Video quality no-reference evaluation method, device and equipment and storage medium

Abstract

The invention relates to a video quality no-reference evaluation method, which comprises the following steps of: receiving and processing an H.264 packaged video code stream, and re-framing through jitter buffering to ensure that a video frame is completely received; analyzing the recombined code stream to obtain a macro block type and a quantization parameter value; determining an average quantization parameter value of non-skip type macro blocks in each frame of graph according to the macro block type and the quantization parameter value; and determining a video quality score according to the average quantization parameter value. The technical problem that an existing video quality no-reference evaluation method technology is difficult to apply in a high-real-time scene is solved, and the method is low in calculation complexity, does not depend on an original video and a complex deep learning model and can efficiently run in the high-real-time scene. The method has the advantages of being high in real-time performance, high in calculation efficiency, wide in application range and the like, and is particularly suitable for high-real-time-performance application scenes such as video calls and video conferences.

Description

Technical Field

[0001] The present invention relates to the technical field of video quality assessment, and in particular to a method, device, equipment and storage medium for video quality assessment without reference. Background Art

[0002] Video quality assessment (VQA) is an important algorithm in the field of computer audio and video. According to the relevant standards for video quality detection established by the Video Quality Experts Group (VQEG) of the International Telecommunication Union (ITU-T), it is divided into subjective evaluation and objective evaluation.

[0003] Subjective evaluation requires the participation of raters, who score the video through human visual senses (MOS) to obtain the evaluation of video quality. This method usually requires a laboratory environment and a large amount of human resources, and its evaluation process is time-consuming and costly.

[0004] Objective quality evaluation is divided into full reference (FR), partial reference (RR) and no reference (NR) methods. Full reference methods require the original video as a reference and compare it with the distorted video to measure the quality. Common full reference indicators include peak signal-to-noise ratio (PSNR), structural similarity (SSIM) and video multi-method assessment fusion (VMAF). While ensuring accuracy, these methods require the storage and transmission of the original video, which limits their practical application. Partial reference methods only require part of the original video's information or features for evaluation, such as comparison after feature extraction. Although it reduces the dependence on the original video, a certain degree of reference information is still required. No reference methods do not require the original video at all and only evaluate the quality by analyzing the distorted video itself. No reference methods are the most flexible in application, but they are also the most technically difficult field.

[0005] In the absence of a reference, there is only a distorted video but no original video. The traditional method is to use machine learning to find the features of the image, and then calculate the final distorted video quality based on a large amount of training data of subjective quality scores evaluated by human eyes. This method has a long process and is usually used for offline video measurement, evaluation and optimization that is not sensitive to latency. However, for high-real-time scenarios such as video calls or video conferences, the latency requirement is usually within 500ms, and the quality assessment method based on deep learning is not applicable. Summary of the invention

[0006] The main purpose of the present invention is to provide a video quality no-reference assessment method, device, equipment and storage medium, aiming to solve the technical problem that the existing video quality no-reference assessment method technology is difficult to apply in high real-time scenarios.

[0007] To achieve the above object, the present invention provides a method for evaluating video quality without reference, and the method for evaluating video quality without reference comprises the following steps:

[0008] Receiving a video code stream encapsulated in H.264, reframing the video code stream through a jitter buffer, and obtaining a recombined code stream;

[0009] Parsing the reorganized bitstream to obtain macroblock type and quantization parameter value;

[0010] An average quantization parameter value of macroblocks whose macroblock type is non-skip type in each frame of the reorganized bitstream is determined according to the macroblock type and the quantization parameter value, and a quality score of the video bitstream is determined according to the average quantization parameter value.

[0011] Optionally, the step of reframing the video code stream through a jitter buffer includes:

[0012] Packet loss, jitter and / or disorder of the video code stream caused by network congestion are processed to complete re-framing, thereby preventing incomplete reception of video frames of the video code stream.

[0013] Optionally, the step of parsing the reorganized bitstream to obtain a macroblock type and a quantization parameter value includes:

[0014] The slice and macroblock structure data are obtained by parsing the recombined code stream, including obtaining each macroblock type and the quantization parameter value.

[0015] Optionally, the step of obtaining the macroblock type and the quantization parameter value includes:

[0016] The macroblock type and the quantization parameter value are obtained, where the quantization parameter value ranges from an integer between 0 and 51.

[0017] Optionally, the step of determining, according to the macroblock type and the quantization parameter value, an average quantization parameter value of macroblocks whose macroblock type is non-skip type in each frame of the reconstructed bitstream comprises:

[0018] The macroblocks whose macroblock type in each frame of the reorganized bitstream are excluded, and the average quantization parameter value is determined according to the quantization parameter values ​​of the remaining non-skip type macroblocks in each frame of the reorganized bitstream.

[0019] Optionally, the step of determining the quality score of the video stream according to the average quantization parameter value includes:

[0020] An adaptive mechanism is established to adjust the quantization parameter value weight of the skip type macroblock according to different video contents in the reconstructed bitstream, determine a comprehensive average quantization parameter value according to the quantization parameter value of the skip type macroblock, the quantization parameter value weight and the average quantization parameter value, and determine the quality score of the video bitstream according to the comprehensive average quantization parameter value.

[0021] Optionally, after the step of determining the quality score of the video stream according to the average quantization parameter value, the method further includes:

[0022] A user feedback mechanism is established, user subjective scores are obtained according to the user feedback mechanism, and a comprehensive score is determined according to the user subjective scores and the quality score.

[0023] Furthermore, in order to achieve the above-mentioned object, the present invention also provides a video quality non-reference evaluation method device, and the video quality non-reference evaluation method device comprises:

[0024] A reassembly module is used to receive a video code stream encapsulated in H.264, reframe the video code stream through a jitter buffer, and obtain a reassembled code stream;

[0025] A parsing module, used for parsing the recombined bitstream to obtain a macroblock type and a quantization parameter value;

[0026] An evaluation module is used to determine an average quantization parameter value of a macroblock whose macroblock type is a non-skip type in each frame of the reconstructed bitstream according to the macroblock type and the quantization parameter value, and determine a quality score of the video bitstream according to the average quantization parameter value.

[0027] Furthermore, to achieve the above-mentioned purpose, the present invention also provides a video quality no-reference assessment method device, the video quality no-reference assessment method device comprising a memory, a processor, and a video quality no-reference assessment method program stored in the memory and executable on the processor, the video quality no-reference assessment method program, when executed by the processor, implements the steps of the video quality no-reference assessment method as described above.

[0028] Furthermore, to achieve the above-mentioned purpose, the present invention also provides a storage medium, on which a video quality no-reference assessment method program is stored, and when the video quality no-reference assessment method program is executed by a processor, the steps of the video quality no-reference assessment method as described above are implemented.

[0029] The present invention relates to a method for evaluating video quality without reference, which receives and processes a video code stream encapsulated in H.264, reframes the video frame through a jitter buffer, and ensures that the video frame is received completely; parses the reorganized code stream to obtain the macroblock type and quantization parameter value; determines the average quantization parameter value of non-skip type macroblocks in each frame of graphics according to the macroblock type and the quantization parameter value; and then determines the video quality score according to the average quantization parameter value. The technical problem that the existing video quality without reference evaluation method technology is difficult to apply in high real-time scenarios is solved. The method has low computational complexity, does not rely on the original video and complex deep learning models, and can run efficiently in high real-time scenarios. The present invention has the advantages of strong real-time performance, high computational efficiency, and a wide range of applications, and is particularly suitable for high real-time application scenarios such as video calls and video conferences. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] Figure 1 It is a structural schematic diagram of the device hardware operating environment involved in the device embodiment of the video quality non-reference evaluation method of the present invention;

[0031] Figure 2 A schematic diagram of a flow chart of a first embodiment of a method for assessing video quality without reference according to the present invention;

[0032] Figure 3 It is a functional module diagram of a preferred embodiment of a method and device for assessing video quality without reference according to the present invention.

[0033] The realization of the purpose, functional features and advantages of the present invention will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0034] It should be understood that the specific embodiments described herein are only used to explain the present invention, and are not used to limit the present invention.

[0035] It should be noted that video quality assessment (VQA) is an important algorithm in the field of computer audio and video. According to the relevant standards for video quality detection established by the Video Quality Experts Group (VQEG) of the International Telecommunication Union (ITU-T), it is divided into subjective evaluation and objective evaluation.

[0036] Subjective evaluation requires the participation of raters, who score the video through human visual senses (MOS) to obtain the evaluation of video quality. This method usually requires a laboratory environment and a large amount of human resources, and its evaluation process is time-consuming and costly.

[0037] Objective quality evaluation is divided into full reference (FR), partial reference (RR) and no reference (NR) methods. Full reference methods require the original video as a reference and compare it with the distorted video to measure the quality. Common full reference indicators include peak signal-to-noise ratio (PSNR), structural similarity (SSIM) and video multi-method assessment fusion (VMAF). While ensuring accuracy, these methods require the storage and transmission of the original video, which limits their practical application. Partial reference methods only require part of the original video's information or features for evaluation, such as comparison after feature extraction. Although it reduces the dependence on the original video, a certain degree of reference information is still required. No reference methods do not require the original video at all and only evaluate the quality by analyzing the distorted video itself. No reference methods are the most flexible in application, but they are also the most technically difficult field.

[0038] In the absence of a reference, there is only a distorted video but no original video. The traditional method is to use machine learning to find the features of the image, and then calculate the final distorted video quality based on a large amount of training data of subjective quality scores evaluated by human eyes. This method has a long process and is usually used for offline video measurement, evaluation and optimization that is not sensitive to latency. However, for high-real-time scenarios such as video calls or video conferences, the latency requirement is usually within 500ms, and the quality assessment method based on deep learning is not applicable.

[0039] Based on the above defects, the present invention provides a video quality non-reference evaluation method and device, referring to Figure 1 , Figure 1 It is a structural schematic diagram of the device hardware operating environment involved in the device embodiment of the video quality non-reference assessment method of the present invention.

[0040] like Figure 1 As shown, the video quality non-reference assessment method device may include: a processor 1001, such as a CPU, a communication bus 1002, a user interface 1003, a network interface 1004, and a memory 1005. Among them, the communication bus 1002 is used to realize the connection and communication between these components. The user interface 1003 may include a display screen (Display), an input unit such as a keyboard (Keyboard), and the optional user interface 1003 may also include a standard wired interface and a wireless interface. The network interface 1004 may optionally include a standard wired interface and a wireless interface (such as a WI-FI interface). The memory 1005 may be a high-speed RAM memory, or a stable memory (non-volatile memory), such as a disk memory. The memory 1005 may also be a storage device independent of the aforementioned processor 1001.

[0041] Those skilled in the art will understand that Figure 1The hardware structure of the video quality no-reference assessment method device shown in the figure does not constitute a limitation of the video quality no-reference assessment method device, and may include more or less components than shown in the figure, or combine certain components, or arrange the components differently.

[0042] like Figure 1 As shown, the memory 1005 as a storage medium may include an operating system, a network communication module, a user interface module, and a video quality no-reference evaluation method program. The operating system is a program for managing and controlling the video quality no-reference evaluation method device and software resources, and supports the operation of the network communication module, the user interface module, the video quality no-reference evaluation method program, and other programs or software; the network communication module is used to manage and control the network interface 1004; and the user interface module is used to manage and control the user interface 1003.

[0043] exist Figure 1 In the hardware structure of the video quality non-reference evaluation method device shown, the network interface 1004 is mainly used to connect to the background server and perform data communication with the background server; the user interface 1003 is mainly used to connect to the client and perform data communication with the client; the processor 1001 can call the video quality non-reference evaluation method program stored in the memory 1005, and perform the following operations:

[0044] Receiving a video code stream encapsulated in H.264, reframing the video code stream through a jitter buffer, and obtaining a recombined code stream;

[0045] Parsing the reorganized bitstream to obtain macroblock type and quantization parameter value;

[0046] An average quantization parameter value of macroblocks whose macroblock type is non-skip type in each frame of the reorganized bitstream is determined according to the macroblock type and the quantization parameter value, and a quality score of the video bitstream is determined according to the average quantization parameter value.

[0047] Furthermore, the step of reframing the video code stream through a jitter buffer includes:

[0048] Packet loss, jitter and / or disorder of the video code stream caused by network congestion are processed to complete re-framing, thereby preventing incomplete reception of video frames of the video code stream.

[0049] Furthermore, the step of parsing the reorganized bitstream to obtain the macroblock type and quantization parameter value includes:

[0050] The slice and macroblock structure data are obtained by parsing the recombined code stream, including obtaining each macroblock type and the quantization parameter value.

[0051] Furthermore, the step of obtaining the macroblock type and the quantization parameter value includes:

[0052] The macroblock type and the quantization parameter value are obtained, where the quantization parameter value ranges from an integer between 0 and 51.

[0053] Further, the step of determining, according to the macroblock type and the quantization parameter value, an average quantization parameter value of macroblocks of a non-skip type in each frame of the reconstructed bitstream comprises:

[0054] The macroblocks whose macroblock type in each frame of the reorganized bitstream are excluded, and the average quantization parameter value is determined according to the quantization parameter values ​​of the remaining non-skip type macroblocks in each frame of the reorganized bitstream.

[0055] Furthermore, the step of determining the quality score of the video stream according to the average quantization parameter value includes:

[0056] An adaptive mechanism is established to adjust the quantization parameter value weight of the skip type macroblock according to different video contents in the reconstructed bitstream, determine a comprehensive average quantization parameter value according to the quantization parameter value of the skip type macroblock, the quantization parameter value weight and the average quantization parameter value, and determine the quality score of the video bitstream according to the comprehensive average quantization parameter value.

[0057] Furthermore, after the step of determining the quality score of the video stream according to the average quantization parameter value, the method further includes:

[0058] A user feedback mechanism is established, user subjective scores are obtained according to the user feedback mechanism, and a comprehensive score is determined according to the user subjective scores and the quality score.

[0059] The specific implementation of the video quality no-reference assessment method device of the present invention is basically the same as the various embodiments of the video quality no-reference assessment method described below, and will not be repeated here.

[0060] The present invention also provides a video quality no-reference assessment method based on the video quality no-reference assessment method device.

[0061] Reference Figure 2 , Figure 2 FIG. 4 is a flow chart of a first embodiment of a method for assessing video quality without reference according to the present invention.

[0062] The embodiment of the present invention provides an embodiment of a method for assessing video quality without reference. It should be noted that although a logical sequence is shown in the flow chart, in some cases, the steps shown or described may be performed in an order different from that shown here.

[0063] In various embodiments of the video quality no-reference assessment method, the execution subject is a video quality no-reference assessment method device.

[0064] The video quality no-reference assessment method is applied to a video quality no-reference assessment method device, and the video quality no-reference assessment method includes:

[0065] Step S10, receiving a video code stream encapsulated in H.264, reframing the video code stream through a jitter buffer, and obtaining a recombined code stream;

[0066] Video quality assessment (VQA) is an important algorithm in the field of computer audio and video. According to the relevant standards for video quality detection established by the Video Quality Experts Group (VQEG) of the International Telecommunication Union (ITU-T), it is divided into subjective evaluation and objective evaluation.

[0067] Subjective evaluation requires the participation of raters, who score the video through human visual senses (MOS) to obtain the evaluation of video quality. This method usually requires a laboratory environment and a large amount of human resources, and its evaluation process is time-consuming and costly.

[0068] Objective quality evaluation is divided into full reference (FR), partial reference (RR) and no reference (NR) methods. Full reference methods require the original video as a reference and compare it with the distorted video to measure the quality. Common full reference indicators include peak signal-to-noise ratio (PSNR), structural similarity (SSIM) and video multi-method assessment fusion (VMAF). While ensuring accuracy, these methods require the storage and transmission of the original video, which limits their practical application. Partial reference methods only require part of the original video's information or features for evaluation, such as comparison after feature extraction. Although it reduces the dependence on the original video, a certain degree of reference information is still required. No reference methods do not require the original video at all and only evaluate the quality by analyzing the distorted video itself. No reference methods are the most flexible in application, but they are also the most technically difficult field.

[0069] In the absence of a reference, there is only a distorted video but no original video. The traditional method is to use machine learning to find the features of the image, and then calculate the final distorted video quality based on a large amount of training data of subjective quality scores evaluated by human eyes. This method has a long process and is usually used for offline video measurement, evaluation and optimization that is not sensitive to latency. However, for high-real-time scenarios such as video calls or video conferences, the latency requirement is usually within 500ms, and the quality assessment method based on deep learning is not applicable.

[0070] In order to solve the technical problem that the existing video quality no-reference evaluation method technology is difficult to apply in high real-time scenarios, a video quality no-reference evaluation method is proposed in an embodiment of the present invention, which aims to receive and process the H.264 encapsulated video stream, reframe through the jitter buffer to ensure the complete reception of the video frame; parse the reorganized stream to obtain the macroblock type and quantization parameter value; determine the average quantization parameter value of the non-skip type macroblock in each frame of the graphic according to the macroblock type and the quantization parameter value; and then determine the video quality score according to the average quantization parameter value. The technical problem that the existing video quality no-reference evaluation method technology is difficult to apply in high real-time scenarios is solved. The method has low computational complexity, does not rely on the original video and complex deep learning models, and can run efficiently in high real-time scenarios. The invention has the advantages of strong real-time performance, high computational efficiency, and a wide range of applications, and is particularly suitable for high real-time application scenarios such as video calls and video conferences.

[0071] The video quality no-reference assessment method in this embodiment is applicable to a device having a video quality no-reference assessment method function.

[0072] In this embodiment, a video stream encapsulated in H.264RTP / UDP is first received through the network. In this embodiment, it is assumed that the video stream is transmitted through a standard network protocol, and the receiving end uses a common network video receiver. The receiver can be a computer, a smart phone, a tablet computer, or other devices with video receiving function.

[0073] For example, a network interface is used to receive RTP / UDP encapsulated video packets, a sequence number and a timestamp of each RTP packet are checked to ensure that the received packets are complete and in order, and then the received RTP packets are placed in a jitter buffer.

[0074] In order to deal with jitter, packet loss and disorder problems in the network environment, the receiving end first performs jitter buffering on the video stream. The main function of the jitter buffer is to reframe and ensure the integrity of the video frame. The specific steps are as follows:

[0075] Sorting and Buffering: The jitter buffer sorts the RTP packets according to their sequence numbers and rearranges the out-of-order packets. At the same time, the buffer will buffer a certain number of packets to handle jitter and packet loss issues.

[0076] Reassembly of video frames: Once complete frame data is received, the jitter buffer reassembles the data into H.264 frames. Assuming that network jitter causes some packets to arrive late, the jitter buffer will wait for a preset time (e.g. 100 milliseconds) to ensure that as many frames as possible are not lost.

[0077] Packet loss handling: For lost packets, the jitter buffer can interpolate or repeat the previous frame to reduce visual discontinuities.

[0078] Step S20, parsing the reorganized bitstream to obtain a macroblock type and a quantization parameter value;

[0079] In this embodiment, the reorganized H.264 code stream is parsed to extract the macroblock structure data in the video frame. The specific steps are as follows:

[0080] Frame header parsing: First, parse the video frame header information to identify the starting position and type of each frame (such as I frame, P frame or B frame). The frame header information is usually contained in the NAL unit of the H.264 video bitstream. By parsing the NAL unit header, the starting position and type of the frame can be identified.

[0081] Slice parsing: further parse the slices in each frame. Slices are the basic unit in H.264 encoding and contain multiple macroblocks. The specific steps are as follows:

[0082] Parse the slice header information, including slice type, starting position, reference frame index, etc.; obtain the number and position of macroblocks in the slice; determine the decoding order and method of the slice according to the slice header information.

[0083] Macroblock parsing: Extract macroblocks one by one from the slices, parse the type of each macroblock (such as Intra, Inter, and Skip, etc.) and the corresponding quantization parameter value, i.e., the quantization parameter qp value (quantization parameter). The specific steps are as follows:

[0084] Macroblock type analysis: Macroblock types are divided into three categories: Intra, Inter, and Skip. Intra macroblocks represent intra-frame prediction, using the information of the current frame for prediction; Inter macroblocks represent inter-frame prediction, using the information of the reference frame for prediction; Skip macroblocks represent residual-free coding, that is, the prediction error is zero. Determine the macroblock type by parsing the macroblock header information. For Intra macroblocks, parse the prediction mode; for Inter macroblocks, parse the motion vector and reference frame index; for Skip macroblocks, no further parsing is required.

[0085] Quantization parameter parsing: The quantization parameter qp value is used to control the quantization level of the macroblock. The higher the value, the higher the compression rate and the greater the distortion. Parse the quantization parameter qp value in the macroblock header, ranging from 0 to 51. The parsing method of the quantization parameter varies according to the macroblock type. For Intra macroblocks and Inter macroblocks, the qp value is parsed; for Skip macroblocks, the qp value is zero. The quantization parameter qp value may change within the slice, so the current qp value needs to be continuously updated during the parsing process.

[0086] According to the ITU-T H.264 standard, the quantization parameter qp ranges from 0 to 51 integers, which actually corresponds to the H.264 encoding quantization step qstep (0.625-224), indicating the degree of quantization refinement. The higher the value, the higher the video compression rate and the greater the distortion. Therefore, the higher the qp value, the lower the corresponding video quality. In frames with an actual average qp value of more than 30, the human eye can basically feel that the image is obviously blurred.

[0087] In addition, during the parsing process, if an erroneous NAL unit or data corruption is encountered, the system should be able to identify and skip these errors and continue to parse subsequent data. The stability and accuracy of the parsing process are ensured through redundant information and error detection mechanisms. The parsed macroblock type and quantization parameter qp value are stored in a data structure for subsequent calculations. For example, an array or linked list structure can be used to store the macroblock information of each frame. In actual implementation, in order to improve the parsing efficiency, parallel processing techniques such as multithreading or GPU acceleration can be used. By parsing slices and macroblocks in parallel, the parsing speed is significantly improved to meet the requirements of real-time processing.

[0088] Step S30, determining an average quantization parameter value of non-skip macroblocks in each frame of the reconstructed bitstream according to the macroblock type and the quantization parameter value, and determining a quality score of the video bitstream according to the average quantization parameter value.

[0089] In this embodiment, after the macroblock type and the quantization parameter value are parsed, further processing is performed as follows:

[0090] Count non-skip macroblocks: Count the number of all non-skip macroblocks in each frame. For example, a frame contains 100 macroblocks, of which 80 are non-skip macroblocks.

[0091] Calculate the sum of the quantization parameter QP values: Calculate the sum of the quantization parameter QP values ​​of these non-skip macroblocks. For example, assume that the sum of the QP values ​​of these 80 macroblocks is 3200.

[0092] Calculate the average quantization parameter qp value: divide the sum of the qp values ​​by the number of non-skip macroblocks to obtain the average quantization parameter value (avg_qp) of the frame. For example, the average quantization parameter value = 3200 / 80 = 40.

[0093] The video quality score of the frame is determined based on the average quantization parameter value. Generally, the higher the quantization parameter value, the worse the video quality. The average quantization parameter value can be mapped to the video quality score by the following method:

[0094] Quality score mapping formula: Set a quality score mapping formula, for example, quality score = 100-avg_qp. In this way, when the avg_qp value is low, the quality score is high, and vice versa.

[0095] Quality score calculation: In the above example, the average quantization parameter value is 40, so the video quality score = 100-40 = 60.

[0096] Of course, the average quantization parameter value may also be directly used as the video quality score, that is, the higher the average quantization parameter value, that is, the higher the video quality score, the lower the actual corresponding video quality.

[0097] Finally, through the above steps, the video quality of each frame can be evaluated without reference and a quality score can be output, thereby achieving real-time quality evaluation of H.264 encoded video.

[0098] In the actual verification, H.264 encoded samples and original videos with several different bit rates were used to calculate the video quality curves using the PSNR and SSIM full-reference evaluation models, and compared with the video quality curves calculated using the scheme without reference evaluation. The results show that the evaluation results of this algorithm are consistent with the curve trends of the full-reference evaluation results, with no obvious difference, and are suitable for evaluating video quality when there is no original video on the H.264 decoding side.

[0099] The present invention receives and processes the video code stream encapsulated in H.264, reframes through a jitter buffer to ensure that the video frame is received completely; parses the reorganized code stream to obtain the macroblock type and quantization parameter value; determines the average quantization parameter value of non-skip type macroblocks in each frame of graphics according to the macroblock type and quantization parameter value; and then determines the video quality score according to the average quantization parameter value. The technical problem that the existing video quality reference-free evaluation method technology is difficult to apply in high real-time scenarios is solved. The method has low computational complexity, does not rely on the original video and complex deep learning models, and can run efficiently in high real-time scenarios. The present invention has the advantages of strong real-time performance, high computational efficiency, and a wide range of applications, and is particularly suitable for high real-time application scenarios such as video calls and video conferences.

[0100] Furthermore, a second embodiment of the video quality non-reference assessment method of the present invention is proposed. In the above step S10, the step of reframing the video code stream through a jitter buffer includes:

[0101] Step a1, processing the packet loss, jitter and / or disorder of the video code stream caused by network congestion, completing re-framing, and preventing incomplete reception of video frames of the video code stream.

[0102] In this embodiment, how to handle packet loss, jitter and disorder caused by network congestion is described in detail to ensure complete reception and re-framing of the video code stream, which may include the following steps:

[0103] Receive RTP packets: Use the network interface to receive RTP / UDP encapsulated video packets. Each RTP packet contains a video data segment and comes with a sequence number and timestamp.

[0104] Initialize the jitter buffer: In order to deal with jitter and packet loss problems in network transmission, a jitter buffer is first initialized. The jitter buffer is used to temporarily store and sort the received RTP packets, and the cache time window of the jitter buffer is set, for example, to 100 milliseconds to ensure that most video packets can arrive within this time window.

[0105] Handling out-of-order: The jitter buffer sorts the received packets according to the sequence number of the RTP packet. By comparing the sequence numbers, the RTP packets that arrive out of order can be identified and arranged; if the sequence number of a packet is missing, the jitter buffer will reserve the position and wait for the arrival of the missing packet. Assume that some packets may arrive late in the network environment.

[0106] Handling packet loss: If the RTP packet corresponding to a certain sequence number is not received within the set cache time window, the packet is considered lost; for the lost packet, interpolation or repetition of the previous valid packet can be used to minimize visual discontinuity in video playback. For example, if the packet with sequence number 5 is lost, the data of the packet with sequence number 4 can be used for interpolation or repetition.

[0107] Handling jitter: Due to network congestion and changes in transmission paths, the received RTP packets may have time jitter. The jitter buffer uses timestamp information to ensure that the packets are cached and processed within the preset time window to eliminate time jitter; the timestamp of each RTP packet is used to determine its correct playback time. By adjusting the buffer time, the jitter of the packet arrival time can be smoothed.

[0108] Reframing: After the RTP packets in the jitter buffer are sorted according to the sequence number and timestamp, reframing begins. According to the frame structure of H.264 encoding, the sorted packet data is reassembled into a complete video frame; ensuring that the NAL units (network abstraction layer units) of each video frame are arranged in the correct order to form a complete frame data.

[0109] Output reassembled bitstream: Output the reframed video data for subsequent video analysis and quality assessment. The reframed bitstream should restore the original video data as completely as possible to ensure the accuracy of subsequent processing; the output reassembled bitstream should be as consistent as possible with the original video bitstream to ensure the smoothness and continuity of video playback.

[0110] Through the above steps, this embodiment handles packet loss, jitter and disorder caused by network congestion or other reasons, completes reframing, and ensures the integrity and continuity of the video code stream. In practical applications, this method can significantly improve the reliability and quality of video transmission and is suitable for high real-time application scenarios such as video calls and video conferences.

[0111] Furthermore, a third embodiment of the video quality non-reference assessment method of the present invention is proposed, wherein the above step S20 comprises:

[0112] Step b1, obtaining slice and macroblock structure data by parsing the reorganized bitstream, including obtaining each macroblock type and the quantization parameter value.

[0113] In this embodiment, how to parse the reorganized H.264 code stream, obtain the slice and macroblock structure data, and extract the type and quantization parameter value of each macroblock is described in detail. The specific steps may include:

[0114] Initialize the parser: Initialize the H.264 code stream parser to parse the reassembled video code stream. The parser can be a software module or hardware device that has the ability to parse the H.264 code stream.

[0115] Frame header parsing: Parse the video frame header information to identify the starting position and type of each frame (such as I frame, P frame or B frame). The frame header information is usually contained in the NAL unit of the H.264 video stream. The NAL unit header contains information about the frame type and priority. By parsing the NAL header, the starting position and type of the frame can be identified.

[0116] Slice parsing: further parse the slices in each frame. H.264 video frames are usually divided into multiple slices, each of which contains several macroblocks. Parse the slice header information, including slice type, starting position, reference frame index, etc. The slice header information can determine the properties and decoding order of the slice. Get the number and position of macroblocks in the slice, and determine the decoding order and method of the slice based on the slice header information.

[0117] Macroblock parsing: Extract macroblocks from slices one by one, and parse the type and quantization parameter qp value of each macroblock. Macroblock is the basic coding unit in H.264 encoding, and each macroblock contains 16x16 pixel image data.

[0118] Macroblock type analysis: Macroblock types are divided into three categories: Intra, Inter, and Skip. Intra macroblock: indicates intra-frame prediction, using the information of the current frame for prediction. Parse prediction modes, such as 4x4 block mode or 16x16 block mode. Inter macroblock: indicates inter-frame prediction, using the information of the reference frame for prediction. Parse motion vectors and reference frame indexes. Skip macroblock: indicates no residual coding, that is, the prediction error is zero, and no further analysis is required.

[0119] Quantization parameter parsing: The quantization parameter qp value is used to control the quantization level of the macroblock. The higher the value, the higher the compression rate and the greater the distortion. Parse the quantization parameter qp value in the macroblock header, ranging from 0 to 51. The parsing method of the quantization parameter varies according to the macroblock type. For Intra macroblocks and Inter macroblocks, the qp value is parsed; for Skip macroblocks, the qp value is zero. The quantization parameter qp value may change within the slice, so the current qp value needs to be continuously updated during the parsing process.

[0120] Error handling: During the parsing process, if an erroneous NAL unit or data corruption is encountered, the system should be able to identify and skip these errors and continue parsing the subsequent data. Use redundant information and error detection mechanisms, such as checksum or cyclic redundancy check (CRC), to ensure the stability and accuracy of the parsing process.

[0121] Data storage: Store the parsed macroblock type and quantization parameter qp value in a data structure for subsequent calculations. For example, an array or linked list structure can be used to store the macroblock information of each frame. The data structure should be designed reasonably to efficiently store and access macroblock types and quantization parameter values ​​to support subsequent quality assessment calculations.

[0122] Performance optimization: In actual implementation, in order to improve parsing efficiency, parallel processing techniques such as multithreading or GPU acceleration can be used. By parsing slices and macroblocks in parallel, the parsing speed can be significantly improved to meet the requirements of real-time processing. The parsing tasks of each slice or macroblock can be assigned to different processing units to achieve parallel parsing.

[0123] This embodiment parses the reorganized bitstream to obtain the type and quantization parameter value of each macroblock. The parsing process should ensure the integrity and accuracy of the data and provide a reliable data basis for subsequent video quality evaluation. This method is suitable for high-real-time video processing applications, such as video calls and video conferences, and can effectively evaluate video quality and ensure user experience.

[0124] Furthermore, a fourth embodiment of the video quality non-reference assessment method of the present invention is proposed. In the above step S20, the step of obtaining the macroblock type and the quantization parameter value includes:

[0125] Step c1, obtaining the macroblock type and the quantization parameter value, wherein the quantization parameter value ranges from an integer between 0 and 51.

[0126] In this embodiment, the quantization parameter qp value is used to control the quantization level of the macroblock. The higher the value, the higher the compression rate and the greater the distortion. The quantization parameter qp value in the macroblock header is parsed, ranging from 0 to 51. The parsing method of the quantization parameter varies according to the macroblock type. For Intra macroblocks and Inter macroblocks, the qp value is parsed; for Skip macroblocks, the qp value is zero. The quantization parameter qp value may change within the slice, so it is necessary to continuously update the current qp value during the parsing process.

[0127] According to the ITU-T H.264 standard, the quantization parameter qp ranges from 0 to 51 integers, which actually corresponds to the H.264 encoding quantization step qstep (0.625-224), indicating the degree of quantization refinement. The higher the value, the higher the video compression rate and the greater the distortion. Therefore, the higher the qp value, the lower the corresponding video quality. In frames with an actual average qp value of more than 30, the human eye can basically feel that the image is obviously blurred.

[0128] This embodiment obtains the type and quantization parameter value of each macroblock by parsing the reorganized bitstream. The parsing process should ensure the integrity and accuracy of the data and provide a reliable data basis for subsequent video quality evaluation.

[0129] Furthermore, a fifth embodiment of the method for assessing video quality without reference according to the present invention is proposed. In the above step S30, the step of determining the average quantization parameter value of the macroblocks of the non-skip type in each frame of the reconstructed bitstream according to the macroblock type and the quantization parameter value comprises:

[0130] Step d1, excluding the skip-type macroblocks in each frame of the reorganized bitstream, and determining the average quantization parameter value according to the quantization parameter values ​​of the remaining non-skip-type macroblocks in each frame of the reorganized bitstream.

[0131] In this embodiment, since the pixel residual of intra-frame prediction and the motion residual of inter-frame prediction of skip-type macroblocks are extremely low and almost do not occupy H.264 code stream data, this embodiment describes in detail how to exclude skip-type macroblocks and determine the average quantization parameter value of each frame of video based on the quantization parameter values ​​of the remaining non-skip-type macroblocks. The specific steps may include:

[0132] Parse the reorganized bitstream: Parse the reorganized H.264 bitstream according to the above steps to obtain the type and quantization parameter value of each macroblock. The specific parsing steps include frame header parsing, slice parsing and macroblock parsing.

[0133] Macroblock type screening: During the parsing process, the type and quantization parameter value of each macroblock are recorded. For each macroblock, first determine whether its type is skip type. Skip macroblocks indicate no residual coding, that is, the prediction error is zero.

[0134] Exclude skip type macroblocks: In each frame of video data, exclude all skip type macroblocks. Since these macroblocks have no actual residual data, they will not affect the quantization parameter value calculation. Only non-skip type macroblocks (including Intra and Inter macroblocks) are retained for subsequent quantization parameter value calculation.

[0135] Count non-skip macroblocks: Count the number of all non-skip macroblocks in each frame. For example, if a frame contains 100 macroblocks, 80 of which are non-skip macroblocks. Add up the quantization parameter values ​​of these non-skip macroblocks and calculate their sum.

[0136] Calculate the sum of the quantization parameter QP values: Calculate the sum of the quantization parameter QP values ​​of these non-skip macroblocks. For example, assume that the sum of the QP values ​​of these 80 macroblocks is 3200.

[0137] Calculate the average quantization parameter qp value: divide the sum of the qp values ​​by the number of non-skip macroblocks to obtain the average quantization parameter value (avg_qp) of the frame. For example, the average quantization parameter value = 3200 / 80 = 40.

[0138] For example, suppose a frame of video contains the following macroblock data:

[0139] Macroblock 1 (Intra): qp value = 30

[0140] Macroblock 2 (Inter): qp value = 25

[0141] Macroblock 3 (Skip): qp value = 0 (skip)

[0142] Macroblock 4 (Intra): qp value = 35

[0143] Macroblock 5 (Inter): qp value = 40

[0144] Macroblock 6 (Skip): qp value = 0 (skip)

[0145] Macroblock 7 (Intra): qp value = 32

[0146] During the parsing process, skip type macroblocks (macroblock 3 and macroblock 6) are first excluded, and only non-skip type macroblocks are retained:

[0147] Macroblock 1 (Intra): qp value = 30

[0148] Macroblock 2 (Inter): qp value = 25

[0149] Macroblock 4 (Intra): qp value = 35

[0150] Macroblock 5 (Inter): qp value = 40

[0151] Macroblock 7 (Intra): qp value = 32

[0152] Calculate the sum of the qp values ​​of these non-skip type macroblocks:

[0153] Total qp = 30 + 25 + 35 + 40 + 32 = 162.

[0154] Calculate the number of non-skip type macroblocks:

[0155] Number of non-skip macroblocks = 5

[0156] Calculate the average quantization parameter value (avg_qp):

[0157] Average quantization parameter value = 162 / 5 = 32.4.

[0158] Through the above steps, this embodiment excludes skip type macroblocks and calculates the average quantization parameter value of non-skip type macroblocks. The average value will be used for subsequent video quality evaluation. This method ensures the accuracy and integrity of data during the quality evaluation process and is particularly suitable for high real-time video processing applications such as video calls and video conferences.

[0159] Furthermore, a sixth embodiment of the method for evaluating video quality without reference according to the present invention is proposed. In the above step S30, the step of determining the quality score of the video stream according to the average quantization parameter value includes:

[0160] Step e1, establishing an adaptive mechanism, adjusting the quantization parameter value weight of the skip type macroblock according to different video contents in the reconstructed bitstream, determining a comprehensive average quantization parameter value according to the quantization parameter value of the skip type macroblock, the quantization parameter value weight and the average quantization parameter value, and determining the quality score of the video bitstream according to the comprehensive average quantization parameter value.

[0161] In this embodiment, it is described in detail how to establish an adaptive mechanism, adjust the quantization parameter value weight of the skip type macroblock according to the different video contents in the reconstructed bitstream, calculate the comprehensive average quantization parameter value in combination with the average quantization parameter value of the non-skip type macroblock, and use it to determine the quality score of the video bitstream. The specific steps may include:

[0162] Initialize the adaptive mechanism: Initialize the adaptive mechanism module, which is used to dynamically adjust the quantization parameter value weight of the skip type macroblock according to the video content. The adaptive mechanism can adopt a rule-based adjustment method or optimize the weight through a machine learning model.

[0163] Analyze video content: Analyze the content of each frame of video to extract the complexity and motion characteristics of the video scene. For example, you can calculate the edge strength, texture complexity, and motion vector features within the frame. According to the video content characteristics, classify different scene types (such as static scenes, high-motion scenes, etc.).

[0164] Dynamically adjust the quantization parameter value weight of skip type macroblocks: Dynamically adjust the quantization parameter value weight of skip type macroblocks according to the video content characteristics. For example, in high-motion scenes, the weight of skip macroblocks can be appropriately reduced; in low-motion scenes, the weight of skip macroblocks can be increased. The specific weight adjustment method can be based on preset rules or through real-time learning and optimization of weight parameters through machine learning models.

[0165] Calculate the comprehensive quantization parameter value of the skip type macroblock: For each skip type macroblock in each frame of video, calculate its comprehensive quantization parameter value. The calculation formula of the comprehensive quantization parameter value is:

[0166]

[0167] Where n is the number of skip macroblocks, skip_qp i is the quantization parameter value of the i-th skip macroblock, weight i is the weight of the i-th skip macroblock.

[0168] Calculate the comprehensive average quantization parameter value: combine the average quantization parameter value of the non-skip macroblock and the comprehensive quantization parameter value of the skip macroblock to calculate the comprehensive average quantization parameter value. The specific calculation formula is:

[0169]

[0170] Determine the quality score of the video bitstream: Determine the quality score of the video bitstream based on the comprehensive average quantization parameter value. The mapping formula of the quality score can be: quality score = 100-comprehensive average quantization parameter value. For example, if the comprehensive average quantization parameter value is 40, the quality score is:

[0171] Mass fraction = 100-40 = 60.

[0172] Through the above steps, this embodiment successfully establishes an adaptive mechanism, adjusts the quantization parameter value weight of the skip type macroblock, and calculates the comprehensive average quantization parameter value and quality score. This method ensures the accuracy and integrity of data during the quality assessment process, and is particularly suitable for high-real-time video processing applications, such as video calls and video conferences.

[0173] Furthermore, a seventh embodiment of the method for assessing video quality without reference according to the present invention is proposed, and after the above step S30, the method further includes:

[0174] Step f1, establish a user feedback mechanism, obtain the user's subjective score according to the user feedback mechanism, and determine the comprehensive score according to the user's subjective score and the quality score.

[0175] In this embodiment, how to establish a user feedback mechanism after calculating the average quantization parameter value and determining the quality score of the video bitstream is described in detail, and how to determine the comprehensive score by combining the user's subjective rating and the quality score. The specific steps may include:

[0176] Calculate the average quantization parameter value: Calculate the average quantization parameter value of the video frame according to the above steps. For example, assume that the average quantization parameter value of a certain frame of video is 32.4.

[0177] Determine the quality score of the video bitstream: Determine the quality score of the video bitstream according to the calculated average quantization parameter value. The mapping formula of the quality score may be: quality score = 100 - average quantization parameter value.

[0178] For example, if the average quantization parameter value is 32.4, the quality score is:

[0179] Mass fraction = 100-32.4 = 67.6.

[0180] Establish a user feedback mechanism: Develop a user feedback mechanism to collect users' subjective ratings of video quality through a user interface (such as a video player or video conferencing software). The user rating range can be set from 1 to 5, with 1 indicating very poor quality and 5 indicating very good quality. During or after the video is played, prompt the user to rate the quality of the current video. Collect user rating data and store it in a feedback database.

[0181] Collect user subjective ratings: Collect multiple users' subjective ratings of the same video and calculate the average of the user ratings. For example, suppose the user subjective rating data of a video is as follows:

[0182] User 1 Rating: 4

[0183] User 2 Rating: 3

[0184] User 3 Rating: 4

[0185] User4 Rating: 5

[0186] User 5 Rating: 3

[0187] The average value of the user's subjective rating is calculated as: 3.8.

[0188] Combine user subjective ratings and quality scores: Combine the quality score calculated by the system with the user's subjective rating to determine the comprehensive score. The calculation formula for the comprehensive score can be designed according to actual needs. Common methods are weighted average or linear combination. Set the weights of the quality score and the user's subjective rating. For example, the weight of the quality score is 0.7 and the weight of the user's subjective rating is 0.3. Calculate the comprehensive score:

[0189] Comprehensive score = quality score × 0.7 + average user subjective score × 20 × 0.3

[0190] Among them, the user's subjective rating is magnified by 20 times to match the scale of the quality score.

[0191] For example, quality score: 67.6, average user subjective rating: 3.8, user subjective rating magnification value: 3.8\times 20=76.

[0192] Comprehensive score calculation: Comprehensive score = 67.6 × 0.7 + 76 × 0.3 = 47.32 + 22.8 = 70.12

[0193] Through the above steps, this embodiment establishes a user feedback mechanism, combines the user's subjective rating and the quality score calculated by the system, and determines a more comprehensive rating. This method not only takes into account the objective quantitative parameter value, but also integrates the user's subjective feelings, thereby improving the accuracy of video quality assessment and user satisfaction.

[0194] The invention also provides a video quality non-reference evaluation method and device.

[0195] Reference Figure 3 , Figure 3 This is a functional module diagram of a first embodiment of a method and apparatus for evaluating video quality without reference according to the present invention. The method and apparatus for evaluating video quality without reference comprises:

[0196] A reassembly module is used to receive a video code stream encapsulated in H.264, reframe the video code stream through a jitter buffer, and obtain a reassembled code stream;

[0197] A parsing module, used for parsing the recombined bitstream to obtain a macroblock type and a quantization parameter value;

[0198] An evaluation module is used to determine an average quantization parameter value of a macroblock whose macroblock type is a non-skip type in each frame of the reconstructed bitstream according to the macroblock type and the quantization parameter value, and determine a quality score of the video bitstream according to the average quantization parameter value.

[0199] In addition, an embodiment of the present invention further provides a storage medium.

[0200] The storage medium stores a video quality no-reference assessment method program, and when the video quality no-reference assessment method program is executed by the processor, the steps of the video quality no-reference assessment method described above are implemented.

[0201] The specific implementation of the storage medium of the present invention is basically the same as the embodiments of the above-mentioned video quality non-reference evaluation method, and will not be repeated here.

[0202] The embodiments of the present invention are described above in conjunction with the accompanying drawings, but the present invention is not limited to the above-mentioned specific implementation methods. The above-mentioned specific implementation methods are merely illustrative and not restrictive. Under the enlightenment of the present invention, ordinary technicians in this field can also make many forms without departing from the scope of protection of the present invention and the claims. All equivalent structures or equivalent process changes made using the contents of the specification and drawings of the present invention, or directly or indirectly used in other related technical fields, are protected by the present invention.

[0203] In the embodiments provided in the present application, it should be understood that the disclosed devices and methods can also be implemented in other ways. The device embodiments described above are merely schematic. For example, the flowcharts and block diagrams in the accompanying drawings show the possible architecture, functions and operations of the devices, methods and computer program products according to the multiple embodiments of the present application. In this regard, each box in the flowchart or block diagram can represent a module, a program segment or a part of a code, and the module, a program segment or a part of a code contains one or more executable instructions for implementing 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 consecutive boxes 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 and / or the flowchart, and the combination of boxes in the block diagram and / or the flowchart can be implemented with a dedicated hardware-based system that performs a specified function or action, or can be implemented with a combination of dedicated hardware and computer instructions.

[0204] In addition, each functional module in each embodiment of the present application can be integrated together to form an independent part, or each module can exist separately, or two or more modules can be integrated to form an independent part. If the function is implemented in the form of a software functional module and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on such an understanding, the technical solution of the present application is essentially or the part that contributes to the prior art or the part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including several instructions to enable a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM, Read-Only Memory), a random access memory (RAM, Random Access Memory), a disk or an optical disk.

[0205] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprise a ..." do not exclude the existence of other identical elements in the process, method, article or device including the elements.

[0206] The above are only various implementations of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Claims

1. A video quality no-reference assessment method, characterized in that: The video quality no-reference assessment method comprises the following steps: Receiving a video code stream encapsulated in H.264, reframing the video code stream through a jitter buffer, and obtaining a recombined code stream; Parsing the reorganized bitstream to obtain macroblock type and quantization parameter value; An average quantization parameter value of macroblocks whose macroblock type is non-skip type in each frame of the reorganized bitstream is determined according to the macroblock type and the quantization parameter value, and a quality score of the video bitstream is determined according to the average quantization parameter value.

2. The video quality non-reference assessment method according to claim 1, characterized in that: The step of reframing the video code stream through a jitter buffer comprises: Packet loss, jitter and / or disorder of the video code stream caused by network congestion are processed to complete re-framing, thereby preventing incomplete reception of video frames of the video code stream.

3. The video quality non-reference assessment method according to claim 1, characterized in that: The step of parsing the reorganized bitstream to obtain the macroblock type and quantization parameter value comprises: The slice and macroblock structure data are obtained by parsing the recombined code stream, including obtaining each macroblock type and the quantization parameter value.

4. The video quality non-reference assessment method according to claim 1, characterized in that: The step of obtaining the macroblock type and the quantization parameter value comprises: The macroblock type and the quantization parameter value are obtained, where the quantization parameter value ranges from an integer between 0 and 51.

5. The video quality non-reference assessment method according to claim 1, characterized in that: The step of determining the average quantization parameter value of the macroblocks whose macroblock type is non-skip type in each frame of the recombined bitstream according to the macroblock type and the quantization parameter value comprises: The macroblocks whose macroblock type in each frame of the reorganized bitstream are excluded, and the average quantization parameter value is determined according to the quantization parameter values ​​of the remaining non-skip type macroblocks in each frame of the reorganized bitstream.

6. The video quality non-reference assessment method according to claim 1, characterized in that: The step of determining the quality score of the video stream according to the average quantization parameter value comprises: An adaptive mechanism is established to adjust the quantization parameter value weight of the skip type macroblock according to different video contents in the reconstructed bitstream, determine a comprehensive average quantization parameter value according to the quantization parameter value of the skip type macroblock, the quantization parameter value weight and the average quantization parameter value, and determine the quality score of the video bitstream according to the comprehensive average quantization parameter value.

7. The video quality non-reference assessment method according to claim 1, characterized in that: After the step of determining the quality score of the video stream according to the average quantization parameter value, the method further includes: A user feedback mechanism is established, user subjective scores are obtained according to the user feedback mechanism, and a comprehensive score is determined according to the user subjective scores and the quality score.

8. A method and device for evaluating video quality without reference, characterized in that: The video quality no-reference assessment method device comprises: A reassembly module is used to receive a video code stream encapsulated in H.264, reframe the video code stream through a jitter buffer, and obtain a reassembled code stream; A parsing module, used for parsing the recombined bitstream to obtain a macroblock type and a quantization parameter value; An evaluation module is used to determine an average quantization parameter value of a macroblock whose macroblock type is a non-skip type in each frame of the reconstructed bitstream according to the macroblock type and the quantization parameter value, and determine a quality score of the video bitstream according to the average quantization parameter value.

9. A video quality non-reference assessment method and device, characterized in that: The video quality no-reference assessment method device includes a memory, a processor, and a video quality no-reference assessment method program stored in the memory and executable on the processor. When the video quality no-reference assessment method program is executed by the processor, the steps of the video quality no-reference assessment method as described in any one of claims 1 to 7 are implemented.

10. A computer storage medium, characterized in that: The storage medium stores a video quality no-reference assessment method program, which, when executed by a processor, implements the steps of the video quality no-reference assessment method according to any one of claims 1 to 7.

Images