Method and System for Improving the Smoothness of Cloud Games

Abstract

The present invention provides a method for improving the smoothness of cloud games, belonging to the technical field of image processing. The method includes receiving a game screen video stream rendered by a server end at a low resolution; decoding the video stream to obtain a video screen; and performing upsampling on the decoded video screen to generate a game screen for display on a client end. By means of a super-resolution method, the present invention reduces the resolution of the game screen on the server end, reduces the rendering amount of the GPU on the server end, and reduces the lag phenomenon caused by overloading of the GPU. By means of the super-resolution method, the present invention reduces the resolution of the game screen on the server end, reduces the amount of data to be transmitted to the client end, and reduces the cloud game lag and delay phenomena caused by network transmission. The resolution of the game screen on the server end can be reduced, the GPU performance requirement on the server end is lowered, and a low-cost and low-performance GPU can be used for rendering games, thereby reducing the equipment cost of cloud game servers.

Description

Technical Field

[0001] The present invention relates to the technical field of image processing, and particularly relates to a method and system for improving the smoothness of cloud games. Background Art

[0002] Cloud game is a game mode based on cloud computing. In the operation mode of cloud games, all games are run on the server side, and the rendered game screens are compressed and transmitted to users through the network. On the client side, the user's game device does not require any high-end processors and graphics cards, and only basic video decompression and display capabilities are needed.

[0003] Image super-resolution reconstruction technology is a technology that uses LR (low-resolution) images to restore the corresponding HR (high-resolution) images to improve the image resolution. Super-resolution reconstruction technology can be divided into two types. One is the non-deep learning method, that is, the traditional super-resolution method is used to achieve image reconstruction. The other is the deep learning method, that is, a convolutional neural network is used to achieve image feature extraction, mapping, and reconstruction work.

[0004] Video is composed of continuous images. Video super-resolution reconstruction technology refers to the technology of generating a corresponding high-resolution video version from the corresponding low-resolution video version. Frame rate refers to the number of frames of an image refreshed per second in a video or game screen, and the unit of measurement is FPS (frames per second). The higher the frame rate, the smoother the video and game screens will be. To ensure a good gaming experience for users, usually the frame rate of the game screen is not less than 30 FPS, and the optimal playing frame rate is 60 FPS or 120 FPS.

[0005] Currently, the implementation method of cloud games is as follows: the game screen is rendered on the server side and encoded into a video stream, and the video stream is transmitted to the client side, where the client side decodes and displays the video stream.

[0006] In order to allow users to obtain a high-quality gaming experience, cloud games use a relatively high frame rate and resolution. Usually, the frame rate of cloud game screens is not less than 30 FPS, and the screen resolution is not less than 1920×1080. High frame rates and high resolutions will generate a large amount of video stream data that needs to be transmitted, which poses very high requirements on the network transmission rate between the server side and the client side. If the network rate between the server side and the client side fails to meet the requirements, cloud games will experience stuttering and latency, thus seriously affecting the gaming experience.

[0007] In addition, if the network condition between the server and the client is unstable and the network transmission rate fluctuates, the cloud game will also experience lag. During the process of game screen rendering and encoding, in order to improve the encoding efficiency, the GPU (Graphics Processing Unit) is usually directly used for encoding. However, the game screen rendering also uses the GPU. Therefore, if the game rendering volume is too large, it will cause the GPU to be overloaded, resulting in a longer GPU encoding time and causing the cloud game to lag. Summary of the Invention

[0008] The object of the present invention is to provide a method and system for improving the smoothness of cloud games, so as to solve at least one of the technical problems existing in the above background technology.

[0009] In order to achieve the above object, the present invention adopts the following technical solutions:

[0010] On the one hand, the present invention provides a method for improving the smoothness of cloud games, including:

[0011] Receiving the game screen video stream rendered by the server in low resolution; decoding the video stream to obtain the video screen; performing upsampling on the decoded video screen to generate an upsampled super-resolution game screen, and displaying it on the client.

[0012] Preferably, the decoded video screen is upsampled using an interpolation upsampling algorithm to generate an upsampled game screen.

[0013] Preferably, after sharpening the upsampled game screen, it is used for client display.

[0014] Preferably, the upsampling ratio of the interpolation upsampling algorithm is the ratio of the resolution size of the screen to be displayed by the client to the size of the game screen rendered by the server.

[0015] Preferably, the interpolation upsampling algorithm is Lanczos interpolation, nearest neighbor interpolation, bilinear quadratic interpolation or bicubic interpolation.

[0016] Preferably, the decoded video screen is upsampled using a super-resolution model to generate a super-resolution game screen, and the super-resolution game screen is used for client display.

[0017] Preferably, the training of the super-resolution model includes:

[0018] Using the same controller to control two groups of cloud game servers, and rendering game screens with different resolutions respectively; among them, the resolution size of the game screen rendered with a higher resolution is the resolution size of the game screen displayed by the client, and the resolution size of the game screen rendered with a lower resolution is the size of the game screen rendered by the server;

[0019] Encode the game screen rendered at the higher resolution and the game screen rendered at the lower resolution into a high-resolution video stream and a low-resolution video stream respectively;

[0020] Decode the high-resolution video stream and the low-resolution video stream into a high-resolution video and a low-resolution video respectively;

[0021] Convert the high-resolution video and the low-resolution video into a high-resolution image sequence and a low-resolution image sequence respectively;

[0022] The high-resolution image sequence and the low-resolution image sequence form a super-resolution data set;

[0023] Crop the super-resolution data set into overlapping image patches of a unified size;

[0024] Determine the network structure and loss function to be used, use the cropped image patches for training, and finally generate a super-resolution model.

[0025] In a second aspect, the present invention provides a system for improving the smoothness of cloud games, including:

[0026] A receiving module for receiving the game screen video stream rendered at the low resolution by the server side; a decoding module for decoding the video stream to obtain a video screen; a sampling module for upsampling the decoded video screen to generate an upsampled super-resolution game screen; a display module for displaying the generated super-resolution game screen on the client side.

[0027] In a third aspect, the present invention provides a non-transitory computer-readable storage medium, which is used to store computer instructions. When the computer instructions are executed by a processor, the method for improving the smoothness of cloud games as described above is implemented.

[0028] In a fourth aspect, the present invention provides an electronic device, including: a processor, a memory, and a computer program; wherein, the processor is connected to the memory, the computer program is stored in the memory, and when the electronic device runs, the processor executes the computer program stored in the memory so that the electronic device executes the instructions for implementing the method for improving the smoothness of cloud games as described above.

[0029] Advantages of the present invention: By using the super-resolution method to reduce the game screen resolution on the server side, the rendering volume of the server-side GPU is reduced, and the lag phenomenon caused by overloaded GPU is reduced; by using the super-resolution method to reduce the game screen resolution on the server side, the amount of data to be transmitted to the client is reduced, and the lag and delay phenomena of cloud games caused by network transmission are reduced; it is possible to reduce the game screen resolution on the server side, lower the GPU performance requirements on the server side, and no longer require a high-performance GPU to render some games. Some cheaper low-performance GPUs can be used instead, thereby reducing the equipment cost of cloud game servers.

[0030] Additional aspects and advantages of the present invention will be given in part in the following description, which will become apparent from the following description, or can be understood through the practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0032] Figure 1 It is a schematic flowchart of the method for improving the smoothness of cloud games described in Embodiment 1 of the present invention;

[0033] Figure 2 It is a schematic flowchart of the method for improving the smoothness of cloud games described in Embodiment 3 of the present invention;

[0034] Figure 3 It is a screenshot of a video recording of running a game at 1080P resolution described in Embodiment 3 of the present invention;

[0035] Figure 4 It is a screenshot of a video recording of running a game at 720P resolution described in Embodiment 3 of the present invention;

[0036] Figure 5 It is a schematic diagram of the upsampled game screen described in Embodiment 3 of the present invention;

[0037] Figure 6 It is a schematic diagram of the game screen after sharpening the upsampled game screen described in Embodiment 3 of the present invention;

[0038] Figure 7 It is a schematic diagram of the screen details of the game screen finally displayed on the client described in Embodiment 3 of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0039] The embodiments of the present invention will be described in detail below. Examples of the embodiments are shown in the accompanying drawings, where the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the drawings are exemplary only for explaining the present invention and should not be construed as limiting the present invention.

[0040] Those skilled in the art of the present technology can understand that, unless otherwise defined, all terms (including technical terms and scientific terms) used herein have the same meaning as the general understanding of those of ordinary skill in the art to which the present invention pertains.

[0041] It should also be understood that terms such as those defined in a general dictionary should be understood to have a meaning consistent with their meaning in the context of the prior art and will not be interpreted in an idealized or overly formal sense unless defined as herein.

[0042] Those skilled in the art of the present technology can understand that, unless specifically stated otherwise, the singular forms "a", "an", "the", and "said" used herein may also include the plural forms. It should be further understood that the term "comprising" used in the description of the present invention means the presence of the stated features, integers, steps, operations, elements, and / or components, but does not exclude the presence or addition of one or more other features, integers, steps, operations, elements, and / or their groups.

[0043] In the description of this specification, the description with reference to terms such as "one embodiment", "some embodiments", "example", "specific example", or "some examples" means that the specific features, structures, materials, or characteristics described in connection with the embodiment or example are included in at least one embodiment or example of the present invention. Moreover, the specific features, structures, materials, or characteristics described may be combined in a suitable manner in any one or more embodiments or examples. In addition, without contradiction, those skilled in the art may combine and combine the different embodiments or examples described in this specification and the features of the different embodiments or examples.

[0044] To facilitate the understanding of the present invention, the present invention will be further explained below with specific examples in conjunction with the accompanying drawings, and the specific examples do not constitute a limitation on the embodiments of the present invention.

[0045] Those skilled in the art should understand that the drawings are only schematic diagrams of the embodiments, and the components in the drawings are not necessarily essential for implementing the present invention.

[0046] Embodiment 1

[0047] Embodiment 1 provides a system for improving the smoothness of cloud games. The system includes:

[0048] A receiving module for receiving a game screen video stream rendered by the server side at a low resolution; a decoding module for decoding the video stream to obtain a video screen; a sampling module for upsampling the decoded video screen to generate an upsampled super-resolution game screen; and a display module for displaying the generated game screen on the client side.

[0049] As Figure 1 shown, in Embodiment 1, a method for improving the smoothness of cloud games is implemented by using the above system for improving the smoothness of cloud games. The method includes:

[0050] Using the receiving module of the client to receive the game screen video stream rendered by the server side at a low resolution; using the decoding module of the client to decode the video stream to obtain a video screen; using the sampling module of the client to upsample the decoded video screen to generate an upsampled super-resolution game screen, and using the display module to display the upsampled super-resolution game screen on the client side.

[0051] In Embodiment 1, the sampling module uses an interpolation upsampling algorithm to upsample the decoded video screen to generate an upsampled super-resolution game screen.

[0052] In Embodiment 1, the system further includes a sharpening module for sharpening the generated upsampled super-resolution game screen.

[0053] The upsampling ratio of the interpolation upsampling algorithm is the ratio of the resolution size of the screen to be displayed by the client to the size of the game screen rendered by the server side.

[0054] In specific applications, the interpolation upsampling algorithm is Lanczos interpolation, nearest neighbor interpolation, bilinear quadratic interpolation, bicubic interpolation, etc.

[0055] In Embodiment 1, taking the Lanczos interpolation algorithm as an example, the improvement of the smoothness of cloud games is described. It is determined that the resolution size of the game screen to be displayed by the simulated client is 1920×1080 (1080P), and the size of the game screen rendered by the simulated server side is 1280×720 (720P).

[0056] The game is run and video recorded at 1080P resolution and 720P resolution respectively. The game video recorded while running at 720P resolution is converted into a 720P image sequence, and the upsampling ratio of the interpolation upsampling is calculated. The upsampling ratio = 1080÷720 = 1.5. The 720P image sequence is upsampled using the Lanczos interpolation algorithm with an upsampling ratio of 1.5 to generate an upsampled image sequence. A 3×3 convolutional kernel is used to perform spatial convolution on the upsampled image sequence to generate a sharpened image sequence.

[0057] In Embodiment 1, the 3×3 convolution kernel is as follows:

[0058] [[0, -0.5, 0],

[0059] [-0.5, 3, -0.5],

[0060] [0, -0.5, 0]]

[0061] The sharpened game image sequences are synthesized into a video, which is the game screen simulated for client display.

[0062] Embodiment 2

[0063] Embodiment 2 provides a system for improving the smoothness of cloud games. The system includes:

[0064] A receiving module for receiving the game screen video stream rendered by the server side at a low resolution; a decoding module for decoding the video stream to obtain a video screen; a sampling module for upsampling the decoded video screen to generate an upsampled super-resolution game screen; and a display module for displaying the generated game screen on the client side.

[0065] In Embodiment 2, a method for improving the smoothness of cloud games is implemented using the above system for improving the smoothness of cloud games. The method includes:

[0066] Using the receiving module of the client to receive the game screen video stream rendered by the server side at a low resolution; using the decoding module of the client to decode the video stream to obtain a video screen; using the sampling module of the client to upsample the decoded video screen to generate an upsampled super-resolution game screen, and using the display module to display the upsampled super-resolution game screen on the client side.

[0067] In Embodiment 2, the decoded video screen is upsampled using a super-resolution model to generate a super-resolution game screen, and the super-resolution game screen is used for client display.

[0068] In Embodiment 2, the training of the super-resolution model includes:

[0069] Using the same controller to control two groups of cloud game servers, and rendering game screens at different resolutions respectively; among them, the resolution size of the game screen rendered at a higher resolution is the resolution size of the game screen displayed on the client side, and the resolution size of the game screen rendered at a lower resolution is the resolution size of the game screen rendered by the server side;

[0070] Encoding the game screen rendered at a higher resolution and the game screen rendered at a lower resolution into a high-resolution video stream and a low-resolution video stream respectively;

[0071] Decode the high-resolution video stream and the low-resolution video stream into a high-resolution video and a low-resolution video respectively;

[0072] Convert the high-resolution video and the low-resolution video into a high-resolution image sequence and a low-resolution image sequence respectively;

[0073] The high-resolution image sequence and the low-resolution image sequence constitute a super-resolution data set;

[0074] Crop the super-resolution data set into overlapping image patches of a unified size;

[0075] Determine the network structure and loss function to be used, use the cropped image patches for training, and finally generate a super-resolution model.

[0076] Embodiment 3

[0077] The present invention proposes a method for improving the smoothness of cloud games based on super-resolution. The overall process is as Figure 2 shown, where the super-resolution method is divided into a cloud game super-resolution method based on interpolation and a cloud game super-resolution method based on deep learning.

[0078] In this Embodiment 3, the cloud game super-resolution method based on interpolation is as follows:

[0079] Determine the resolution size of the game screen to be displayed by the client and the size of the game screen rendered by the server; the server renders the game screen and transmits the video stream to the client, and the client decodes the video stream to obtain a video screen; use the interpolation upsampling algorithm to upsample the video screen decoded by the client to generate an upsampled game screen; perform sharpening processing on the upsampled game screen, and use the generated game screen for client display.

[0080] Using the cloud game super-resolution method based on deep learning, including the server using low-resolution to render the game screen and transmitting the video stream to the client, and the client decoding the video stream to obtain a video screen; using the generated super-resolution model to upsample the video screen decoded by the client to generate a super-resolution game screen; using the super-resolution game screen for client display.

[0081] In this Embodiment 3, when using the cloud game super-resolution method based on deep learning, the training of the super-resolution model includes:

[0082] Use the same controller to control two groups of cloud game servers and perform game operations; for the two groups of cloud game servers, one group renders the game screen at a low resolution and the other group renders the game screen at a high resolution; the resolution size of the high-resolution game screen is the resolution size of the game screen displayed by the client in the present invention; the resolution size of the low-resolution game screen is the resolution size of the game screen rendered by the server side in the present invention; encode the high- and low-resolution game screens into high- and low-resolution video streams; decode the high- and low-resolution video streams into high- and low-resolution videos; convert the high- and low-resolution videos into high- and low-resolution image sequences; the high- and low-resolution image sequences constitute a super-resolution data set. Crop the super-resolution data set into overlapping image blocks of the same size; determine the network structure and loss function to be used and perform training, and finally generate a super-resolution model.

[0083] In this embodiment, selectable network structures, such as SRCNN, RLSP, OVSR, etc., and this network structure needs to meet: the speed of inferring the low-resolution video received by the client should be fast enough, at least reaching 30 FPS. Selectable loss functions, such as L1 loss, L2 loss, Perceptual loss, GAN loss.

[0084] In Embodiment 3 of the present invention, the game demonstrated is "Red Dead Redemption 2", and the effect of the cloud game super-resolution method based on interpolation is shown through simulation. The following are the specific steps in Embodiment 3:

[0085] 1. Determine that the resolution size of the game screen to be displayed by the simulated client is 1920×1080 (1080P), and the resolution size of the game screen rendered by the simulated server side is 1280×720 (720P);

[0086] 2. Run the game at 1080P resolution and 720P resolution respectively and perform video recording, Figure 3 and Figure 4 show screenshots of the recorded videos respectively;

[0087] 3. Convert the game video recorded while running at 720P resolution into a 720P image sequence;

[0088] 4. Calculate the upsampling ratio of interpolation upsampling, upsampling ratio = 1080÷720 = 1.5

[0089] 5. Use the Lanczos interpolation algorithm to upsample the 720P image sequence with an upsampling ratio of 1.5 to generate an upsampled image sequence. Figure 5 is an example of the upsampled image;

[0090] 6. Use a 3×3 convolutional kernel to perform spatial convolution on the upsampled image sequence to generate a sharpened image sequence;

[0091] 7. The 3×3 convolutional kernel is:

[0092] [[0, -0.5, 0],

[0093] [-0.5, 3, -0.5],

[0094] [0, -0.5, 0]]

[0095] 8. Synthesize the sharpened game image sequence into a video, which is the game screen simulated for client display. Figure 6 Is the game screen simulated for client display. Figure 7 Is the comparison diagram of game screen details. The left figure is the 720P game details, the middle figure is the game details after interpolation super-resolution, and the right figure is the 1080P game details.

[0096] Example 4

[0097] Example 4 of the present invention provides a non-transitory computer-readable storage medium, which is used to store computer instructions. When the computer instructions are executed by a processor, the method for improving the smoothness of cloud games as described above is implemented. The method includes:

[0098] Receive the game screen video stream rendered by the server side at a low resolution; decode the video stream to obtain a video screen; perform upsampling on the decoded video screen to generate a super-resolution upsampled game screen and display it on the client.

[0099] Example 5

[0100] Example 5 of the present invention provides a computer program (product), including a computer program, which is used to implement the method for improving the smoothness of cloud games as described above when running on one or more processors. The method includes:

[0101] Receive the game screen video stream rendered by the server side at a low resolution; decode the video stream to obtain a video screen; perform upsampling on the decoded video screen to generate a super-resolution upsampled game screen and display it on the client.

[0102] Example 6

[0103] Embodiment 6 of the present invention provides an electronic device, including: a processor, a memory, and a computer program; wherein, the processor is connected to the memory, the computer program is stored in the memory, and when the electronic device runs, the processor executes the computer program stored in the memory, so that the electronic device executes instructions for implementing the method for improving the smoothness of cloud games as described above. The method includes:

[0104] Receiving a game screen video stream rendered by the server side at a low resolution; decoding the video stream to obtain a video screen; performing upsampling on the decoded video screen to generate a super-resolution upsampled game screen, and displaying it on the client side.

[0105] Those skilled in the art should understand that the embodiments of the present invention can be provided as a method, a system, or a computer program product. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present invention can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0106] The present invention is described with reference to the flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to the embodiments of the present invention. It should be understood that each flow and / or block in the flowchart and / or block diagram can be implemented by computer program instructions, and the combination of flows and / or blocks in the flowchart and / or block diagram can also be implemented by computer program instructions. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing devices to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing devices generate means for implementing the functions specified in one Figure 1 one flow or multiple flows and / or blocks Figure 1 one block or multiple blocks.

[0107] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer-readable memory generate a manufactured product including instruction means, and the instruction means implements the functions specified in one Figure 1 one flow or multiple flows and / or blocks Figure 1 one block or multiple blocks.

[0108] These computer program instructions can also be loaded onto a computer or other programmable data processing device, and a series of operation steps are executed on the computer or other programmable device to generate a computer-implemented process. Thus, the instructions executed on the computer or other programmable device provide for implementing the functions in the processFigure 1 One process or multiple processes and / or boxes Figure 1 Steps of functions specified in one box or multiple boxes.

[0109] Although the specific embodiments of the present invention have been described in conjunction with the accompanying drawings, they are not limitations on the protection scope of the present invention. Those skilled in the art should understand that, based on the technical solutions disclosed in the present invention, various modifications or deformations that can be made by those skilled in the art without creative efforts should be covered within the protection scope of the present invention.

Claims

1. A method for improving the smoothness of cloud games, characterized in that, Including: Receiving a video stream of a game screen rendered by the server side at a low resolution; Decoding the video stream to obtain a video screen; Upsampling the decoded video screen to generate a super-resolution upsampled game screen for display on the client side. Specifically: using a super-resolution model to upsample the decoded video screen to generate a super-resolution game screen, and using the super-resolution game screen for client-side display. The training of the super-resolution model includes: Using the same controller to control two groups of cloud game servers to render game screens at different resolutions respectively; among them, the resolution size of the game screen rendered at a higher resolution is the resolution size of the game screen displayed on the client side, and the resolution size of the game screen rendered at a lower resolution is the resolution size of the game screen rendered on the server side; Encoding the game screen rendered at a higher resolution and the game screen rendered at a lower resolution into a high-resolution video stream and a low-resolution video stream respectively; Decoding the high-resolution video stream and the low-resolution video stream into a high-resolution video and a low-resolution video respectively; Converting the high-resolution video and the low-resolution video into a high-resolution image sequence and a low-resolution image sequence respectively; The high-resolution image sequence and the low-resolution image sequence form a super-resolution data set; Cropping the super-resolution data set into overlapping image blocks of the same size; Determining the network structure and loss function to be used, training with the cropped image blocks, and finally generating a super-resolution model.

2. The method for improving the smoothness of cloud games according to claim 1, characterized in that Upsampling the decoded video screen using an interpolation upsampling algorithm to generate an upsampled game screen.

3. The method for improving the smoothness of cloud games according to claim 2, wherein Performing sharpening processing on the upsampled game screen and then using it for client-side display.

4. The method for improving the smoothness of cloud games according to claim 2, wherein The upsampling magnification of the interpolation upsampling algorithm is the ratio of the resolution size of the screen to be displayed on the client side to the resolution size of the game screen rendered on the server side.

5. The method for improving the smoothness of cloud games according to any one of claims 2-4, characterized in that, The interpolation upsampling algorithm is Lanczos interpolation, nearest neighbor interpolation, bilinear quadratic interpolation or bicubic interpolation.

6. A system for improving the smoothness of cloud games, characterized in that, Including: A receiving module for receiving a video stream of a game screen rendered by the server side at a low resolution; A decoding module for decoding the video stream to obtain a video screen; A sampling module for upsampling the decoded video screen to generate an upsampled super-resolution game screen; A display module for displaying the generated game screen on the client side. Specifically: using a super-resolution model to upsample the decoded video screen to generate a super-resolution game screen, and using the super-resolution game screen for client-side display. The training of the super-resolution model includes: Using the same controller to control two groups of cloud game servers to render game screens at different resolutions respectively; among them, the resolution size of the game screen rendered at a higher resolution is the resolution size of the game screen displayed on the client side, and the resolution size of the game screen rendered at a lower resolution is the resolution size of the game screen rendered on the server side; Encoding the game screen rendered at a higher resolution and the game screen rendered at a lower resolution into a high-resolution video stream and a low-resolution video stream respectively; Decode the high-resolution video stream and the low-resolution video stream into a high-resolution video and a low-resolution video respectively; Convert the high-resolution video and the low-resolution video into a high-resolution image sequence and a low-resolution image sequence respectively; The high-resolution image sequence and the low-resolution image sequence constitute a super-resolution data set; Crop the super-resolution data set into overlapping image patches of a unified size; Determine the network structure and loss function to be used, use the cropped image patches for training, and finally generate a super-resolution model.

7. A non-transitory computer-readable storage medium, characterized in that, The non-transitory computer-readable storage medium is used to store computer instructions, and when the computer instructions are executed by a processor, the method for improving the smoothness of cloud games according to any one of claims 1-5 is implemented.

8. An electronic device, characterized in that, Including: A processor, a memory, and a computer program; wherein, the processor is connected to the memory, the computer program is stored in the memory, and when the electronic device runs, the processor executes the computer program stored in the memory so that the electronic device executes the instructions for implementing the method for improving the smoothness of cloud games according to any one of claims 1-5.

Images