Video quality restoration method and device, electronic equipment and computer program product

Abstract

The invention relates to a video quality restoration method and device, electronic equipment and a computer program product, and belongs to the technical field of video processing. The method comprises the following steps: inputting a first quality video into a quality perception model to obtain a low-quality problem type of the first quality video and a low-quality degree coefficient corresponding to the low-quality problem type; aiming at each low-quality problem type and the low-quality degree coefficient corresponding to the low-quality problem type, respectively determining a corresponding target repair model; and repairing the first quality video through the target repairing model to obtain a repaired second quality video. According to the invention, the quality perception model is utilized to analyze the problem type and severity of the low-quality video, so that the video restoration process is guided, the processing effect for the low-quality problem of different dimensions can be improved, and better video watching experience is provided for a user.

Description

Technical Field

[0001] The present disclosure relates to the field of video processing technology, and in particular to a video quality restoration method, a video quality restoration device, an electronic device, and a computer program product. Background Art

[0002] In recent years, social media and short video platforms have developed rapidly, and massive amounts of user-generated videos have been watched and shared on these platforms.

[0003] Due to differences in shooting equipment and techniques, many videos have image quality issues, such as blur, block effects, noise, etc. These problems may appear alone or in combination in different videos, seriously affecting the user's viewing experience.

[0004] In view of this, there is an urgent need in the art for a video quality restoration method that can enhance video quality and provide users with a better video viewing experience.

[0005] It should be noted that the information disclosed in the above background technology section is only used to enhance the understanding of the background of the present disclosure, and therefore may include information that does not constitute the prior art known to ordinary technicians in the field. Summary of the invention

[0006] The purpose of the present disclosure is to provide a video quality restoration method, a video quality restoration device, an electronic device and a computer program product, thereby enhancing the video quality at least to a certain extent and providing users with a better video viewing experience.

[0007] According to a first aspect of the present disclosure, a video quality restoration method is provided, comprising:

[0008] Inputting the first quality video into the quality perception model to obtain the low-quality problem type of the first quality video and the low-quality degree coefficient corresponding to the low-quality problem type;

[0009] For each of the low-quality problem types and the low-quality degree coefficients corresponding to the low-quality problem types, respectively determine a corresponding target repair model;

[0010] The first quality video is repaired using the target repair model to obtain a repaired second quality video.

[0011] In an exemplary embodiment of the present disclosure, the quality perception model is trained based on the training data corresponding to each of the low-quality problem types;

[0012] The training data corresponding to each of the low-quality problem types includes each set of video data and a corresponding data label, and the data label is used to indicate a low-quality degree coefficient corresponding to the video data.

[0013] In an exemplary embodiment of the present disclosure, the quality perception model includes a quality perception model corresponding to each of the low-quality problem types.

[0014] In an exemplary embodiment of the present disclosure, the quality perception model includes a plurality of model branches, each of which is used to predict each of the low-quality problem types;

[0015] The quality perception model is trained based on the training data corresponding to all the low-quality problem types.

[0016] In an exemplary embodiment of the present disclosure, the data label is a numerical value within a preset interval.

[0017] In an exemplary embodiment of the present disclosure, the training data includes first training data and second training data;

[0018] The first training data includes real data corresponding to each of the low-quality question types, and data labels corresponding to the real data;

[0019] The second training data includes video data of target quality corresponding to each of the low-quality problem types and data labels corresponding to the video data of target quality, and the video data of target quality is obtained by processing with adjustable quality adjustment parameters. In an exemplary embodiment of the present disclosure, the target restoration model includes video restoration models corresponding to different low-quality degrees in each of the low-quality problem types.

[0020] According to a second aspect of the present disclosure, a video quality restoration device is provided, comprising:

[0021] A problem type prediction module is configured to input a first quality video into a quality perception model to obtain a low-quality problem type of the first quality video and a low-quality degree coefficient corresponding to the low-quality problem type;

[0022] A repair model determination module is configured to determine a corresponding target repair model for each of the low-quality problem types and the low-quality degree coefficient corresponding to the low-quality problem type;

[0023] The video quality restoration module is configured to perform restoration of the first quality video using the target restoration model to obtain a restored second quality video.

[0024] According to a third aspect of the present disclosure, an electronic device is provided, comprising: a processor; and a memory for storing instructions executable by the processor; wherein the processor is configured to execute the instructions to implement any one of the above-described video quality restoration methods.

[0025] According to a fourth aspect of the present disclosure, a computer program product is provided, including a computer program, wherein when the computer program is executed by a processor, the video quality restoration method described in any one of the above is implemented.

[0026] The exemplary embodiments of the present disclosure may have the following beneficial effects:

[0027] In the video quality repair method of the exemplary implementation mode of the present disclosure, by inputting the first quality video into the quality perception model, the low-quality problem type of the first quality video and the low-quality degree coefficient corresponding to the low-quality problem type are obtained, and then for each low-quality problem type and the low-quality degree coefficient corresponding to the low-quality problem type, the corresponding target repair model is determined respectively, and the first quality video is repaired by the target repair model. In the video quality repair method of the exemplary implementation mode of the present disclosure, on the one hand, the low-quality problems existing in the low-quality video are detected and analyzed in multiple dimensions by the quality perception model, and the low-quality problem type and the severity of the problem existing in the low-quality video are determined, so as to guide the subsequent video repair process and make the video repair more targeted; on the other hand, different repair models are used for joint processing for the existing problems. Since the number of parameters of the dedicated small model for specific image quality problems is small and the inference speed is fast, multiple specialized lightweight models are combined for processing. Under the condition of achieving the same video quality enhancement effect, the model scale and processing speed can be effectively reduced, computing resources are saved, and the processing effect for low-quality problems of different dimensions is improved, providing users with a better video viewing experience. At the same time, the model optimization goal is single and clear, and it is easier to train.

[0028] 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

[0029] The accompanying drawings herein are incorporated into the specification and constitute a part of the specification, illustrate embodiments consistent with the present disclosure, and together with the specification are used to explain the principles of the present disclosure. Obviously, the accompanying drawings described below are only some embodiments of the present disclosure, and for ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without creative work.

[0030] Figure 1 A schematic diagram showing a flow chart of a video quality restoration method according to an exemplary embodiment of the present disclosure;

[0031] Figure 2 A flowchart of a method for training a quality perception model according to an exemplary embodiment of the present disclosure is shown;

[0032] Figure 3A schematic diagram of a quality perception model of multiple tasks according to a specific embodiment of the present disclosure is schematically shown;

[0033] Figure 4 A schematic diagram of a process of a video quality restoration method based on quality perception in a specific implementation of the present disclosure is shown;

[0034] Figure 5 A schematic diagram schematically shows a video quality restoration result according to a specific implementation of the present disclosure;

[0035] Figure 6 A block diagram showing a video quality restoration apparatus according to an exemplary embodiment of the present disclosure is shown;

[0036] Figure 7 A schematic diagram of the structure of a computer system suitable for implementing the electronic device of the embodiment of the present disclosure is shown. DETAILED DESCRIPTION

[0037] 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.

[0038] 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 numbers 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.

[0039] The following example embodiments can be implemented in a variety of forms and should not be construed as being limited to the examples set forth herein; on the contrary, these embodiments are provided so that the present disclosure will be more comprehensive and complete, and the concepts of the example embodiments are fully conveyed to those skilled in the art. The described features, structures, or characteristics may be combined in one or more embodiments in any suitable manner. In the following description, many specific details are provided to provide a full understanding of the embodiments of the present disclosure. However, those skilled in the art will appreciate that the technical solutions of the present disclosure may be practiced while omitting one or more of the specific details, or other methods, components, devices, steps, etc. may be adopted. In other cases, known technical solutions are not shown or described in detail to avoid obscuring various aspects of the present disclosure.

[0040] In addition, the accompanying drawings are only schematic illustrations of the present disclosure and are not necessarily drawn to scale. The same reference numerals in the figures represent the same or similar parts, and their repeated description will be omitted. Some of the block diagrams shown in the accompanying drawings are functional entities and do not necessarily correspond to physically or logically independent entities. These functional entities can be implemented in software form, or implemented in one or more hardware modules or integrated circuits, or implemented in different networks and / or processor devices and / or microcontroller devices.

[0041] In some related embodiments, a blind restoration solution may be used to address the video quality restoration problem. The blind restoration solution generates high-quality and low-quality training data pairs by online mixed degradation, and then trains a restoration model to process all different types of low-quality videos, thereby obtaining a video blind restoration result.

[0042] The blind repair solution in the above-mentioned related embodiments uses one model to handle all different types of low-quality problems, which may cause the following problems:

[0043] (1) The unified model requires a large number of model parameters, consumes a lot of computing resources, and has a relatively slow calculation speed.

[0044] (2) The unified hybrid repair model requires multi-objective optimization, which makes model training difficult and it is difficult to achieve good results for all low-quality problems.

[0045] (3) The model cannot adjust the processing intensity according to the low quality of the input content to achieve the best image quality.

[0046] Based on the above problems, this exemplary embodiment first provides a video quality repair method. Figure 1 As shown, the above video quality repair method may include the following steps:

[0047] Step S110: Input the first quality video into the quality perception model to obtain the low-quality problem type of the first quality video and the low-quality degree coefficient corresponding to the low-quality problem type.

[0048] Step S120: for each low-quality problem type and the low-quality degree coefficient corresponding to the low-quality problem type, determine the corresponding target repair model respectively.

[0049] Step S130: Repair the first quality video using the target repair model to obtain a repaired second quality video.

[0050] In the video quality repair method of the exemplary implementation mode of the present disclosure, by inputting the first quality video into the quality perception model, the low-quality problem type of the first quality video and the low-quality degree coefficient corresponding to the low-quality problem type are obtained, and then for each low-quality problem type and the low-quality degree coefficient corresponding to the low-quality problem type, the corresponding target repair model is determined respectively, and the first quality video is repaired by the target repair model. In the video quality repair method of the exemplary implementation mode of the present disclosure, on the one hand, the low-quality problems existing in the low-quality video are detected and analyzed in multiple dimensions by the quality perception model, and the low-quality problem type and the severity of the problem existing in the low-quality video are determined, so as to guide the subsequent video repair process and make the video repair more targeted; on the other hand, different repair models are used for joint processing for the existing problems. Since the number of parameters of the dedicated small model for specific image quality problems is small and the inference speed is fast, multiple specialized lightweight models are combined for processing. Under the condition of achieving the same video quality enhancement effect, the model scale and processing speed can be effectively reduced, computing resources are saved, and the processing effect for low-quality problems of different dimensions is improved, providing users with a better video viewing experience. At the same time, the model optimization goal is single and clear, and it is easier to train.

[0051] Next, combine Figures 2 to 5 The above steps of this exemplary embodiment are described in more detail.

[0052] In step S110, the first quality video is input into a quality perception model to obtain a low-quality problem type of the first quality video and a low-quality degree coefficient corresponding to the low-quality problem type.

[0053] In this example implementation, the first quality video refers to a low-quality video with low video clarity or poor video quality. The quality perception model is used to analyze the types of low-quality problems and the corresponding low-quality degrees in the low-quality video, where the low-quality problem types may include noise, block effects, blur, oversharpness and other problem types. The low-quality degree coefficient is used to reflect the severity of the problem corresponding to a certain problem type of the low-quality video, which can be a specific value, and the larger the value, the more serious the problem.

[0054] In this example implementation, Figure 2 As shown in FIG. 1 , the training method of the quality perception model may specifically include the following steps:

[0055] Step S210. Construct corresponding training data for each low-quality problem type, wherein each set of video data in the training data carries a corresponding data label, and the data label is used to indicate the low-quality degree coefficient corresponding to the video data.

[0056] In this example implementation, the quality perception model is trained based on the training data corresponding to each low-quality problem type, wherein the training data corresponding to each low-quality problem type includes each set of video data and a corresponding data label, and the data label is used to represent the low-quality degree coefficient corresponding to the video data.

[0057] In the training of the quality perception model, we first determine all the types of low-quality problems to be analyzed, such as noise, blockiness, blur, oversharpness, etc. Then, we construct labeled training data for each type of low-quality problem. The data label is the severity of the low-quality problem, which can be a value within a preset range, for example, it can be normalized to a floating point number between 0 and 1.

[0058] In this example implementation, the training data includes first training data and second training data. The first training data includes real data corresponding to each low-quality problem type and data labels corresponding to the real data; the second training data includes video data of target quality corresponding to each low-quality problem type and data labels corresponding to the video data of target quality, and the video data of target quality is obtained by processing with adjustable quality adjustment parameters.

[0059] The original data of the training data for building the quality perception model can include two parts: real data and synthetic data. The real data can be annotated by professionals to obtain the first training data. The synthetic data can be simulated by adding noise, blur and other quality adjustment parameters with controllable parameters to the high-quality data by the algorithm to obtain the corresponding low-quality data, and then the data is annotated to obtain the second training data.

[0060] Step S220: Train a quality perception model based on the training data corresponding to each low-quality problem type.

[0061] After constructing the training data, the labeled training data can be used to train the quality-aware model.

[0062] In this example implementation, the quality-aware model includes a quality-aware model corresponding to each low-quality problem type.

[0063] For each low-quality problem type, a separate quality-aware model can be trained based on the training data of that type, specifically for detecting that low-quality problem type and predicting the severity of the problem under that type.

[0064] In this example implementation, the quality perception model may also include multiple model branches, each model branch is used to predict a respective low-quality problem type; the quality perception model is trained based on the training data corresponding to all low-quality problem types.

[0065] For all low-quality problem types, a comprehensive quality-aware model can be trained to predict the extent of different low-quality problems in the video.

[0066] Figure 3 A schematic diagram of a multi-task quality perception model according to a specific embodiment of the present disclosure is schematically shown. The multi-task quality perception model adopts a multi-task training method to obtain a multi-branch perception model, and each branch is used to predict different low-quality problem types and the problem severity of the problem type.

[0067] In step S120 , for each low-quality problem type and the low-quality degree coefficient corresponding to the low-quality problem type, a corresponding target repair model is determined respectively.

[0068] In this example implementation, a corresponding video restoration model may be trained for each low-quality problem type and each low-quality degree in the low-quality problem type.

[0069] In order to reduce the model size and improve the restoration effect, each video restoration model is only used to restore one type of low-quality problem. Different degrees of the same low-quality problem use different video restoration models. For example, for strong noise, a stronger noise suppression model is needed to achieve the purpose of denoising, while for weak noise, a weaker noise suppression model is used to avoid damaging video details. Therefore, for each type of low-quality problem, multiple restoration models of different degrees can be trained.

[0070] In step S130, the first quality video is restored by using the target restoration model to obtain a restored second quality video.

[0071] The second quality video refers to a high-quality video with higher video definition or better video quality. By repairing the first quality video, a repaired second quality video can be obtained.

[0072] In this example implementation, for a low-quality video to be processed, a quality perception model is first used to analyze the low-quality problems and degrees, and then the corresponding lightweight target repair models are selected according to the problems and degrees to process them in sequence, thereby achieving the best repair and enhancement effects.

[0073] like Figure 4 The figure is a complete flow chart of a quality-aware video quality restoration method in a specific implementation of the present disclosure, which is an example of the above steps in this example implementation. The specific steps of the flow chart are as follows:

[0074] First, a low-quality video is input into the quality perception model 401 to obtain the low-quality problems and degrees of the low-quality video. Then, according to the problems and degrees, a corresponding target restoration model is selected from the video restoration model 402. For example, in the noise suppression model, a corresponding noise suppression model (such as a medium noise suppression model) is selected based on the different degrees of the noise problem. In the deblurring model, a corresponding deblurring model (such as a strong deblurring model) is selected based on the different degrees of the blurring problem. In the deblocking model, a corresponding deblocking effect model (such as a weak deblocking effect model) is selected based on the different degrees of the blocking effect problem. Then, the low-quality video is processed in turn according to the selected lightweight models to obtain the final restored video.

[0075] Figure 5 The schematic diagram of the video quality restoration result according to a specific embodiment of the present disclosure is schematically shown. The left figure is the input low-quality video. The middle figure is the restoration model after processing using a single model, and its output result is too sharp. The right figure is the restoration model after processing using the video quality restoration method in this example embodiment, which can sense the sharpness of the picture and automatically select the model, and the output result is more natural.

[0076] It should be noted that although the steps of the method in the present disclosure are described in a specific order in the drawings, this does not require or imply that the steps must be performed in this specific order, or that all the steps shown must be performed to achieve the desired results. Additionally or alternatively, some steps may be omitted, multiple steps may be combined into one step, and / or one step may be decomposed into multiple steps, etc.

[0077] Furthermore, the present disclosure also provides a video quality restoration device. Figure 6 As shown, the video quality repair device may include a problem type prediction module 610, a repair model determination module 620, and a video quality repair module 630. Among them:

[0078] The problem type prediction module 610 is configured to input the first quality video into the quality perception model to obtain the low-quality problem type of the first quality video and the low-quality degree coefficient corresponding to the low-quality problem type;

[0079] The repair model determination module 620 is configured to determine the corresponding target repair model for each low-quality problem type and the low-quality degree coefficient corresponding to the low-quality problem type;

[0080] The video quality restoration module 630 is configured to perform restoration on the first quality video through a target restoration model to obtain a restored second quality video.

[0081] In some exemplary embodiments of the present disclosure, the present disclosure further provides a quality perception model training device, which may include a training data construction unit and a quality perception model training unit.

[0082] A training data construction unit is configured to construct corresponding training data for each low-quality problem type, wherein each set of video data in the training data has a corresponding data label, and the data label is used to indicate a low-quality degree coefficient corresponding to the video data;

[0083] The quality perception model training unit is configured to execute training data corresponding to each low-quality problem type to train a quality perception model.

[0084] In some exemplary embodiments of the present disclosure, the quality-aware model training apparatus may include a single quality-aware model training unit configured to execute training data corresponding to each low-quality problem type to obtain a quality-aware model corresponding to the low-quality problem type.

[0085] In some exemplary embodiments of the present disclosure, the quality-aware model training device may further include a comprehensive quality-aware model training unit, which is configured to execute training data corresponding to all low-quality problem types, and jointly train to obtain a quality-aware model, wherein the quality-aware model includes multiple model branches, and each model branch is used to predict each low-quality problem type.

[0086] In some exemplary embodiments of the present disclosure, the training data construction unit may include a first training data construction unit, a second training data construction unit, and a training data construction unit.

[0087] A first training data construction unit is configured to obtain real data corresponding to each low-quality question type, and add corresponding data labels to the real data to obtain first training data;

[0088] A second training data construction unit is configured to obtain target quality video data for each low-quality problem type, and process the target quality video data through an adjustable quality adjustment parameter to obtain second training data;

[0089] The training data constructing unit is configured to obtain training data according to the first training data and the second training data.

[0090] In some exemplary embodiments of the present disclosure, a video quality restoration device provided by the present disclosure may also include a video restoration model training module, which is configured to execute training of corresponding video restoration models for each low-quality problem type and each low-quality degree in the low-quality problem type.

[0091] The specific details of each module / unit in the above-mentioned video quality restoration device and quality perception model training device have been described in detail in the corresponding method embodiment part and will not be repeated here.

[0092] Figure 7 A schematic diagram of the structure of a computer system suitable for implementing an electronic device of an embodiment of the present disclosure is shown.

[0093] It should be noted that Figure 7 The computer system 700 of the electronic device shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present disclosure.

[0094] like Figure 7 As shown, computer system 700 includes a central processing unit (CPU) 701, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 702 or a program loaded from a storage part 708 into a random access memory (RAM) 703. Various programs and data required for system operation are also stored in RAM 703. CPU 701, ROM 702 and RAM 703 are connected to each other via a bus 704. An input / output (I / O) interface 705 is also connected to bus 704.

[0095] The following components are connected to the I / O interface 705: an input section 706 including a keyboard, a mouse, etc.; an output section 707 including a cathode ray tube (CRT), a liquid crystal display (LCD), etc., and a speaker, etc.; a storage section 708 including a hard disk, etc.; and a communication section 709 including a network interface card such as a LAN card, a modem, etc. The communication section 709 performs communication processing via a network such as the Internet. A drive 710 is also connected to the I / O interface 705 as needed. A removable medium 711, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc., is installed on the drive 710 as needed, so that a computer program read therefrom is installed into the storage section 708 as needed.

[0096] In particular, according to an embodiment of the present disclosure, the process described above with reference to the flowchart can be implemented as a computer software program. For example, an embodiment of the present disclosure includes a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program includes a program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from a network through a communication section 709, and / or installed from a removable medium 711. When the computer program is executed by a central processing unit (CPU) 701, various functions defined in the system of the present disclosure are executed.

[0097] The exemplary embodiments of the present disclosure also provide a computer program product, which includes a computer program, and when the computer program is executed by a processor, the video quality restoration method is implemented.

[0098] In one embodiment, the computer program product may be a tangible product containing a computer program, such as a computer-readable storage medium storing a computer program. The readable storage medium may be a storage medium based on electrical, magnetic, optical, electromagnetic, infrared, or other signals, including but not limited to: random access memory (RAM), read-only memory (ROM), magnetic tape, floppy disk, flash memory (Flash), mechanical hard disk (HDD), solid-state drive (SSD), and the like. Exemplarily, the computer program product may be implemented as a non-volatile storage medium storing a computer program, such as a read-only memory, a NAND flash memory, and the like.

[0099] In one embodiment, the computer program product may be an intangible product including a computer program. Exemplarily, the computer program product may be implemented as a virtual digital product, such as a digital file storing an executable file, an installation package, etc. of the computer program.

[0100] The code of the computer program can be written in one or more programming languages. Programming languages ​​such as C language, Java, C++, etc. The program code can be executed entirely on the user computing device, or partially on the user computing device, or as a separate software package, or partially on the user computing device and partially on a remote computing device, or entirely on a remote computing device or server. In the case of a remote computing device, the remote computing device can be connected to the user computing device through any type of network, such as a local area network (LAN), a wide area network (WAN), etc., or can be connected to an external computing device (e.g., an Internet connection provided by an operator).

[0101] The computer program may be carried or transmitted via electrical, magnetic, optical, electromagnetic, infrared, or other signals. The electronic device may convert the signal carrying the computer program into a digital signal, thereby running the computer program. When the computer program is run on an electronic device, its code is used to enable the electronic device to execute (more specifically, the processor of the electronic device may execute) the method steps of various exemplary embodiments of the present disclosure, such as the above-mentioned video quality restoration method.

[0102] The flow charts and block diagrams in the accompanying drawings illustrate the possible architecture, functions and operations of the systems, methods and computer program products according to various embodiments of the present disclosure. In this regard, each box in the flow chart or block diagram can represent a module, a program segment, or a part of a code, and the above-mentioned module, program segment, or a part of a code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in a different order from the order marked in the accompanying drawings. For example, two boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram or flow chart, and the combination of the boxes in the block diagram or flow chart can be implemented with a dedicated hardware-based system that performs a specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0103] It should be noted that, although several modules of the device for action execution are mentioned in the above detailed description, this division is not mandatory. In fact, according to the embodiments of the present disclosure, the features and functions of two or more modules described above can be concretized in one module. Conversely, the features and functions of one module described above can be further divided into multiple modules to be concretized.

[0104] 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 technical means in the art that are not disclosed in the present disclosure.

[0105] 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. A video quality restoration method, characterized in that: include: Inputting the first quality video into the quality perception model to obtain the low-quality problem type of the first quality video and the low-quality degree coefficient corresponding to the low-quality problem type; For each of the low-quality problem types and the low-quality degree coefficients corresponding to the low-quality problem types, respectively determine a corresponding target repair model; The first quality video is repaired using the target repair model to obtain a repaired second quality video.

2. The video quality restoration method according to claim 1, characterized in that: The quality perception model is obtained by training based on the training data corresponding to each of the low-quality problem types; The training data corresponding to each of the low-quality problem types includes each set of video data and a corresponding data label, and the data label is used to indicate a low-quality degree coefficient corresponding to the video data.

3. The video quality restoration method according to claim 2, characterized in that: The quality-aware model includes a quality-aware model corresponding to each of the low-quality problem types.

4. The video quality restoration method according to claim 2, characterized in that: The quality perception model comprises a plurality of model branches, each of which is used to predict each of the low-quality problem types; The quality perception model is trained based on the training data corresponding to all the low-quality problem types.

5. The video quality restoration method according to claim 2, characterized in that: The data label is a value within a preset range.

6. The video quality restoration method according to claim 2, characterized in that: The training data includes first training data and second training data; The first training data includes real data corresponding to each of the low-quality question types, and data labels corresponding to the real data; The second training data includes video data of target quality corresponding to each of the low-quality problem types and data labels corresponding to the video data of target quality, and the video data of target quality is obtained by processing with adjustable quality adjustment parameters.

7. The video quality restoration method according to claim 1, characterized in that: The target restoration model includes video restoration models corresponding to different low-quality degrees in each of the low-quality problem types.

8. A video quality restoration device, characterized in that: include: A problem type prediction module is configured to input a first quality video into a quality perception model to obtain a low-quality problem type of the first quality video and a low-quality degree coefficient corresponding to the low-quality problem type; A repair model determination module is configured to determine a corresponding target repair model for each of the low-quality problem types and the low-quality degree coefficient corresponding to the low-quality problem type; The video quality restoration module is configured to perform restoration of the first quality video using the target restoration model to obtain a restored second quality video.

9. 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 video quality restoration method according to any one of claims 1 to 7.

10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the video quality restoration method according to any one of claims 1 to 7 is implemented.

Images