:

[0010] The image acquisition module is used to acquire an image to be processed, wherein the image to be processed is a first-resolution image with an alpha channel;

[0011] The image reconstruction module is used to add the image to be processed into a trained image reconstruction model, and the image reconstruction model performs super-resolution reconstruction processing on the image to be processed to generate a target image. The target image is a second resolution image with an alpha channel, wherein the resolution of the second resolution image is higher than that of the first resolution image.

[0012] The image output module is used to output the target image generated by the image reconstruction model.

[0013] In a third aspect, embodiments of this application provide an image super-resolution processing apparatus, including: a memory and one or more processors;

[0014] The memory is used to store one or more programs;

[0015] When the one or more programs are executed by the one or more processors, the one or more processors implement the image super-resolution processing method as described in the first aspect.

[0016] In a fourth aspect, embodiments of this application provide a storage medium for storing computer-executable instructions, which, when executed by a computer processor, are used to perform the image super-resolution processing method as described in the first aspect.

[0017] This application embodiment inputs the image to be processed with a transparency channel and a first resolution into an image reconstruction model. The image reconstruction model performs super-resolution reconstruction processing on the image to be processed to obtain a target image with a transparency channel and a second resolution, and outputs the target image. This automatically generates a super-resolution image of the image to be processed with a transparency channel, effectively improving the generation efficiency of the super-resolution image. Attached Figure Description

[0018] Figure 1 This is a flowchart of an image super-resolution processing method provided in an embodiment of this application;

[0019] Figure 2 This is a schematic diagram of a second resolution sample image provided in an embodiment of this application;

[0020] Figure 3 This is a schematic diagram of a first-resolution sample image provided in an embodiment of this application;

[0021] Figure 4 This is a schematic diagram of super-resolution reconstruction processing of an image to be processed according to an embodiment of this application;

[0022] Figure 5 This is a schematic diagram of the structure of an image super-resolution processing device provided in an embodiment of this application;

[0023] Figure 6 This is a schematic diagram of the structure of an image super-resolution processing device provided in an embodiment of this application. Detailed Implementation

[0024] To make the objectives, technical solutions, and advantages of this application clearer, specific embodiments of this application will be described in further detail below with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are merely for explaining this application and not for limiting it. It should also be noted that, for ease of description, only the parts relevant to this application are shown in the drawings, not all of them. Before discussing exemplary embodiments in more detail, it should be mentioned that some exemplary embodiments are described as processes or methods depicted as flowcharts. Although the flowcharts describe operations (or steps) as sequential processes, many of these operations can be performed in parallel, concurrently, or simultaneously. Furthermore, the order of the operations can be rearranged. The process can be terminated when its operation is completed, but additional steps not included in the drawings may also be present. The above processes can correspond to methods, functions, procedures, subroutines, subroutines, etc.

[0025] Figure 1 A flowchart of an image super-resolution processing method provided in this application embodiment is given. The image super-resolution processing method provided in this application embodiment can be executed by an image super-resolution processing device, which can be implemented by hardware and / or software and integrated into an image super-resolution processing device (such as a mobile phone, tablet, computer, etc.).

[0026] The following description uses an image super-resolution processing device to perform an image super-resolution processing method as an example.

[0027] refer to Figure 1 The image super-resolution processing method includes:

[0028] S101: Obtain the image to be processed, which is a first-resolution image with an alpha channel.

[0029] The image to be processed provided in this embodiment is an image with a transparency channel (alpha channel) and a first resolution, where the transparency channel can be used to indicate the transparency and semi-transparency of the corresponding image. It should be explained that the image to be processed provided in this solution is an image that requires super-resolution reconstruction to obtain a super-resolution image with a higher resolution and the same transparency channel.

[0030] In related technologies, after performing super-resolution reconstruction on the image to be processed, the resulting image is a super-resolution image without an alpha channel. If a high-resolution image with an alpha channel is required, it is usually necessary to add an alpha channel to the super-resolution image after super-resolution reconstruction (e.g., manually adding an alpha channel to the super-resolution image obtained through image processing software such as Photoshop), or to redraw a new image with an alpha channel and a higher resolution based on the image to be processed. These methods of obtaining a target image with an alpha channel and a second-highest resolution based on a first-resolution image to be processed require manual image creation or reconstruction, resulting in a large workload and low efficiency in generating super-resolution images. To solve these technical problems, this solution uses an image reconstruction model to perform super-resolution reconstruction on the image to be processed, automatically generating a target image with an alpha channel and a higher resolution. This eliminates the need for manual processing of the reconstructed image, effectively improving the efficiency of super-resolution image generation.

[0031] For example, when it is necessary to perform super-resolution reconstruction processing on an image to be processed with a first resolution and an alpha channel, the image to be processed is provided to the image super-resolution processing device. The image to be processed can be obtained externally (e.g., downloaded via a network or provided by other terminal devices via wired or wireless means) or selected from local storage locations. After obtaining the image to be processed, the image super-resolution processing device can submit it to the image reconstruction model for further processing.

[0032] S102: The image to be processed is added to the trained image reconstruction model, and the image reconstruction model performs super-resolution reconstruction processing on the image to be processed to generate a target image. The target image is a second resolution image with an alpha channel, wherein the resolution of the second resolution image is higher than that of the first resolution image.

[0033] This solution utilizes an image reconstruction model to perform super-resolution reconstruction of the image to be processed. The image reconstruction model can be built based on a neural network and trained using collected sample data (including sample images of various resolutions, all with alpha channels) until the loss function of the image reconstruction model is within a set threshold or the accuracy of the image reconstruction model reaches a set requirement. After training, the image reconstruction model is configured in an image super-resolution processing device.

[0034] For example, after acquiring the image to be processed, the image to be processed is input into the image reconstruction model, which performs super-resolution reconstruction on the image to be processed to generate a target image with a second resolution and an alpha channel.

[0035] In this scheme, the resolution of the second-resolution image is higher than that of the first-resolution image. It should be clarified that the first and second resolutions provided in this scheme (including the resolutions corresponding to the first and second-resolution sample images during training) are not specified resolutions; they are relative concepts. For example, when performing super-resolution reconstruction on a 128*128 image to obtain a 256*256 image, 128*128 is the first resolution, and 256*256 is the second resolution. Similarly, when performing super-resolution reconstruction on a 256*256 image to obtain a 512*512 image, 256*256 is the first resolution, and 512*512 is the second resolution.

[0036] In one possible embodiment, the image reconstruction model provided by this scheme is trained based on a first-resolution sample image with an alpha channel and a second-resolution sample image with an alpha channel. Furthermore, the second-resolution sample image with an alpha channel and the first-resolution sample image with an alpha channel are set in pairs, wherein the resolution of the second-resolution sample image is higher than the resolution of the first-resolution sample image.

[0037] For example, after building the image reconstruction model, a pair of first-resolution sample images and second-resolution sample images are obtained. The first-resolution sample image is used as input and the corresponding second-resolution sample image is used as output. The first-resolution sample image and the second-resolution sample image are then input into the image reconstruction model for training.

[0038] The first-resolution sample image and the second-resolution sample image are set in pairs. This can be understood as the first-resolution sample image and the second-resolution sample image displaying the same content, only the resolution of the second-resolution sample image is higher than that of the first-resolution sample image. The first-resolution image can be obtained by downsampling the second-resolution sample image, or by adding the corresponding transparency channel to the first-resolution sample image after performing super-resolution reconstruction without the transparency channel, or by creating a map based on the same content using the paired first-resolution sample image and the second-resolution sample image.

[0039] In one possible embodiment, the second-resolution sample image with an alpha channel provided by this solution is obtained by configuring the alpha channel of pre-collected second-resolution sample images without an alpha channel. That is, according to the training sample requirements of the image reconstruction model, multiple second-resolution sample images (high-resolution sample images) without an alpha channel are collected. After obtaining the second-resolution sample images without an alpha channel, the alpha channel is configured on these second-resolution sample images to obtain the second-resolution sample image with an alpha channel corresponding to each second-resolution sample image.

[0040] Correspondingly, when configuring the transparency channel for a second-resolution sample image without a transparency channel, the process includes: adding a transparency channel to the second-resolution sample image without a transparency channel to obtain a second-resolution sample image with a transparency channel; and configuring the transparency information for the transparency channel of the second-resolution sample image with a transparency channel.

[0041] For example, after acquiring multiple second-resolution sample images without an opacity channel, an opacity channel is added to each second-resolution sample image without an opacity channel, and the opacity information of the corresponding opacity channel of each second-resolution sample image with an added opacity channel is configured.

[0042] Among them, when configuring the transparency information, a second-resolution sample image without a transparency channel can be modified by adding a transparency channel to configure different regions and / or transparency parameters of the transparency channel, resulting in multiple combinations of different regions and / or transparency parameters to obtain a second-resolution sample image with a transparency channel.

[0043] In one possible embodiment, the transparency information includes the transparency region of the corresponding image (representing the range of transparency in the image) and a transparency parameter (representing the transparency or opacity of the image within the transparency region). The transparency information configuration for the second resolution sample image can be based on a combination of different transparency regions and transparency parameters. Therefore, when configuring the transparency information for the transparency channel of the second resolution sample image with a transparency channel, specifically: random region transparency processing is performed on the transparency channel of the second resolution sample image with a transparency channel.

[0044] For example, after adding an alpha channel to a second-resolution sample image without an alpha channel, alpha regions are defined within the second-resolution sample image based on multiple randomly determined areas and / or shapes, and a alpha parameter corresponding to each alpha region is randomly determined. It can be understood that after configuring the alpha channel of a second-resolution sample image with an alpha channel, multiple second-resolution sample images with alpha channels and combinations of different alpha regions and alpha parameters can be obtained. This expands the number of samples for training the image reconstruction model, improves the efficiency of obtaining training samples, and greatly enhances the diversity of training samples, thereby improving the training efficiency and accuracy of the image reconstruction model.

[0045] In one possible embodiment, after obtaining multiple second-resolution sample images with alpha channels, first-resolution sample images with alpha channels can be generated based on these second-resolution sample images with alpha channels, thereby obtaining paired first-resolution sample images and second-resolution sample images. Based on this, the first-resolution sample image with alpha channels provided in this solution is obtained by downsampling the second-resolution sample images with alpha channels.

[0046] For example, after adding and configuring the transparency channel to a second-resolution sample image without a transparency channel to obtain a second-resolution sample image with a transparency channel, each second-resolution sample image with a transparency channel is downsampled to obtain the corresponding first-resolution sample image with a transparency channel.

[0047] In one possible embodiment, the image reconstruction model provided by this solution is built based on a super-resolution network (e.g., Enhanced Super-Resolution Generative Adversarial Networks, ESRGAN). The input parameters of the super-resolution network include image width, image height, image color channels, and image transparency channels. For example, changing the input and output of the super-resolution network from (w, h, 3) to (w, h, 4) allows the super-resolution network to support the image transparency channel, where w represents the image width, h represents the image height, 3 represents the three color channels of the image (e.g., the three RGB color channels, i.e., red, green, and blue), and 4 represents one image transparency channel and three color channels (e.g., the four ARGB channels, i.e., the image transparency channel, red, green, and blue).

[0048] like Figure 2A schematic diagram of a second-resolution sample image is provided, where B1 is a second-resolution sample image without an alpha channel, and B21 and B22 are second-resolution sample images with alpha channels. Assuming the resolution of this second-resolution sample image is 512*512, an alpha channel is added to the second-resolution sample image without an alpha channel, and random area alpha processing is applied to the alpha channel of the second-resolution sample image with an alpha channel, resulting in the following... Figure 2 The diagram shows second-resolution sample images B21 and B22 with alpha channels. Assuming two alpha regions are randomly assigned to each second-resolution sample image, second-resolution sample images with corresponding alpha regions are obtained. In the diagram, C21() represents the alpha region corresponding to second-resolution sample image B21, and C22 represents the alpha region corresponding to second-resolution sample image B22. Furthermore, multiple different alpha parameters are randomly configured for the alpha channels of each second-resolution sample image with an alpha channel, resulting in multiple second-resolution sample images with alpha channels with different combinations of alpha regions and alpha parameters. For example, if 10 different alpha parameters are configured for the alpha channels of two second-resolution sample images, then the second-resolution sample image without an alpha channel corresponds to multiple second-resolution sample images with alpha channels with alpha channels, resulting in 20 different combinations of alpha regions and alpha parameters.

[0049] Furthermore, such as Figure 3 As shown in the schematic diagram of a first-resolution sample image, the second-resolution sample image with an alpha channel generated above is downsampled to obtain one or more (which can be downsampled multiple times based on one or more downsampling control parameters) first-resolution sample images with an alpha channel whose resolution is lower than the second resolution. Figure 3 The first resolution sample images A11 and A12 with the alpha channel are obtained by... Figure 2 The second-resolution sample image B21 with an alpha channel is obtained by downsampling. The first-resolution sample image A11 has a resolution of 256*256, and the first-resolution sample image A12 has a resolution of 128*128. One second-resolution sample image with an alpha channel can correspond to one or more other second-resolution sample images with alpha channels; that is, one second-resolution sample image with an alpha channel can be paired with one or more other second-resolution sample images with alpha channels.

[0050] Figure 4 This is a schematic diagram of super-resolution reconstruction processing of an image to be processed according to an embodiment of this application, wherein the resolution of the image to be processed, D3, is 128*128, and C31 is the transparency area corresponding to the image to be processed, D3. Figure 4As shown, after obtaining the image D3 to be processed for super-resolution reconstruction, the image D3 is input into the image reconstruction model. The image reconstruction model performs super-resolution reconstruction on the image D3 and outputs the target image D4 corresponding to the image D3. The resolution of the target image D4 is 512*512, and C41 is the transparency area corresponding to the target image D4.

[0051] S103: Output the target image generated by the image reconstruction model.

[0052] For example, the target image generated by the image reconstruction model is acquired and output. The output method for the target image can be: saving the target image to a designated image storage location, directly displaying the generated target image, sending it to a terminal device requesting super-resolution reconstruction of the image to be processed, or opening the target image in designated image processing software.

[0053] As described above, by inputting the image to be processed with an alpha channel and a first resolution into the image reconstruction model, the model performs super-resolution reconstruction on the image to be processed, obtaining a target image with an alpha channel and a second resolution, and outputting the target image, the super-resolution image of the image to be processed with an alpha channel is automatically generated, effectively improving the generation efficiency of super-resolution images. Simultaneously, by adding an image alpha channel to the super-resolution network to build the image reconstruction model, and training the model based on paired second-resolution sample images and first-resolution sample images with an alpha channel, the target image with an alpha channel can be automatically generated, improving the generation efficiency of super-resolution images. Furthermore, by randomly configuring the alpha channels on collected second-resolution sample images without alpha channels to obtain second-resolution sample images with alpha channels, and by downsampling the second-resolution sample images with alpha channels to obtain first-resolution sample images paired with the second-resolution sample images with alpha channels, the diversity of training samples is effectively improved, enhancing the training efficiency and accuracy of the image reconstruction model.

[0054] Figure 5 A schematic diagram of an image super-resolution processing apparatus provided in an embodiment of this application is given. (Reference) Figure 5 The image super-resolution processing device includes an image acquisition module 51, an image reconstruction module 52, and an image output module 53, wherein:

[0055] The image acquisition module 51 is used to acquire an image to be processed, which is a first-resolution image with an alpha channel; the image reconstruction module 52 is used to add the image to be processed into a trained image reconstruction model, which performs super-resolution reconstruction on the image to be processed and generates a target image, which is a second-resolution image with an alpha channel, wherein the resolution of the second-resolution image is higher than that of the first-resolution image; the image output module 53 is used to output the target image generated by the image reconstruction model.

[0056] As described above, by inputting the image to be processed with a transparency channel and a first resolution into the image reconstruction model, the image reconstruction model performs super-resolution reconstruction processing on the image to be processed to obtain a target image with a transparency channel and a second resolution, and outputs the target image, automatically generating a super-resolution image of the image to be processed with a transparency channel, effectively improving the generation efficiency of super-resolution images.

[0057] In one possible embodiment, the image super-resolution processing device further includes a model training module, through which the image reconstruction model is trained based on a first resolution sample image with an alpha channel and a second resolution sample image with an alpha channel. The second resolution sample image with an alpha channel and the first resolution sample image with an alpha channel are set in pairs, wherein the resolution of the second resolution sample image is higher than the resolution of the first resolution sample image.

[0058] In one possible embodiment, the image super-resolution processing device further includes a sample acquisition module, wherein the second resolution sample image with transparency channel is obtained by configuring the transparency channel of a pre-collected second resolution sample image without transparency channel through the sample acquisition module.

[0059] In one possible embodiment, the sample acquisition module, when configuring the transparency channel for the second-resolution sample image without a transparency channel, includes:

[0060] Adding a transparency channel to a second-resolution sample image without a transparency channel yields a second-resolution sample image with a transparency channel.

[0061] The transparency information of the transparency channel of the second-resolution sample image with the transparency channel is configured.

[0062] In one possible embodiment, when configuring the transparency information of the transparency channel of the second-resolution sample image with a transparency channel, the sample acquisition module includes:

[0063] Random area transparency processing is performed on the transparency channel of the second-resolution sample image with the transparency channel.

[0064] In one possible embodiment, the image super-resolution processing device further includes a sample acquisition module, wherein the first resolution sample image with a transparency channel is obtained by downsampling the second resolution sample image with a transparency channel through the sample acquisition module.

[0065] In one possible embodiment, the image reconstruction model is built based on a super-resolution network, wherein the input parameters of the super-resolution network include image width, image height, image color channels, and image transparency channels.

[0066] It is worth noting that in the embodiments of the above-mentioned image super-resolution processing device, the various units and modules included are only divided according to functional logic, but are not limited to the above division, as long as the corresponding functions can be achieved; in addition, the specific names of each functional unit are only for easy distinction between each other and are not used to limit the protection scope of the embodiments of the present invention.

[0067] This application also provides an image super-resolution processing device, which can integrate the image super-resolution processing apparatus provided in this application. Figure 6 This is a schematic diagram of the structure of an image super-resolution processing device provided in an embodiment of this application. (Reference) Figure 6 The image super-resolution processing device includes: an input device 63, an output device 64, a memory 62, and one or more processors 61; the memory 62 is used to store one or more programs; when one or more programs are executed by one or more processors 61, the one or more processors 61 implement the image super-resolution processing method provided in the above embodiments. The input device 63, output device 64, memory 62, and processors 61 can be connected via a bus or other means. Figure 6 The bus connection is taken as an example.

[0068] The memory 62, as a computing device readable storage medium, can be used to store software programs, computer-executable programs, and modules, such as program instructions / modules corresponding to the image super-resolution processing method provided in any embodiment of this application (e.g., image acquisition module 51, image reconstruction module 52, and image output module 53 in the image super-resolution processing apparatus). The memory 62 may primarily include a program storage area and a data storage area. The program storage area may store the operating system and at least one application program required for a function; the data storage area may store data created based on the use of the device, etc. Furthermore, the memory 62 may include high-speed random access memory and may also include non-volatile memory, such as at least one disk storage device, flash memory device, or other non-volatile solid-state storage device. In some instances, the memory 62 may further include memory remotely located relative to the processor 61, and these remote memories can be connected to the device via a network. Examples of such networks include, but are not limited to, the Internet, corporate intranets, local area networks, mobile communication networks, and combinations thereof.

[0069] Input device 63 can be used to receive input digital or character information, and to generate key signal inputs related to user settings and function control of the device. Output device 64 may include display devices such as a display screen.

[0070] The processor 61 executes various functional applications and data processing of the device by running software programs, instructions and modules stored in the memory 62, thereby realizing the above-mentioned image super-resolution processing method.

[0071] The image super-resolution processing apparatus, device, and computer provided above can be used to execute the image super-resolution processing method provided in any of the above embodiments, and have corresponding functions and beneficial effects.

[0072] This application embodiment also provides a storage medium for storing computer-executable instructions. When executed by a computer processor, the computer-executable instructions are used to perform the image super-resolution processing method provided in the above embodiment. The image super-resolution processing method includes: acquiring an image to be processed, wherein the image to be processed is a first resolution image with an alpha channel; adding the image to be processed into a trained image reconstruction model, wherein the image reconstruction model performs super-resolution reconstruction processing on the image to be processed and generates a target image, wherein the target image is a second resolution image with an alpha channel, wherein the resolution of the second resolution image is higher than the resolution of the first resolution image; and outputting the target image generated by the image reconstruction model.

[0073] Storage medium – any type of memory device or storage device. The term “storage medium” is intended to include: mounting media, such as CD-ROMs, floppy disks, or magnetic tape devices; computer system memory or random access memory, such as DRAM, DDR RAM, SRAM, EDO RAM, Rambus RAM, etc.; non-volatile memory, such as flash memory, magnetic media (e.g., hard disks or optical storage); registers or other similar types of memory elements, etc. Storage media may also include other types of memory or combinations thereof. Furthermore, storage media may reside in a first computer system in which a program is executed, or may reside in a different second computer system connected to the first computer system via a network (such as the Internet). The second computer system can provide program instructions to the first computer for execution. The term “storage medium” can include two or more storage media that may reside in different locations (e.g., in different computer systems connected via a network). Storage media may store program instructions (e.g., specifically implemented as a computer program) executable by one or more processors.

[0074] Of course, the computer-executable instructions stored in the storage medium provided in the embodiments of this application are not limited to the image super-resolution processing method provided above, but can also perform related operations in the image super-resolution processing method provided in any embodiment of this application.

[0075] The image super-resolution processing apparatus, device, and storage medium provided in the above embodiments can execute the image super-resolution processing method provided in any embodiment of this application. For technical details not described in detail in the above embodiments, please refer to the image super-resolution processing method provided in any embodiment of this application.

[0076] The above description is merely a preferred embodiment and the technical principles employed in this application. This application is not limited to the specific embodiments provided herein, and various obvious changes, readjustments, and substitutions that can be made by those skilled in the art will not depart from the scope of protection of this application. Therefore, although this application has been described in detail through the above embodiments, this application is not limited to the above embodiments, and may include many other equivalent embodiments without departing from the concept of this application, the scope of which is determined by the scope of the claims.

Claims

1. An image super-resolution processing method, characterized in that, include: Acquire an image to be processed, wherein the image to be processed is a first-resolution image with an alpha channel; The image to be processed is added to a trained image reconstruction model, which performs super-resolution reconstruction on the image to be processed and generates a target image. The target image is a second-resolution image with an alpha channel, wherein the resolution of the second-resolution image is higher than that of the first-resolution image. The image reconstruction model is trained based on a first-resolution sample image with an alpha channel and a second-resolution sample image with an alpha channel. The second-resolution sample image with an alpha channel and the first-resolution sample image with an alpha channel are set in pairs. Output the target image generated by the image reconstruction model.

2. The image super-resolution processing method according to claim 1, characterized in that, The second-resolution sample image with an opacity channel is obtained by configuring the opacity channel on a pre-collected second-resolution sample image without an opacity channel.

3. The image super-resolution processing method according to claim 2, characterized in that, When configuring the transparency channel for the second-resolution sample image without a transparency channel, the following steps are included: Adding an alpha channel to a second-resolution sample image without an alpha channel yields a second-resolution sample image with an alpha channel. The transparency information of the transparency channel of the second-resolution sample image with the transparency channel is configured.

4. The image super-resolution processing method according to claim 3, characterized in that, When configuring the transparency information for the transparency channel of the second-resolution sample image with the transparency channel, the following steps are included: Random area transparency processing is performed on the transparency channel of the second-resolution sample image with the transparency channel.

5. The image super-resolution processing method according to claim 1, characterized in that, The first resolution sample image with a transparency channel is obtained by downsampling the second resolution sample image with a transparency channel.

6. The image super-resolution processing method according to claim 1, characterized in that, The image reconstruction model is built based on a super-resolution network. The input parameters of the super-resolution network include image width, image height, image color channels, and image transparency channels.

7. An image super-resolution processing apparatus, characterized in that, It includes an image acquisition module, an image reconstruction module, and an image output module, wherein: The image acquisition module is used to acquire an image to be processed, wherein the image to be processed is a first-resolution image with an alpha channel; The image reconstruction module is used to add the image to be processed into a trained image reconstruction model, and the image reconstruction model performs super-resolution reconstruction processing on the image to be processed to generate a target image. The target image is a second resolution image with an alpha channel, wherein the resolution of the second resolution image is higher than that of the first resolution image. The image reconstruction model is trained based on a first resolution sample image with an alpha channel and a second resolution sample image with an alpha channel. The second resolution sample image with an alpha channel and the first resolution sample image with an alpha channel are set in pairs. The image output module is used to output the target image generated by the image reconstruction model.

8. An image super-resolution processing device, characterized in that, include: Memory and one or more processors; The memory is used to store one or more programs; When the one or more programs are executed by the one or more processors, the one or more processors implement the image super-resolution processing method as described in any one of claims 1-6.

9. A storage medium for storing computer-executable instructions, characterized in that, The computer-executable instructions, when executed by a computer processor, are used to perform the image super-resolution processing method as described in any one of claims 1-6.

Images