Video quality assessment method, device, computer equipment and storage medium

Abstract

The present application relates to artificial intelligence technology and provides a video quality assessment method, device, computer equipment and storage medium. The method comprises: playing a video file through a first thread; collecting video images in the video file being played; evaluating the collected video images through a second thread to obtain an image quality score; and determining a video quality assessment result corresponding to the video file according to the image quality score. The method can improve the accuracy of video quality assessment.

Description

Technical Field

[0001] The present application relates to the field of multimedia technology, and in particular to a video quality assessment method, apparatus, computer equipment and storage medium. Background Art

[0002] With the development of multimedia technology, there are more and more video files to choose from. The video quality of video files and the image quality of video images in video files will directly affect users' choice of video files and their subjective feelings when watching videos. Therefore, how to evaluate the quality of video files and video images in video files is a problem worthy of attention.

[0003] Currently, source video files are usually evaluated in the background by a server, which has the problem of low evaluation accuracy. Summary of the invention

[0004] Based on this, it is necessary to provide a video quality assessment method, apparatus, computer equipment and storage medium that can improve the accuracy of video quality assessment in response to the above technical problems.

[0005] A video quality assessment method, the method comprising:

[0006] Play the video file through the first thread;

[0007] Capture video images from the video file being played;

[0008] The collected video image is evaluated by a second thread to obtain an image quality score;

[0009] A video quality assessment result corresponding to the video file is determined according to the image quality score.

[0010] A video quality assessment device, comprising:

[0011] A playing module, used for playing the video file through the first thread;

[0012] An acquisition module, used for acquiring video images from a video file being played;

[0013] An evaluation module, used for evaluating the collected video image through a second thread to obtain an image quality score;

[0014] An evaluation result determination module is used to determine a video quality evaluation result corresponding to the video file according to the image quality score.

[0015] A computer device comprises a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the following steps are implemented:

[0016] Play the video file through the first thread;

[0017] Capture video images from the video file being played;

[0018] The collected video image is evaluated by a second thread to obtain an image quality score;

[0019] A video quality assessment result corresponding to the video file is determined according to the image quality score.

[0020] A computer-readable storage medium stores a computer program, which, when executed by a processor, implements the following steps:

[0021] Play the video file through the first thread;

[0022] Capture video images from the video file being played;

[0023] The collected video image is evaluated by a second thread to obtain an image quality score;

[0024] A video quality assessment result corresponding to the video file is determined according to the image quality score.

[0025] The above-mentioned video quality assessment method, device, computer equipment and storage medium collect the video image to be evaluated from the video file being played by the first thread, and evaluate the collected video image through the second thread to obtain the image quality score, and then determine the video quality assessment result corresponding to the video file being played according to the image quality score corresponding to the collected video image. In this way, what is evaluated by the second thread is the video image being played, that is, the video image that the user can directly watch. Correspondingly, the image quality score obtained by the evaluation is also a quality score that conforms to the user's subjective feelings. Therefore, on the basis of realizing the real-time evaluation of the image quality score, the evaluation accuracy of the image quality score is improved. Based on the image quality score with higher accuracy, a video quality assessment result with higher accuracy can be obtained, thereby improving the evaluation accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] Figure 1 A diagram showing an application environment of a video quality assessment method in an embodiment;

[0027] Figure 2 is a flow chart of a video quality assessment method in one embodiment;

[0028] Figure 3 is a flow chart of a video quality assessment method in another embodiment;

[0029] Figure 4A schematic diagram of a principle of evaluating video quality through a video player in one embodiment;

[0030] Figure 5 A schematic diagram of an interface for evaluating video quality and displaying video quality evaluation results in one embodiment;

[0031] Figure 6 A schematic diagram of an interface for manually scoring a target image in one embodiment;

[0032] Figure 7 is a structural block diagram of a video quality assessment device in one embodiment;

[0033] Figure 8 FIG. 4 is a diagram showing the internal structure of a computer device in one embodiment. DETAILED DESCRIPTION

[0034] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0035] Artificial Intelligence (AI) is the theory, method, technology and application system that uses digital computers or machines controlled by digital computers to simulate, extend and expand human intelligence, perceive the environment, acquire knowledge and use knowledge to obtain the best results. In other words, artificial intelligence is a comprehensive technology in computer science that attempts to understand the essence of intelligence and produce a new intelligent machine that can respond in a similar way to human intelligence. Artificial intelligence is to study the design principles and implementation methods of various intelligent machines so that machines have the functions of perception, reasoning and decision-making.

[0036] Artificial intelligence technology is a comprehensive discipline that covers a wide range of fields, including both hardware-level and software-level technologies. The basic technologies of artificial intelligence generally include sensors, dedicated artificial intelligence chips, cloud computing, distributed storage, big data processing technology, operation / interaction systems, mechatronics and other technologies. Artificial intelligence software technology mainly includes computer vision technology, speech processing technology, natural language processing technology, and machine learning / deep learning.

[0037] Furthermore, Machine Learning (ML) is a multi-disciplinary subject involving probability theory, statistics, approximation theory, convex analysis, algorithm complexity theory and other disciplines. It specializes in studying how computers simulate or implement human learning behavior to acquire new knowledge or skills and reorganize existing knowledge structures to continuously improve their performance. Machine learning is the core of artificial intelligence and the fundamental way to make computers intelligent. Its applications are spread across all areas of artificial intelligence. Machine learning and deep learning usually include artificial neural networks, belief networks, reinforcement learning, transfer learning, inductive learning, and self-learning.

[0038] The video quality assessment method provided in this application can be applied to Figure 1 In the application environment shown. Among them, the terminal 102 communicates with the server 104 through the network. While the terminal 102 plays the video file through the first thread, the terminal 102 collects the video image in the video file being played, evaluates the collected video image through the second thread, obtains the image quality score, and determines the video quality evaluation result corresponding to the video file according to the image quality score corresponding to the collected video image. Among them, the terminal 102 can be but is not limited to various personal computers, laptops, smart phones, tablet computers and portable wearable devices, and the server 104 can be implemented by an independent server or a server cluster composed of multiple servers.

[0039] In one embodiment, Figure 2 As shown, a video quality assessment method is provided, which is applied to Figure 1 The terminal in is used as an example to illustrate, including the following steps:

[0040] S202: Play the video file through the first thread.

[0041] A thread is a single sequential control process in a computer program and is a basic unit that can run and schedule independently. A video file refers to a source video file to be played. The video file can be a local file or a network file, which is not specifically limited here.

[0042] Specifically, the terminal obtains a video file to be played, and plays the video file through a first thread.

[0043] In one embodiment, the terminal may obtain the video file to be played locally, may obtain the video file from an online network such as a browser and a website, or may obtain the video file to be played from a server.

[0044] In one embodiment, the terminal obtains and analyzes the video file to be played, obtains multiple video images corresponding to the video file and sorted in the play time sequence, and plays the obtained video images in sequence through the first thread. It is understandable that the terminal can obtain multiple video images corresponding to the video file through a single analysis before playing the video file, and can also obtain and play each video image in the video file by performing analysis and playback synchronously during the process of playing the video file.

[0045] In one embodiment, the terminal calls the first thread to play the video file through a video player. The video player may be a client running on the terminal.

[0046] In one embodiment, the terminal detects the video playback instruction for the video file in real time, and when the video playback instruction is detected, the video file is played through the first thread. The terminal can obtain the video file in advance, and display the playback entry information corresponding to the video file, and then detect the video playback instruction for the video file. The terminal can also obtain and display the playback entry information corresponding to the video file in advance, and when the video playback instruction for the video file is detected, the corresponding video file is obtained according to the video playback instruction, and then the obtained video file is played through the first thread. The playback entry information is the entry information that can be used to access and play the corresponding video file, and specifically may include at least one of the promotional video image, video introduction, video title and video link of the video file. It can be understood that the terminal can specifically obtain the video file through the video player and display the corresponding playback entry information of the video file.

[0047] S204, capturing video images in the video file being played.

[0048] The video image is an image extracted from a video file and can be directly displayed on a display screen. The video image can specifically refer to a video frame or an image frame in a video file.

[0049] Specifically, during the process of playing the video file through the first thread, the terminal collects the video file to be evaluated from the video file being played.

[0050] In one embodiment, during the process of playing a video file through the first thread, the terminal may collect the video images played through the first thread in the video file in real time as the video images to be evaluated, or may collect the video images played through the first thread from the video file by sampling as the video images to be evaluated. It is understood that the sampling may be performed according to a preset sampling frequency or a preset sampling period, or may be performed by collecting one or more video images from a preset position of the video file, such as collecting one video image from the start position and the end position of the video file, respectively, which is not specifically limited here.

[0051] In one embodiment, when the terminal plays a video image in a video file on a display screen through a first thread, the video image being played is collected as the video image to be evaluated. The terminal may also collect the video image to be played as the video image to be evaluated after parsing the video file to obtain the video image to be played and before playing the video image through the first thread.

[0052] S206, evaluating the collected video image through the second thread to obtain an image quality score.

[0053] The image quality score refers to the score value obtained by performing a quality assessment on a video image and is used to characterize the image quality. Specifically, it can be used to characterize the subjective quality of the video image directly perceived by the user. The value range of the image quality score can be customized, such as 1 to 5 points, or 1 to 10 points, etc. It can be understood that the "first thread" and "second thread" in this application are only used to distinguish between the thread for playing video files and the thread for evaluating the captured video images, and are not used for specific limitations.

[0054] Specifically, the terminal evaluates the video image collected from the video file being played through the second thread to obtain an image quality score corresponding to the video image.

[0055] In one embodiment, the terminal evaluates the image quality scores of the collected video images in sequence according to the collection order of the video images through the second thread. The terminal evaluates the collected video images in real time through the second thread, so as to achieve synchronous evaluation of the video images being played while playing the video images through the first thread.

[0056] In one embodiment, the terminal calls a trained image quality assessment model through a second thread, and uses the image quality assessment model to assess the captured video image to obtain a corresponding image quality score. The image quality assessment model is a model trained based on a pre-acquired training sample set and can be used to assess the image quality score of the video image. The training sample set includes a target image and a corresponding target image quality score manually annotated.

[0057] In one embodiment, after the terminal captures the video image to be evaluated from the video file being played through the video player, the terminal calls the second thread through the video player to evaluate the image quality score of the captured video image.

[0058] In one embodiment, after evaluating and obtaining the image quality score of the captured video image, the terminal records the obtained image quality score locally corresponding to the video file, so as to determine the video quality evaluation result of the corresponding video file according to the recorded image quality score.

[0059] S208: Determine a video quality assessment result corresponding to the video file according to the image quality score.

[0060] The video quality assessment result refers to an assessment result obtained by performing a quality assessment on a currently playing video file and used to characterize the video quality. Specifically, it can be used to characterize the subjective quality of the video file intuitively perceived by the user during the playback of the video file.

[0061] Specifically, the terminal analyzes the image quality score corresponding to the video image collected from the video file being played, and obtains the video quality evaluation result corresponding to the video file.

[0062] In one embodiment, after evaluating and obtaining the image quality score of the currently captured video image, the terminal determines the video quality score corresponding to the video file according to the image quality scores corresponding to the respective video images already captured from the video file. The terminal records the image quality score obtained by evaluating the video image locally in a cache manner, and determines the video quality evaluation result of the corresponding video file according to the recorded image quality score.

[0063] In one embodiment, the terminal can dynamically determine the video quality assessment result corresponding to the video file according to the image quality score of the collected video image during the playback of the video file. The terminal can also determine the video quality assessment result corresponding to the video file according to the image quality score corresponding to the video image pixels collected from the video file when the video file is finished playing.

[0064] The above-mentioned video quality assessment method collects the video image to be evaluated from the video file being played by the first thread, and evaluates the collected video image through the second thread to obtain the image quality score, and then determines the video quality assessment result corresponding to the video file being played according to the image quality score corresponding to the collected video image. In this way, what is evaluated by the second thread is the video image being played, that is, the video image that the user can directly watch. Correspondingly, the image quality score obtained by the evaluation is also a quality score that conforms to the user's subjective feelings, thereby improving the evaluation accuracy of the image quality score. Based on the image quality score with higher accuracy, a video quality assessment result with higher accuracy can be obtained, thereby improving the evaluation accuracy.

[0065] In one embodiment, step S204 includes: acquiring the video image to be evaluated from the video file being played according to a preset sampling frequency.

[0066] Among them, the preset sampling frequency is the frequency of sampling the video images to be evaluated from the video images in the video file, which can be customized according to actual needs. The preset sampling frequency can specifically refer to selecting a video image to be evaluated every preset number of video images, that is, from the video images continuously played through the first thread, one video image to be evaluated is selected every preset number of video images, and the preset number is 5, for example. The preset sampling frequency can also refer to collecting a video image to be evaluated every preset time length, and the preset time length is 2 seconds. The preset sampling frequency can also refer to the number of video images to be evaluated sampled per second, such as 1 image / second.

[0067] Specifically, the terminal is locally configured with a preset sampling frequency, and collects the video image to be evaluated from the video file being played through the first thread according to the preset sampling frequency.

[0068] In one embodiment, during the process of sequentially playing video images in a video file through a first thread, the terminal collects video images to be evaluated from video images that are being played or to be played through the first thread according to a preset sampling frequency.

[0069] In one embodiment, the terminal collects the video image to be evaluated from the video file being played by the video player according to a preset sampling frequency through the video player. In this way, the video image to be evaluated is collected by sampling, which can reduce the CPU (central processing unit) operating pressure of the video player.

[0070] In the above embodiment, the video images to be evaluated are collected by sampling to avoid the evaluation operation on the video images affecting the normal playback of the video files. By setting a preset sampling frequency, it is possible to ensure that the image quality score based on the sampled video images can obtain a relatively accurate video quality evaluation result, thereby improving the accuracy of video quality evaluation while reducing the pressure of video image evaluation.

[0071] In one embodiment, step S208 includes: when the video file is finished playing, determining a video quality assessment result corresponding to the video file according to an image quality score corresponding to a video image collected from the video file; and displaying the video quality assessment result.

[0072] Specifically, the terminal detects the progress of the video file being played in real time. When it is detected that the video file has finished playing, the terminal analyzes the image quality scores corresponding to each video image collected from the video file to obtain the video quality evaluation result corresponding to the video file. The terminal displays the video quality evaluation result corresponding to the video file on a display screen.

[0073] In one embodiment, when it is detected that the video file has been played normally, or when a video switching instruction or a video exit instruction is detected that is triggered for the video file being played, the terminal determines that the video file has been played. The video switching instruction is used to instruct the terminal to switch from the video file being played to another video file. The video exit instruction is used to instruct the terminal to end the playback process of the video file currently being played. The terminal can also determine that the video file being played has been played when a player exit instruction for the video player is detected. The player exit instruction is used to instruct the terminal to exit the video player and end the playback of the video file currently being played.

[0074] In one embodiment, the terminal displays the video quality evaluation result corresponding to the currently played video file to the user through the display screen, and can also store the video quality evaluation result locally corresponding to the video file. It can be understood that when the video file is finished playing, the terminal obtains the image quality score evaluated and recorded for the video image collected from the video file, summarizes and analyzes the obtained image quality score, and obtains the video quality evaluation result corresponding to the video file. The video quality evaluation result can be used to characterize the subjective video quality of the video file played by the video player this time.

[0075] In the above embodiment, when the video file is finished playing, a relatively accurate video quality evaluation result corresponding to the video file can be obtained according to the image quality score corresponding to the video image collected from the video file, so as to realize the imperceptible evaluation of the video quality, thereby improving the efficiency and accuracy of the video quality evaluation. Furthermore, by displaying the video quality evaluation result, the user can quickly and intuitively obtain the video quality of the currently finished video file.

[0076] In one embodiment, a video quality assessment result corresponding to a video file is determined based on an image quality score corresponding to a video image collected from a video file, including: analyzing the image quality score corresponding to each video image collected from the video file to obtain a maximum score, a minimum score, an average score, and a score variance; and determining a video quality assessment result corresponding to the video file based on the maximum score, the minimum score, the average score, and the score variance.

[0077] Specifically, when it is determined that the video file being played has been played, the terminal obtains the image quality scores corresponding to each video image collected from the video file, and performs statistical analysis on the obtained image quality scores to obtain the highest score and the lowest score among the obtained image quality scores, as well as the corresponding average score and score variance. The terminal can directly use the highest score, the lowest score, the average score and the score variance obtained by analysis as the video quality assessment result corresponding to the video file. The terminal can also average the highest score, the lowest score, the average score and the score variance obtained by analysis, and use the calculated mean as the video quality assessment result corresponding to the video file. It can be understood that the average can be an arithmetic average or a weighted average, and the weight of the weighted average can be customized and is not specifically limited here.

[0078] In one embodiment, when the video file is played, the terminal obtains the image quality scores corresponding to each of the multiple video images collected from the video file, and obtains multiple image quality scores. The terminal selects the highest score and the lowest score from the multiple image quality scores, and calculates the average score and score variance of the multiple image quality scores. It can be understood that if the terminal only collects one video image from the video file when the video file is played, the video quality evaluation result of the video file is determined in the above manner according to the image quality score of the video image.

[0079] In the above embodiment, the image quality scores corresponding to the respective video images collected from the video file are analyzed to obtain the video quality assessment result corresponding to the video file, which can improve the accuracy of video quality assessment.

[0080] In one embodiment, step S202 includes: parsing the video file to obtain an initial video image; performing image post-processing on the initial video image to obtain a video image to be played; playing the video image according to the playback sequence through the first thread; step S204 includes: collecting the video image to be evaluated from the video image played according to the playback sequence.

[0081] The initial video image is a video image obtained by parsing the video file and has not been subjected to image post-processing. Image post-processing refers to adjusting the initial video image obtained by parsing the video file, which may include at least one of brightening and super-resolution, which is not specifically limited here.

[0082] Specifically, after obtaining the video file to be played, the terminal parses the obtained video file to obtain the corresponding initial video image, and performs image post-processing on each of the obtained initial video images to obtain the corresponding video image to be played. After obtaining the video image to be played corresponding to the video file, the terminal plays the obtained video images in sequence according to the play sequence through the first thread. In the process of playing the video images according to the play sequence through the first thread, the terminal collects the video images to be evaluated from the video images played according to the play sequence.

[0083] In one embodiment, the terminal parses the video file in real time to obtain one or more corresponding initial video images, performs image post-processing on the initial video images obtained by the analysis in sequence according to the playback sequence to obtain the video images to be played, and plays the video images to be played in real time according to the playback sequence through the first thread. Correspondingly, the terminal collects the video images to be evaluated from the video images played according to the playback sequence, so as to synchronously evaluate the video images being played through the second thread. In this way, while the terminal plays the video images that are earlier in the playback sequence through the first thread, it parses the video file in real time to obtain the video images that are later in the playback sequence, and evaluates the image quality scores of the video images being played by the first thread in parallel through the second thread, so as to achieve imperceptible evaluation of the currently played video images while ensuring the smoothness of the video file playback, thereby achieving the evaluation of the video file.

[0084] In one embodiment, the step of the terminal parsing the video file to obtain the initial video image includes: the terminal demultiplexing and decoding the video file to obtain the initial video image. Specifically, the terminal demultiplexes the video file to obtain corresponding audio data and video data, and decodes the audio data to obtain the initial audio or the audio to be played, and decodes the video data to obtain the initial video image.

[0085] In one embodiment, the terminal parses the video file through a video player to obtain an initial video image. The terminal can specifically demultiplex the video file through a demultiplexing module in the video player to obtain video data and audio data, and decode the audio data through an audio decoding module in the video player, and decode the video data through a video decoding module in the video player to obtain an initial video image. The terminal performs image post-processing on the initial video image through a post-processing module in the video player to obtain a video image to be played. It can be understood that the demultiplexing module, audio decoding module, video decoding module and post-processing module are all functional modules in the video player, and specifically may refer to a computer program that implements a specific function.

[0086] In one embodiment, the image post-processing operation performed on the initial video image is determined by the video player that is playing the corresponding video file. The corresponding image post-processing strategy parameters are pre-configured for the video player. After the terminal parses the video file through the video player to obtain the initial video image, the image post-processing operation is performed on the initial video image according to the pre-configured image post-processing strategy parameters. Further, the video image to be evaluated is collected from the video image obtained by image post-processing, so that the image quality score of the video image after the real-time image post-processing performed by the video player can be evaluated. The image quality score obtained by the evaluation will be more in line with the user's subjective feelings, thereby improving the evaluation accuracy. It can be understood that if there is no need to perform image post-processing on the initial video image obtained by parsing the video file, the terminal directly determines the initial video image as the video image to be played, and plays the video image according to the playback sequence through the first thread, and evaluates the video image collected from the video image played according to the playback sequence through the second thread, thereby achieving the quality evaluation of the initial video image in the video file that has not been image post-processed.

[0087] It can be understood that the video quality assessment method provided in one or more embodiments of the present application can not only accurately assess the image quality of the video image that has been post-processed by the video player, that is, can accurately assess the video quality of the video file that has been post-processed by the video player, but also can assess the video quality of the user's local video file and the source video file that has not been post-processed. Moreover, through this video quality assessment method, there is no need to deploy independent quality assessment equipment, which can save assessment costs.

[0088] In the above embodiment, before playing the initial video image obtained by parsing the video file, the initial video image is subjected to image post-processing to adaptively adjust the played video image based on the playback requirements. Accordingly, the video image used to evaluate the image quality score is also the video image subjected to image post-processing, that is, the image quality score of the video image directly displayed to the user is evaluated, and the image quality score representing the subjective quality and true feeling of the video image directly perceived by the user can be evaluated, thereby improving the accuracy of the image quality score. Based on the image quality score with higher accuracy, a video quality evaluation result with higher accuracy can be obtained, thereby improving the evaluation accuracy.

[0089] In one embodiment, step S206 includes: using the second thread to evaluate the collected video image through a trained image quality evaluation model to obtain an image quality score.

[0090] Specifically, the terminal calls a trained image quality assessment model through the second thread, and evaluates the collected video image through the image quality assessment model to obtain a corresponding image quality score.

[0091] In one embodiment, the terminal is locally pre-configured with a trained image quality assessment model. After acquiring the video image to be evaluated, the terminal calls the locally pre-configured image quality assessment model through a second thread to predict the image quality score of the video image. The terminal may also dynamically download the trained image quality assessment model after acquiring the video image to be evaluated, and evaluate the video image through the image quality assessment model. It is understandable that the terminal may dynamically download the trained image quality assessment model after acquiring the first video image to be evaluated in the video file, and cache the downloaded image quality assessment model locally, so as to facilitate calling the cached image quality assessment model to evaluate the subsequently acquired video images.

[0092] In one embodiment, after the terminal obtains the video image to be played corresponding to the video file through the post-processing module in the video player, the video image to be played is rendered according to the playback timing through the rendering module in the video player, and the rendered video image is displayed through the display screen to achieve the playback of the video image, thereby achieving the playback of the corresponding video file. Before rendering and displaying the video image, the terminal collects the video image to be evaluated from the video image to be played through the rendering module. The terminal can specifically collect the video image to be evaluated from the video image to be rendered and displayed through the rendering module. The terminal can also collect the video image to be evaluated from the video image that has been rendered and is being displayed through the rendering module. It can be understood that the terminal collects the video image to be evaluated in real time through the rendering model, or collects the video image to be evaluated in a sampling manner.

[0093] Furthermore, the terminal pushes the collected video image to the inference engine through the rendering module, and calls the trained image quality assessment model through the inference engine to evaluate the image quality score of the video image. The terminal can push the collected video image to the evaluation scheduling module through the rendering module, and call the inference engine through the evaluation scheduling module using a second thread to evaluate the image quality score of the video image through the trained image quality assessment model in the inference engine. The terminal can specifically push the video image to the inference engine or the evaluation scheduling module through the rendering module in an asynchronous push manner, so as to asynchronously evaluate the image quality score of the video image through the inference engine, thereby realizing imperceptible evaluation of image quality.

[0094] In one embodiment, the initial video image obtained by parsing the video file by the terminal can be stored in YUV (a color encoding method) format. The terminal renders the initial video image in YUV format through a rendering module, and displays the rendered video image through a display screen.

[0095] In one embodiment, after the terminal obtains the image quality score of the video image through the inference engine evaluation, the image quality score is fed back to the evaluation scheduling module through the inference engine, so that the evaluation scheduling module can determine the video quality evaluation result of the corresponding video file according to the image quality score corresponding to the collected video image.

[0096] In one embodiment, the inference engine and the evaluation scheduling module may be functional modules within the video player, or may be independent modules that are independent of the video player and interact with the rendering module in the video player for data, which is not specifically limited here.

[0097] In one embodiment, the inference engine is an inference engine based on artificial intelligence technology. The inference engine uses a trained image quality assessment model to perform quality inference, i.e., assessment, on each captured video image to obtain a corresponding image quality score.

[0098] In the above embodiment, the image quality score of the video image is evaluated by using a trained image quality evaluation model, which can improve the efficiency and accuracy of the image quality evaluation.

[0099] In one embodiment, the training steps of the image quality assessment model include: obtaining a target image; manually scoring the target image to obtain a target image quality score; using the target image as an input feature and the corresponding target image quality score as a desired output feature for model training to obtain a trained image quality assessment model.

[0100] The target image may be an image captured from a target video file, or may be an image captured by other means, such as a single shot image.

[0101] Specifically, the training step of the image quality assessment model can be performed by a terminal or a server. The terminal or the server obtains various types of target images, and manually scores each of the obtained target images to obtain a corresponding target image quality score. The terminal or the server uses the obtained target image as an input feature and the corresponding target image quality score as the desired output feature for model training to obtain a trained image quality assessment model. The types of target images are divided according to the image content, such as animation, landscape, character, etc., so that the trained image quality assessment model can accurately evaluate various types of video images.

[0102] In one embodiment, the terminal or server obtains a target video file, and collects a video image from the obtained target video file, and uses the collected video image as the target image, which can also be understood as a target video image. The terminal or server can obtain various types of target video files so that various types of target images can be collected. The image quality assessment model is trained based on the target image collected from the target video file, so that when the video image collected from the video file is evaluated based on the trained image quality assessment model, the accuracy of the assessment can be improved.

[0103] In one embodiment, the terminal or server collects and displays the target image through a preconfigured data collection webpage, so that the user can manually score the displayed target image through the data collection webpage to obtain a target image quality score, which is obtained by the user manually scoring the intuitively perceived image, that is, a quality score that conforms to the user's subjective feelings. The terminal can configure a corresponding scoring control for each target image, so that the user can manually score based on the scoring control.

[0104] In one embodiment, after acquiring the target image and the corresponding target image quality score, the terminal or server performs model training according to the target image and the target image quality score through a preconfigured training engine.

[0105] In one embodiment, the machine learning operators involved in training the image quality assessment model include but are not limited to neural networks, decision trees, and support vector machines.

[0106] In the above embodiment, the target image is manually scored to obtain a target image quality score representing the user's subjective feeling about the target image, so that an image quality assessment model with higher assessment accuracy can be trained based on the target image quality score and the corresponding target image.

[0107] like Figure 3 As shown, in one embodiment, a video quality assessment method is provided, and the method specifically includes the following steps:

[0108] S302, parsing the video file to obtain an initial video image.

[0109] S304: performing image post-processing on the initial video image to obtain a video image to be played.

[0110] S306: Play the video images according to the play sequence through the first thread.

[0111] S308 , collecting the video images to be evaluated from the video images played according to the play sequence according to a preset sampling frequency.

[0112] S310: Use the second thread to evaluate the collected video image through the trained image quality evaluation model to obtain an image quality score.

[0113] S312, when the video file is finished playing, analyzing the image quality scores corresponding to the video images collected from the video file to obtain the highest score, the lowest score, the average score and the score variance.

[0114] S314, determining a video quality assessment result corresponding to the video file according to the highest score, the lowest score, the average score and the score variance.

[0115] S316, display the video quality assessment result.

[0116] In the above embodiment, the first thread plays the video image obtained by parsing the video file and performing image post-processing according to the playback sequence, and the second thread calls the trained image quality assessment model to evaluate the video image collected from the video image played according to the playback sequence, which can achieve imperceptible evaluation of the video image and improve the evaluation efficiency and accuracy, and then determine the video quality assessment result of the corresponding video file according to the image quality score of each video image, which can improve the evaluation efficiency of the video quality and obtain the video quality assessment result that meets the user's subjective feelings, thereby improving the accuracy of the video quality assessment. The sampling method can reduce the additional pressure brought by the image quality assessment to the terminal and the video player that plays the video file, and can avoid the problem of reducing the smoothness of the video playback due to the quality assessment, so as to ensure the accuracy of the assessment while reducing the evaluation cost.

[0117] Figure 4 The schematic diagram is a principle diagram of evaluating video quality through a video player in an embodiment. After startup, the video player obtains the video file to be played, demultiplexes the video file to obtain video data, decodes the video data to obtain an initial video image, performs image post-processing on the initial video image to obtain the video image to be played, renders and displays the video image according to the playback sequence through a first thread, and before rendering and displaying the video image, collects the video image played according to the playback sequence, that is, collects the video image to be evaluated, and calls the trained image quality evaluation model through a second thread to evaluate the image quality score of the video image to be evaluated. It can be understood that the video player determines the video quality evaluation result corresponding to the video file based on the image quality score of the video image collected from the video file.

[0118] Figure 5 FIG. 1 is a schematic diagram of an interface for evaluating video quality and displaying video quality evaluation results in one embodiment. Figure 5As shown, during the process of playing a video file, the terminal displays the video image currently being played in the video file in real time on the video playback interface 501 through the display screen, and at the same time, collects and evaluates the image quality score of the video image in the video file being played. When the video file is finished playing, the terminal determines the video quality evaluation result of the video file based on the image quality score of the video image collected from the video file, and displays the video quality evaluation result on the video end interface 502 through the display screen. Figure 5 The interface contents shown in the video playback interface 501 and the video end interface 502, as well as the position and form of the video quality assessment results displayed in the video end interface 502 are only examples and are not used for specific limitations. The video quality assessment results displayed in the video end interface 502 include the highest score, the lowest score, the average score and the score variance corresponding to the video file. The video quality assessment result can also be a video quality assessment score determined according to the four parameter values, or a video quality assessment score calculated in other ways based on the image quality scores of each collected video image. Other methods, such as weighted averaging or arithmetic averaging of each image quality score, are not specifically limited here.

[0119] Furthermore, the terminal plays the video file through the first thread, and evaluates the image quality score of the collected video image through the second thread, which can avoid the image quality evaluation operation from occupying the thread resources of the video player, thereby avoiding the image quality evaluation from affecting the normal operation of the video player and the normal playback of the video file. The terminal can record the image quality score of each video image to be evaluated locally, which can be recorded in memory or disk. When the video file is finished playing, the terminal can send the video quality evaluation result corresponding to the video file to the server, so that the server can perform statistics and analysis in the background. The terminal can also send the image quality score of the video image collected from the video file to the server.

[0120] In one embodiment, during the playback of the video file, the terminal can dynamically determine the video quality assessment result of the video file based on the image quality score of the video image collected from the video file, and can also display the dynamically determined video quality assessment result while displaying the currently playing video image through the video playback interface 501.

[0121] Figure 6 FIG. 1 is a schematic diagram of an interface for manually scoring a target image in one embodiment. Figure 6As shown, target images to be scored, such as target images 1-6, are collected and displayed through a data collection webpage, and a corresponding scoring control is displayed for each target image. When a user triggers a corresponding scoring trigger operation through the data collection webpage for the scoring control corresponding to the displayed target image, a quality score scoring box is displayed at a corresponding position of the target image currently being scored in the data collection webpage, so that the user can enter the target image quality score obtained by manually scoring the target image in the quality score scoring box, such as Figure 6 Enter the target image quality score at the position "XX" shown in the quality score scoring box. It is worth noting that Figure 6 The number of target images shown, the display mode of each target image and the display mode of the corresponding scoring control, and the display form and content of the quality score scoring box are only examples and are not intended to be specific. Figure 6 Taking the acquisition and display of the target image through a data acquisition webpage as an example, it is not intended to specifically limit the acquisition and display form of the target image. For example, the target image can be acquired and displayed through a data acquisition application.

[0122] It should be understood that although Figure 2-3 The steps in the flowchart are shown in sequence as indicated by the arrows, but these steps are not necessarily executed in the order indicated by the arrows. Unless otherwise specified in this document, there is no strict order restriction for the execution of these steps, and these steps can be executed in other orders. Moreover, Figure 2-3 At least part of the steps may include multiple steps or multiple stages. These steps or stages are not necessarily performed at the same time, but can be performed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed in turn or alternately with other steps or at least part of the steps or stages in other steps.

[0123] In one embodiment, Figure 7 As shown, a video quality assessment device 700 is provided. The device can adopt a software module or a hardware module, or a combination of the two to become a part of a computer device. The device specifically includes: a playback module 702, a collection module 704, an assessment module 706 and an assessment result determination module 708, wherein:

[0124] The playing module 702 is used to play the video file through the first thread;

[0125] The acquisition module 704 is used to acquire the video image in the video file being played;

[0126] An evaluation module 706, configured to evaluate the acquired video image through a second thread to obtain an image quality score;

[0127] The evaluation result determination module 708 is used to determine the video quality evaluation result corresponding to the video file according to the image quality score.

[0128] In one embodiment, the acquisition module 704 is further configured to acquire the video image to be evaluated from the video file being played according to a preset sampling frequency.

[0129] In one embodiment, the evaluation result determination module 708 is also used to determine the video quality evaluation result corresponding to the video file according to the image quality score corresponding to the video image collected from the video file when the video file is finished playing; and display the video quality evaluation result.

[0130] In one embodiment, the evaluation result determination module 708 is also used to analyze the image quality scores corresponding to each video image collected from the video file to obtain the highest score, the lowest score, the average score and the score variance; based on the highest score, the lowest score, the average score and the score variance, determine the video quality evaluation result corresponding to the video file.

[0131] In one embodiment, the playback module 702 is further used to parse the video file to obtain an initial video image; perform image post-processing on the initial video image to obtain a video image to be played; play the video image according to the playback sequence through the first thread; the acquisition module 704 is further used to acquire the video image to be evaluated from the video image played according to the playback sequence.

[0132] In one embodiment, the evaluation module 706 is used to use the second thread to evaluate the collected video image through the trained image quality evaluation model to obtain an image quality score.

[0133] In one embodiment, the video quality assessment device 700 further includes a training module;

[0134] The training module is used to obtain a target image; manually score the target image to obtain a target image quality score; use the target image as an input feature and the corresponding target image quality score as a desired output feature to perform model training to obtain a trained image quality assessment model.

[0135] For the specific definition of the video quality assessment device, please refer to the definition of the video quality assessment method above, which will not be repeated here. Each module in the above-mentioned video quality assessment device can be implemented in whole or in part by software, hardware and a combination thereof. The above-mentioned modules can be embedded in or independent of the processor in the computer device in the form of hardware, or can be stored in the memory of the computer device in the form of software, so that the processor can call and execute the operations corresponding to the above modules.

[0136] In one embodiment, a computer device is provided. The computer device may be a terminal, and its internal structure diagram may be as follows: Figure 8 As shown. The computer device includes a processor, a memory, a communication interface, a display screen and an input device connected through a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The communication interface of the computer device is used to communicate with an external terminal in a wired or wireless manner, and the wireless manner can be achieved through WIFI, an operator network, NFC (near field communication) or other technologies. When the computer program is executed by the processor, a video quality assessment method is implemented. The display screen of the computer device can be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device can be a touch layer covered on the display screen, or a key, trackball or touchpad set on the computer device housing, or an external keyboard, touchpad or mouse, etc.

[0137] Those skilled in the art will understand that Figure 8 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0138] In one embodiment, a computer device is further provided, including a memory and a processor, wherein a computer program is stored in the memory, and the processor implements the steps in the above method embodiments when executing the computer program.

[0139] In one embodiment, a computer-readable storage medium is provided, storing a computer program, which implements the steps in the above method embodiments when executed by a processor.

[0140] Those of ordinary skill in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory or optical memory, etc. Volatile memory can include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM).

[0141] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0142] The above-mentioned embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the invention patent. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the patent of the present application shall be subject to the attached claims.

Claims

1. A video quality assessment method, the method comprising: Parsing the video file through a video player to obtain an initial video image; The video player is pre-configured with corresponding image post-processing strategy parameters; Performing image post-processing on the initial video image according to the image post-processing strategy parameters by the video player to obtain a video image to be played; Play the video image according to the play sequence through the first thread; Capture video images from the video file being played; The second thread uses the trained image quality assessment model to evaluate the collected video images to obtain an image quality score; The training steps of the image quality assessment model include: Acquire a target image; manually score the target image to obtain a target image quality score; use the target image as an input feature and the corresponding target image quality score as a desired output feature to perform model training to obtain the trained image quality assessment model; Determine the video quality assessment result corresponding to the video file according to the image quality score; during the playback of the video file, dynamically determine the video quality assessment result of the video file according to the image quality score of the video image collected from the video file; and display the dynamically determined video quality assessment result while displaying the currently played video image through the video playback interface; The video quality evaluation result represents the subjective video quality of the video file played by the video player.

2. The method according to claim 1, characterized in that The step of collecting the video image in the video file being played includes: The video image to be evaluated is acquired from the video file being played according to a preset sampling frequency.

3. The method according to claim 1, characterized in that The determining, according to the image quality score, a video quality assessment result corresponding to the video file includes: When the video file is finished playing, determining a video quality assessment result corresponding to the video file according to an image quality score corresponding to a video image collected from the video file; The video quality assessment result is displayed.

4. The method according to claim 3, characterized in that The step of determining a video quality assessment result corresponding to the video file according to an image quality score corresponding to a video image collected from the video file comprises: Analyzing the image quality scores corresponding to each video image collected from the video file to obtain the highest score, the lowest score, the average score and the score variance; A video quality assessment result corresponding to the video file is determined according to the highest score, the lowest score, the average score and the score variance.

5. The method according to claim 1, characterized in that The step of collecting the video image in the video file being played includes: The video images to be evaluated are acquired from the video images played in a play time sequence.

6. A video quality assessment device, characterized in that: The device comprises: A playback module, configured to parse a video file through a video player to obtain an initial video image; the video player is pre-configured with corresponding image post-processing strategy parameters; the video player performs image post-processing on the initial video image according to the image post-processing strategy parameters to obtain a video image to be played; and the video image is played through a first thread according to a playback sequence; An acquisition module, used for acquiring video images from a video file being played; An evaluation module, used to evaluate the collected video images by using a trained image quality evaluation model using a second thread to obtain an image quality score; A training module is used to obtain a target image; manually score the target image to obtain a target image quality score; use the target image as an input feature and the corresponding target image quality score as a desired output feature to perform model training to obtain the trained image quality assessment model; An evaluation result determination module is used to determine the video quality evaluation result corresponding to the video file according to the image quality score; during the playback of the video file, dynamically determine the video quality evaluation result of the video file according to the image quality score of the video image collected from the video file; and display the dynamically determined video quality evaluation result while displaying the currently played video image through the video playback interface; The video quality evaluation result represents the subjective video quality of the video file played by the video player.

7. The device according to claim 6, characterized in that The acquisition module is also used to acquire the video image to be evaluated from the video file being played according to a preset sampling frequency.

8. The device according to claim 6, characterized in that The evaluation result determination module is also used to determine the video quality evaluation result corresponding to the video file according to the image quality score corresponding to the video image collected from the video file when the video file is finished playing; and display the video quality evaluation result.

9. The device according to claim 8, characterized in that The evaluation result determination module is also used to analyze the image quality scores corresponding to each video image collected from the video file to obtain the highest score, the lowest score, the average score and the score variance; based on the highest score, the lowest score, the average score and the score variance, determine the video quality evaluation result corresponding to the video file.

10. The device according to claim 6, characterized in that The acquisition module is further used to acquire the video images to be evaluated from the video images played according to the play time sequence.

11. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 5 are implemented.

12. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 5 are implemented.

Images