Evaluation and early warning method and system for video playback quality based on the middle number platform

Abstract

The present invention discloses an evaluation and early warning method and system for video playback quality based on an intermediate number platform. The method of the present invention comprises obtaining video data to be evaluated on the intermediate number platform; extracting data type features of a test data set; inputting the data type features of the test data set into a trained video quality evaluation model to perform video quality classification evaluation to obtain a first video quality evaluation result; and performing a video quality early warning according to the first video quality evaluation result to obtain a video quality early warning result. The present invention can use the signaling and media stream features collected by the intermediate number platform to perform model training, so that the trained model can effectively evaluate and warn the playback quality of the video advertisements of the intermediate number when the video playback of the terminal device cannot be obtained.

Description

Technical Field

[0001] The present invention relates to the field of communication technology, and in particular to an evaluation and early warning method, system, device and storage medium for video playback quality based on an intermediate number platform. Background Art

[0002] At present, the Internet has profoundly changed people's lives. People's food, clothing, housing and transportation are all covered by the Internet. All traditional enterprises are moving towards the Internet, and Internet+ has become the country's major strategy and direction. How to protect one's personal privacy in the Internet wave has gradually become a top priority. Virtual intermediate number technology can provide a solution for the O2O (online to offline) industry to hide the real numbers of both parties in communication, and enable merchants and customers to communicate with each other by phone or text message through a temporary number.

[0003] With the advancement of technology, virtual intermediate numbers have the ability to play customized video ads to the caller's mobile phone through SIP signaling when dialing the intermediate number. For example, when a customer who orders food contacts the takeaway boy through the intermediate number, a video promotional ad is played for the company. Intermediate number corporate customers are usually very concerned about the playback quality of intermediate number video ads. Due to the influence of transcoding, compression, transmission, and differences in playback equipment, intermediate number video ads will produce different types and degrees of distortion, resulting in a decrease in video quality. Therefore, studying efficient and accurate intermediate number video ad playback quality evaluation methods is of great significance for the iteration of intermediate number video technology and improving corporate customer satisfaction.

[0004] Existing video quality evaluation methods mainly include subjective video quality evaluation methods and objective video quality evaluation methods. Subjective video quality evaluation methods usually play the image sequence to be evaluated for the evaluator to watch in a controlled test environment, and then the evaluator subjectively gives a score based on the evaluation criteria. Finally, the scores of all evaluators are statistically analyzed and the average score is obtained as the final score of the video. Objective video quality evaluation methods (full reference methods) usually compare the video to be evaluated with the original video to obtain the similarity or fidelity of the image. The image algorithms commonly used are: mean square error / peak signal-to-noise ratio method to measure the distortion of the video sequence. Video quality evaluation method based on the physiological characteristics of HVS (human visual system). Image and video quality evaluation method based on structural distortion.

[0005] The subjective video quality evaluation method is the most realistic method to reflect human visual perception. However, due to the influence of objective conditions, differences in human subjective factors, and differences in human visual sensitivity, different observers will have differences in evaluating videos, resulting in inaccurate results. In addition, the subjective video quality evaluation method requires repeated experiments, and has the disadvantages of complex steps to build a test environment, long duration, and high cost. The objective video quality evaluation method is actually based on image similarity to evaluate quality. Take the mean square error as an example, two videos to be evaluated have the same mean square error value. Since the real users of the video ads in the middle are real users served by corporate customers, it is difficult to obtain the real playback situation on the real users' terminals. Compatibility tests can be performed on some models in the laboratory environment, but the real playback situation in the production environment cannot be evaluated. Summary of the invention

[0006] The present invention aims to solve one of the technical problems in the related art at least to a certain extent.

[0007] To this end, the present invention proposes a method for evaluating and warning the video playback quality based on the intermediate number platform, and uses the signaling and media stream features collected by the intermediate number platform to train the model. The trained model can evaluate and warn the playback quality of the video advertisements of the intermediate number when the video playback of the terminal device cannot be obtained.

[0008] Another object of the present invention is to propose an evaluation and early warning system for video playback quality based on an intermediate number platform.

[0009] A third object of the present invention is to provide a computer device.

[0010] A fourth object of the present invention is to provide a non-transitory computer-readable storage medium.

[0011] To achieve the above-mentioned purpose, the present invention proposes, on one hand, a method for evaluating and warning the video playback quality based on an intermediate number platform, comprising:

[0012] Obtaining video data to be evaluated from the intermediate platform; wherein the video data to be evaluated includes a training data set and a test data set;

[0013] Extracting data type characteristics of the test data set;

[0014] Inputting the data type characteristics of the test data set into a trained video quality assessment model for video quality classification and evaluation to obtain a first video quality assessment result; wherein the trained video quality assessment model is obtained by performing model training based on the data type characteristics of the training data set as the input layer of the classifier and the second video quality assessment result as the classification result layer;

[0015] A video quality warning is performed according to the first video quality evaluation result to obtain a video quality warning result.

[0016] The video playback quality evaluation and early warning method based on the intermediate number platform of the embodiment of the present invention may also have the following additional technical features:

[0017] In one embodiment of the present invention, the data type characteristics of the test data set include: SIP signaling characteristics and media stream characteristics; wherein the SIP signaling characteristics include the number of routes through which the signaling passes, the time difference between signaling sending and ringing, SDP characteristic parameter fusion parameters, and delay characteristic parameters of the SIP signaling process; the media stream characteristics include the number of media negotiation times before playing the video advertisement and the number of media RTP packets transmitted when playing the video advertisement.

[0018] In one embodiment of the present invention, if the intermediate number is the same as the home domain of the called terminal; extracting the number of routes through which the signaling passes includes:

[0019] The calling terminal initiates an INVITE request to the called terminal;

[0020] The INVITE request is sent to the S-CSCF route to which the calling terminal belongs through the P-CSCF route to which the calling terminal belongs;

[0021] The INVITE request is sent to the S-CSCF route to which the intermediate number belongs through the I-CSCF to which the intermediate number belongs, and the INVITE request is sent to the P-CSCF route to which the called party belongs through the S-CSCF route to which the intermediate number belongs;

[0022] The INVITE request is sent to the called terminal through the P-CSCF to which the called party belongs.

[0023] In one embodiment of the present invention, if the intermediate number is different from the home domain of the called terminal; extracting the number of routes through which the signaling passes includes:

[0024] The calling terminal initiates an INVITE request to the called terminal;

[0025] The INVITE request is sent to the S-CSCF route to which the calling terminal belongs through the P-CSCF route to which the calling terminal belongs;

[0026] The INVITE request is sent to the S-CSCF route to which the intermediate number belongs through the I-CSCF to which the intermediate number belongs, and the INVITE request is sent to the I-CSC route to which the called terminal belongs through the S-CSCF route to which the intermediate number belongs;

[0027] The INVITE request is sent to the S-CSCF route of the called terminal through the I-CSCF route of the called terminal. The S-CSCF route of the called terminal sends the INVITE request to the P-CSCF route of the called terminal. The INVITE request is sent to the called terminal through the P-CSCF route of the called terminal.

[0028] In one embodiment of the present invention, extracting the time difference between sending the signaling and ringing includes:

[0029] Obtaining a first time point based on call signaling; wherein the first time point includes a time point when the called terminal information is obtained through a transfer query after the call signaling enters the intermediate number platform;

[0030] Obtain a second time point based on the ringing of the called terminal; wherein the second time point includes the time point when the intermediate number platform returns the ringing of the called terminal to the calling terminal;

[0031] The time difference between signaling sending and ringing is calculated based on the first time point and the second time point.

[0032] In one embodiment of the present invention, extracting the SDP characteristic parameter fusion parameter includes:

[0033] The SDP_flow characteristic parameter fusion parameter is used to characterize the application layer control protocol SDP flow parameters, and the SDP flow parameters are defined as:

[0034]

[0035] b is the value in the b-bandwidth row, indicating the bandwidth value of this media negotiation, b=<type>:<bandwidth>.

[0036] In one embodiment of the present invention, extracting the delay characteristic parameter of the SIP signaling process includes:

[0037] The intermediate number application server initiates a media negotiation request to the calling terminal;

[0038] Based on the media negotiation signal, when a media negotiation request returned by the calling terminal is received, it is recorded as a request response process, and the corresponding time difference is recorded;

[0039] The time difference corresponding to all request-response processes during the entire call is recorded, so that SIP_delay_time is recorded as the delay characteristic parameter of the SIP signaling process.

[0040] In one embodiment of the present invention, extracting the number of media negotiation times before playing the video advertisement includes:

[0041] Obtain an INVITE request including first media coding information of a calling terminal;

[0042] In response to the INVITE request, the intermediate number application server re-filters and sorts the second media coding information supported by the intermediate number platform and sends it to the called terminal;

[0043] The third media coding information supported by the called terminal is returned to complete a media negotiation request based on the returned result.

[0044] In one embodiment of the present invention, extracting the number of media RTP packets transmitted when playing the video advertisement includes:

[0045] Gets the number of media RTP packets sent by the media server to the calling terminal.

[0046] In one embodiment of the present invention, the classifier is a neural network model, and the second video quality evaluation result is an artificial scoring value obtained by scoring a training data set using a subjective video quality evaluation method; training the video quality evaluation model includes:

[0047] Extracting data type features of the training data set;

[0048] Using the data type characteristics of the training data set as the input layer of the neural network model, and using the manual scoring value as the classification result layer of the neural network model to perform network model training;

[0049] The trained video quality assessment model is obtained based on the trained neural network model.

[0050] To achieve the above-mentioned purpose, the present invention proposes, on the other hand, a video playback quality evaluation and early warning system based on an intermediate number platform, comprising:

[0051] A video data acquisition module is used to acquire the video data to be evaluated from the intermediate number platform; wherein the video data to be evaluated includes a training data set and a test data set;

[0052] A data feature extraction module, used to extract data type features of the test data set;

[0053] A video quality evaluation module, used for inputting the data type characteristics of the test data set into a trained video quality evaluation model for video quality classification evaluation to obtain a first video quality evaluation result; wherein the trained video quality evaluation model is obtained by performing model training based on the data type characteristics of the training data set as the input layer of the classifier and the second video quality evaluation result as the classification result layer;

[0054] The video quality warning module is used to perform a video quality warning according to the first video quality evaluation result to obtain a video quality warning result.

[0055] The video playback quality evaluation and early warning method and system based on the intermediate number platform of the embodiment of the present invention can effectively evaluate and warn the video services on the intermediate number line to obtain the current video playback quality and improve the user's comfort.

[0056] To achieve the above-mentioned purpose, the third aspect embodiment of the present application proposes a computer device, including: a processor and a memory; wherein the processor runs a program corresponding to the executable program code by reading the executable program code stored in the memory, so as to implement the video playback quality evaluation and early warning method based on the intermediate number platform as described in the first aspect embodiment.

[0057] To achieve the above-mentioned purpose, the fourth aspect embodiment of the present application proposes a non-temporary computer-readable storage medium, on which a computer program is stored. When the program is executed by the processor, it implements the video playback quality evaluation and early warning method based on the intermediate number platform as described in the first aspect embodiment.

[0058] Additional aspects and advantages of the present invention will be given in part in the following description and in part will be obvious from the following description, or will be learned through practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] The above and / or additional aspects and advantages of the present invention will become apparent and easily understood from the following description of the embodiments in conjunction with the accompanying drawings, in which:

[0060] Figure 1 is a flow chart of a method for evaluating and warning the video playback quality based on an intermediate number platform according to an embodiment of the present invention;

[0061] Figure 2 It is a framework diagram of a method for evaluating and warning the video playback quality based on an intermediate number platform according to an embodiment of the present invention;

[0062] Figure 3 is a schematic diagram of a continuous dual-stimulus quality measurement method according to an embodiment of the present invention;

[0063] Figure 4 This is a schematic diagram of signaling routing when the intermediate number and the called party have the same home domain according to an embodiment of the present invention;

[0064] Figure 5 This is a schematic diagram of signaling routing according to an embodiment of the present invention in which the intermediate number and the called party have different attribution domains;

[0065] Figure 6is a schematic diagram of a method for calculating the signaling time difference of a middle video advertisement according to an embodiment of the present invention;

[0066] Figure 7 is a structural schematic diagram of a video playback quality evaluation and early warning system based on an intermediate number platform according to an embodiment of the present invention;

[0067] Figure 8 is a computer device according to an embodiment of the present invention. DETAILED DESCRIPTION

[0068] It should be noted that, in the absence of conflict, the embodiments of the present invention and the features in the embodiments can be combined with each other. The present invention will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0069] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present invention.

[0070] The following describes, with reference to the accompanying drawings, a video playback quality evaluation and early warning method, system, device, and storage medium based on an intermediate number platform proposed according to an embodiment of the present invention.

[0071] First, the professional terms that may be used in the present invention are explained, as shown in Table 1:

[0072] Table 1

[0073]

[0074]

[0075]

[0076] The framework diagram of the method of the present invention is roughly as follows: first, a series of intermediate number video advertisements played by terminals and their corresponding signaling and media stream characteristic parameters passing through the intermediate number platform are collected in a laboratory environment as a training set. The signaling and media stream characteristic parameters of the training set video are input into the classifier as the input layer (an artificial neural network can be used as a classifier), and then the training set video quality (a subjective video quality evaluation method can be used) is input into the classifier as the classification result layer and trained. The trained classification model can evaluate and warn the intermediate number video advertisements for the signaling and media stream characteristic parameters of the new intermediate number platform. Figure 2 shown.

[0077] Figure 1 is a flow chart of a method for evaluating and warning the video playback quality based on an intermediate number platform according to an embodiment of the present invention. Figure 1 As shown, the method includes but is not limited to the following steps:

[0078] S1, obtaining the video data to be evaluated from the intermediate platform; wherein the video data to be evaluated includes a training data set and a test data set.

[0079] It is understandable that when the present invention performs actual testing on the video data to be evaluated on the intermediate number platform, it is necessary to first obtain a training data set of the video data to be evaluated.

[0080] It is understandable that the training data set is mainly used for subsequent model training, and the test set is then used to perform actual video quality evaluation based on the trained model.

[0081] It is understandable that the video data to be evaluated mainly includes a series of intermediate video advertisements played by terminals collected in a laboratory environment.

[0082] S2, extract the data type characteristics of the test dataset.

[0083] S3, inputting the data type characteristics of the test data set into the trained video quality assessment model for video quality classification evaluation to obtain a first video quality assessment result; wherein the trained video quality assessment model is obtained by performing model training based on the data type characteristics of the training data set as the input layer of the classifier and the second video quality assessment result as the classification result layer.

[0084] It is understandable that, before extracting data features from the test data set, the present invention first needs to extract data features from the training data set so as to perform model training using the data features of the training data set.

[0085] In one embodiment of the present invention, a video quality evaluation training set is established, and the data types in the training set are determined therefrom, including SIP signaling feature types and media stream feature types.

[0086] Among them, the video scenes in the training set are composed of "video pairs" consisting of the original video and several videos to be evaluated that are collected on mobile phones in the laboratory and played to the calling mobile phone before the call is made through the intermediate number. The video played through the intermediate number will introduce different types and degrees of distortion due to reasons such as mobile phone device compatibility, differences in three-network signals, and cross-network elements.

[0087] Among them, each original video corresponds to no less than 10 collected intermediate videos to be evaluated, and the length of the videos to be evaluated is no less than 10 seconds. The Double Stimulus Continuous Quality Scale (DSCQS) is used for the videos to be evaluated. This method is a way to evaluate video quality based on the subjective visual / auditory feelings of experts. The specific method is to let experts watch a "video pair" consisting of multiple original reference videos and videos to be evaluated (distorted videos). The display order of the original videos and the videos to be evaluated is random. The experts score each video in the "video pair" and finally take the average value to obtain the expert score classification category of the videos to be evaluated, which has 5 levels: very poor, poor, general, good, and very good. In addition, the video scenes in the training set include but are not limited to buildings, people, indoors, macro close-ups, static objects, natural scenery, sports competitions, literary performances, large-scale mass activities and other types of videos.

[0088] In one embodiment of the present invention, during the scoring process, the expert is given a video of Figure 3 The video sequence shown is divided into a preliminary test phase and a scoring phase. The preliminary test phase does not require scoring, and scoring is performed when the video is played again. Among them, T1 and T3 time periods are the time for playing the video, and T2 and T4 are the rest time (displayed in gray on the monitor). T1 and T3 play the same video, T1 is randomly the original video and the video to be evaluated, and T3 is another video. The experts cannot know the viewing order in advance before the test.

[0089] It can be understood that the video quality evaluation method in the present invention includes but is not limited to DSCQS, DSIS (dual stimulus evaluation method), SS (single stimulus evaluation method), SC (stimulus comparison method), SSCQE (single stimulus continuous quality evaluation method), SAMVIQ (single stimulus continuous quality evaluation method), SAMVIQ (Subjective Assessment Method for Video Quality evaluation) and other methods in subjective video quality evaluation methods, and includes but is not limited to MSE (mean square error method), PSNR (peak signal-to-noise ratio), SSIM (method based on structural similarity), Video-SSIM (video quality evaluation model based on SSIM), methods based on visual features, methods based on perceptual features and other methods in objective video quality evaluation methods.

[0090] In one embodiment of the present invention, the SIP signaling feature types of the training data set include (1) the number of routes through which the signaling passes, (2) the time difference between the signaling sending and the ringing, (3) the SDP feature parameter fusion parameter (SDP_flow), and (4) the delay feature parameter of the SIP signaling process (SIP_delay_time), so as to train the model ANN using the extracted SIP signaling feature types of the training data set.

[0091] It is understandable that, considering that the prior art does not consider obtaining corresponding features for evaluating video quality from the perspective of SIP signaling, and due to the reliability of SIP signaling record information, the present invention evaluates video quality based on information features from this perspective.

[0092] Since the video to be evaluated is transmitted to the calling mobile phone through the SIP signaling of the intermediate number system, its SIP signaling feature parameters will also affect its video quality. The SIP signaling feature parameters of the video played based on the intermediate number in the training set are extracted for subsequent model training.

[0093] Preferably, the signaling passes through a routing number.

[0094] It is understandable that SIP signaling contains a series of header fields, among which the value of the Record-Route header field (Proxies) is inserted when the signaling passes through the Proxy, so that subsequent requests with the same Call-ID will be routed to these Proxies. A SIP call will pass through different numbers of routes depending on the different home domains of the caller and the callee.

[0095] For example, Figure 4 As shown in the figure, if the calling party is in Guangzhou, the intermediate number is in Xi'an, and the called party is also in Xi'an, and the intermediate number and the called party are in the same home domain, the routing path of the signaling request is as follows Figure 4 As shown, the process is:

[0096] 1: The calling terminal initiates an INVITE request to the called party, which first passes through the P-CSCF to which the calling terminal belongs.

[0097] 2: Routing from the P-CSCF to which the caller belongs to the S-CSCF to which the caller belongs.

[0098] 3.1: Since the calling party and the intermediate number belong to different domains, the first request will need to be sent to the I-CSCF to which the intermediate number belongs.

[0099] 3.2: The I-CSCF to which the intermediate number belongs continues to send a request to the S-CSCF to which the intermediate number belongs.

[0100] 3: Due to the existence of Record-Route, subsequent requests initiated by the caller do not need to be sent to the I-CSCF to which the intermediate number belongs, and the I-CSCF to which the intermediate number belongs can be directly found.

[0101] 4: The intermediate number application server (AS) modifies the SIP signaling to enable the intermediate number to have the video business card function.

[0102] 5: The request is sent to the P-CSCF of the called party. It is assumed that the intermediate number and the called party belong to the same S-CSCF.

[0103] 6: The called terminal receives the request from the calling terminal.

[0104] For example, Figure 5 As shown in the figure, if the calling party is in Guangzhou, the middle number is in Xi'an, and the called party is in Beijing, the middle number and the called party's home network are different. If the middle number and the called party's home domain are different, the following process will be added if they are the same:

[0105] 5.1: Since the intermediate number and the called party's home domain are different, when the first request is made, the S-CSCF to which the intermediate number belongs forwards the request to the I-CSCF to which the called party belongs.

[0106] 5.2: The I-CSCF to which the called party belongs sends a request to the S-CSCF to which the called party belongs.

[0107] 5: Due to the existence of Record-Route, subsequent requests initiated by the calling party do not need to pass through the I-CSCF of the called party again.

[0108] It can be known that because the calling party, the intermediate number, and the called party belong to different domains (across provinces and operators), a request will pass through several routes, and when passing through the Proxy, its own information may be added to the Record-Route header field (voluntary participation). The subsequent requests initiated by the calling party will use the routing information in the Record-Route header field.

[0109] It is understandable that routing is divided into strict routing (Strict Routing) and loose routing (LooseRouting, lr). Strict routing requires that the data packet must strictly follow the path specified by the sender through each route. These routes need to be connected one by one, and the order of the routers cannot be changed. If the data packet cannot find the route specified by the next hop during transmission, it will return an error of source routing failure and unreachable destination to the sender. If there is no lr parameter in the Route header field, it is strict routing. Loose routing does not require a complete and strict routing path to be listed, but only certain key points in the path need to be given. Routing can be performed between key points through the router's automatic routing selection function. Loose routing is carried out in Record-Route with the lr parameter.

[0110] For example, Route_params is defined as a route-related parameter, and its specific calculation formula is:

[0111] Route_params=αβ(Route_st+γRoute_lr)

[0112] Among them, α is the weight related to cross-province, β is the weight related to cross-operator, γ is the loose routing weight (γ>1), Route_st is the number of strict routes of Record-Route in the call, and Route_lr is the number of loose routes of Record-Route in the call.

[0113] Preferably, the time difference between signaling sending and ringing.

[0114] In one embodiment of the present invention, after a call signaling enters the intermediate number platform and obtains the called party information through the transfer query, the signaling will be connected out, and the recorded time point at this time is T1.

[0115] After the media negotiation process, the called device returns the 180 signaling (indicating ringing), and the intermediate number platform returns the 180 signaling to the calling device. The time point is T2.

[0116] The calculation formula for the time difference between signaling sending and ringing is:

[0117] δ=T2-T1

[0118] For example, the signaling interaction process of a middle number video advertisement is as follows: Figure 6 shown.

[0119] Preferably, the SDP characteristic parameters are fused parameters (SDP_flow).

[0120] It is understandable that SDP (Session Description Protocol) is an application layer control protocol used to describe multimedia sessions. It is text-based and is used to negotiate media types and encodings during session establishment in the communication field. SDP_flow is used to characterize SDP flow parameters. Considering that its channel is similar to a water pipe, the SDP flow parameters are defined as:

[0121]

[0122] b is the value in row b (bandwidth row), indicating the bandwidth value negotiated for this media, b=<type>:<bandwidth>, where bandwidth (in kbit / s) can be the value negotiated in update(video) before playing the video advertisement.

[0123] In the m line (media line), video represents the video line m, m= <media> <port> <proto> <fmt>, and further, framerate and framesize represent the conditions under which framerate represents the video frame rate in frames per second, and framesize represents the frame size.

[0124] Preferably, the delay characteristic parameter (SIP_delay_time) of the SIP signaling process.

[0125] It is understandable that the video advertisement based on the middle number is a function that plays a video to the caller during the ringing stage of the called party when the caller dials the middle number and the caller is connected to the called party. The delay parameter between the calling terminal and the middle number application server (AS) has the greatest impact on the quality of video advertisement playback.

[0126] The method of collecting the delay characteristic parameter is the time difference between when the AS actively initiates a request to the calling terminal and when it receives a response from the calling terminal. Figure 4 As shown in the figure, AS actively initiates media negotiation update (video) to the calling terminal. When AS receives 200update from the calling terminal, it is a complete request response process, and the time difference is recorded as δ i There will be several request-response processes during the entire call. SIP_delay_time is recorded as the delay characteristic parameter of the SIP signaling process. The three-point estimation method based on Beta distribution commonly used in management is used to calculate the most likely delay:

[0127]

[0128] Where T p is max(δ i ), which is the maximum value among them; T m for That is, the mean value; T p is min(δ i ), which is the minimum value among them.

[0129] In one embodiment of the present invention, the model is trained using the media stream feature types of the above-mentioned training data set, including (1) the number of media negotiations before playing a video advertisement, and (2) the number of media RTP packets (RTP_number) transmitted when playing a video advertisement.

[0130] It is understandable that the above-extracted characteristic parameters related to SIP signaling are actually played by the SIP-controlled media server, so the present invention will also extract media stream characteristic parameters for subsequent ANN training.

[0131] Preferably, the number of media negotiations (Negotiate_time) before playing the video advertisement.

[0132] It can be understood that the intermediate number controls the call through SIP signaling, and SDP, as the message body, is responsible for media negotiation and support capability negotiation during the SIP interaction process. It can negotiate the voice and video media formats so that the calling device and the called device and the network elements and routers in between support a certain media encoding for communication.

[0133] The m line indicates the media description, where audio indicates audio, video indicates video, and the following are the media encoding types it supports. Dynamic media encoding needs to be described using [a=rtpmap] later. The SDP protocol achieves negotiation through an offer / answer mechanism. The specific steps are as follows:

[0134] 1. The calling terminal carries its own media coding information through the INVITE message.

[0135] 2. The intermediate number application server (AS) receives the INVITE message, re-filters and sorts the media coding information supported by the intermediate number platform, and continues to send it to the called terminal.

[0136] 3. The called terminal returns the media coding information it supports through 200 / 180 of reliable transmission, completing a media negotiation.

[0137] Since one or more media negotiations may be required before playing a video ad, and multiple media negotiations may be related to different network elements or signal strength, the present invention selects the number of media negotiations (Negotiate_time) before playing a video ad as a characteristic parameter. The present invention shows a complete call signaling, / v1 / session / update, indicating that the encoding changes after the media negotiation in the call, in which there are 4 media negotiations in total.

[0138] Preferably, the number of media RTP packets (RTP_number) transmitted when playing the video advertisement.

[0139] It is understandable that, since the video advertisement is played by the media server sending RTP packets to the calling device, if the terminal determines that the packet is lost during the receiving process, a NACK message will be generated to notify the media server to retransmit. Therefore, in the present invention, the number of media RTP packets (RTP_number) sent by the media server to the calling mobile phone is also selected as a characteristic parameter.

[0140] It can be understood that artificial neural network (ANN) uses mathematical models to simulate the processing mechanism of complex information by the human brain's nervous system. Its characteristics are parallel distributed processing capabilities, high fault tolerance, intelligence and self-learning capabilities.

[0141] Furthermore, the present invention uses 4 characteristic parameters based on SIP signaling and 2 characteristic parameters based on media stream of the video to be evaluated in the training data, a total of 6 characteristic parameters, as the input layer of the artificial neural network, and the classification result layer is the expert's score (5 points) for the video to be evaluated, with a total of 5 nodes (1-5 points respectively). Three convolution layers are set, the loss function adopts the cross entropy loss function, the learning rate is 0.01, and the maximum number of iterations is 200. The trained artificial neural network model obtains a trained video quality evaluation model.

[0142] Finally, the present invention inputs the data type features of the test data set into the trained video quality evaluation model to perform video quality classification evaluation to obtain a first video quality evaluation result, which is also the score value output by the model.

[0143] It can be understood that the data type characteristics of the test set are the same as the characteristics extracted from the training data set mentioned above.

[0144] It is understandable that the classification methods that can be used in the present invention include but are not limited to least squares method, Bayesian classification method, principal component analysis method, artificial neural network, nearest neighbor algorithm, support vector machine and other classification algorithms.

[0145] S4, performing a video quality warning according to the first video quality evaluation result to obtain a video quality warning result.

[0146] Specifically, according to the video quality evaluation result, that is, the score value of the corresponding middle-number video advertisement quality, a corresponding video quality warning is performed, and users can selectively experience the played video according to the warning information to improve the satisfaction of corporate customers and other users.

[0147] According to the video playback quality evaluation and early warning method based on the intermediate number platform according to the embodiment of the present invention, the call signaling and media stream features collected by the intermediate number platform are used as the classification basis through the real device video scores collected in the laboratory. After the classification model is trained, it can automatically score and warn the video playback quality of all calls on the intermediate number platform. It can also automatically evaluate video ads without consuming labor costs, which helps to find technical directions that can be iterated and continuously improve customer satisfaction and the stability of platform services.

[0148] In order to implement the above embodiment, Figure 7 As shown, this embodiment also provides a video playback quality evaluation and early warning system 10 based on the intermediate number platform. The system 10 includes a video data acquisition module 100, a data feature extraction module 200, a video quality evaluation module 300 and a video quality early warning module 400.

[0149] The video data acquisition module 100 is used to acquire the video data to be evaluated from the intermediate number platform; wherein the video data to be evaluated includes a training data set and a test data set;

[0150] A data feature extraction module 200 is used to extract data type features of a test data set;

[0151] The video quality evaluation module 300 is used to input the data type characteristics of the test data set into the trained video quality evaluation model to perform video quality classification evaluation to obtain a first video quality evaluation result; wherein the trained video quality evaluation model is obtained by performing model training based on the data type characteristics of the training data set as the input layer of the classifier and the second video quality evaluation result as the classification result layer;

[0152] The video quality warning module 400 is used to perform a video quality warning according to the first video quality evaluation result to obtain a video quality warning result.

[0153] According to the video playback quality evaluation and early warning system based on the intermediate number platform of the embodiment of the present invention, the call signaling and media stream features collected by the intermediate number platform are used as the classification basis by using the real device video scores collected in the laboratory. After the classification model is trained, it can automatically score and warn the video playback quality of all calls on the intermediate number platform. It can also automatically evaluate video ads without consuming labor costs, which helps to find technical directions that can be iterated and continuously improve customer satisfaction and the stability of platform services.

[0154] In order to implement the method of the above embodiment, the present invention also provides a computer device, such as Figure 8 As shown, the computer device 600 includes a memory 601 and a processor 602; wherein the processor 602 runs a program corresponding to the executable program code by reading the executable program code stored in the memory 601, so as to implement each step of the method described above.

[0155] In order to implement the above embodiments, the present application also proposes a non-temporary computer-readable storage medium on which a computer program is stored. When the program is executed by a processor, the method described in the above embodiments is implemented.

[0156] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "example", "specific example", or "some examples" etc. means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or more embodiments or examples in a suitable manner. In addition, those skilled in the art may combine and combine the different embodiments or examples described in this specification and the features of the different embodiments or examples, without contradiction.

[0157] In addition, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined as "first" and "second" may explicitly or implicitly include at least one of the features. In the description of the present invention, "plurality" means at least two, such as two, three, etc., unless otherwise clearly and specifically defined.< / fmt> < / proto> < / port> < / media>

Claims

1. A video playback quality evaluation and early warning method based on an intermediate number platform, characterized in that: The method comprises: Obtain the video data to be evaluated from the intermediate number platform; wherein the video data to be evaluated includes a training data set and a test data set, and the video to be evaluated is transmitted to the calling mobile phone through SIP signaling performed by the intermediate number platform; Extracting data type characteristics of the test data set; Inputting the data type characteristics of the test data set into a trained video quality assessment model for video quality classification and evaluation to obtain a first video quality assessment result; wherein the trained video quality assessment model is obtained by training the model based on the data type characteristics of the training data set as the input layer of the classifier and the second video quality assessment result as the classification result layer, the first video quality assessment result is a score value output by the trained video quality assessment model, and the second video quality assessment result is an artificial score value obtained by scoring the training data set using a subjective video quality assessment method; Performing a video quality warning according to the first video quality evaluation result to obtain a video quality warning result; The data type characteristics of the test data set include: SIP signaling characteristics and media stream characteristics. The SIP signaling characteristics include the number of routes through which the signaling passes, the time difference between signaling sending and ringing, SDP characteristic parameter fusion parameters, and delay characteristic parameters of the SIP signaling process; the media stream characteristics include the number of media negotiation times before playing the video advertisement and the number of media RTP packets transmitted when playing the video advertisement. The data type characteristics of the test data set are the same as the data type characteristics of the training data set; If the intermediate number is the same as the domain of the called terminal; extracting the number of routes through which the signaling passes, including: The calling terminal initiates an INVITE request to the called terminal; The INVITE request is sent to the S-CSCF route to which the calling terminal belongs through the P-CSCF route to which the calling terminal belongs; The INVITE request is sent to the S-CSCF route to which the intermediate number belongs through the I-CSCF to which the intermediate number belongs, and the INVITE request is sent to the P-CSCF route to which the called party belongs through the S-CSCF route to which the intermediate number belongs; The INVITE request is sent to the called terminal through the P-CSCF to which the called party belongs; If the intermediate number is different from the domain of the called terminal, extracting the number of routes through which the signaling passes, including: The calling terminal initiates an INVITE request to the called terminal; The INVITE request is sent to the S-CSCF route to which the calling terminal belongs through the P-CSCF route to which the calling terminal belongs; The INVITE request is sent to the S-CSCF route to which the intermediate number belongs through the I-CSCF route to which the intermediate number belongs, and the INVITE request is sent to the I-CSC route to which the called terminal belongs through the S-CSCF route to which the intermediate number belongs; The INVITE request is sent to the S-CSCF route of the called terminal through the I-CSCF route of the called terminal, and the S-CSCF route of the called terminal sends the INVITE request to the P-CSCF route of the called terminal, and the INVITE request is sent to the called terminal through the P-CSCF route of the called terminal; Extracting the delay characteristic parameters of the SIP signaling process includes: The intermediate number application server initiates a media negotiation request to the calling terminal; Based on the media negotiation request, when receiving the media negotiation request returned by the calling terminal, it is recorded as a request response process, and the corresponding time difference is recorded; The time difference corresponding to all request and response processes during the entire call process is recorded to record SIP_delay_time as the delay characteristic parameter of the SIP signaling process.

2. The method according to claim 1, characterized in that Extracting the time difference between the sending of the signaling and the ringing, including: Obtaining a first time point based on call signaling; wherein the first time point includes a time point when the called terminal information is obtained through a transfer query after the call signaling enters the intermediate number platform; Obtain a second time point based on the ringing of the called terminal; wherein the second time point includes the time point when the intermediate number platform returns the ringing of the called terminal to the calling terminal; The time difference between signaling sending and ringing is calculated based on the first time point and the second time point.

3. The method according to claim 1, characterized in that Extracting the SDP characteristic parameter fusion parameter includes: The SDP_flow characteristic parameter fusion parameter is used to characterize the application layer control protocol SDP flow parameters, and the SDP flow parameters are defined as: b is the value in the b-bandwidth row, indicating the bandwidth value of this media negotiation, b=<type>:<bandwidth>.

4. The method according to claim 1, characterized in that: The number of media negotiations before playing the video ad is extracted, including: Obtain an INVITE request including first media coding information of a calling terminal; In response to the INVITE request, the intermediate number application server re-filters and sorts the second media coding information supported by the intermediate number platform and sends it to the called terminal; The third media coding information supported by the called terminal is returned to complete a media negotiation request based on the returned result.

5. The method according to claim 3, characterized in that: Extracting the number of media RTP packets transmitted when playing the video advertisement includes: Gets the number of media RTP packets sent by the media server to the calling terminal.

6. The method according to claim 1, characterized in that The classifier is a neural network model, and the second video quality evaluation result is an artificial scoring value obtained by scoring the training data set using a subjective video quality evaluation method; Training the video quality assessment model includes: Extracting data type features of the training data set; Using the data type characteristics of the training data set as the input layer of the neural network model, and using the manual scoring value as the classification result layer of the neural network model to perform network model training; The trained video quality assessment model is obtained based on the trained neural network model.

7. A video playback quality evaluation and early warning system based on the intermediate number platform, characterized in that: include: A video data acquisition module is used to acquire the video data to be evaluated from the intermediate number platform; wherein the video data to be evaluated includes a training data set and a test data set, and the video to be evaluated is transmitted to the calling mobile phone through SIP signaling performed by the intermediate number platform; A data feature extraction module, used to extract data type features of the test data set; A video quality evaluation module, used for inputting the data type characteristics of the test data set into a trained video quality evaluation model for video quality classification evaluation to obtain a first video quality evaluation result; wherein the trained video quality evaluation model is obtained by training the model based on the data type characteristics of the training data set as the input layer of the classifier and the second video quality evaluation result as the classification result layer, the first video quality evaluation result is the score value output by the trained video quality evaluation model, and the second video quality evaluation result is an artificial score value obtained by scoring the training data set using a subjective video quality evaluation method; A video quality warning module, configured to perform a video quality warning according to the first video quality evaluation result to obtain a video quality warning result; The data type characteristics of the test data set include: SIP signaling characteristics and media stream characteristics. The SIP signaling characteristics include the number of routes through which the signaling passes, the time difference between signaling sending and ringing, SDP characteristic parameter fusion parameters, and delay characteristic parameters of the SIP signaling process; the media stream characteristics include the number of media negotiation times before playing the video advertisement and the number of media RTP packets transmitted when playing the video advertisement. The data type characteristics of the test data set are the same as the data type characteristics of the training data set; If the intermediate number is the same as the called terminal's home domain; the data feature extraction module is also used to: The calling terminal initiates an INVITE request to the called terminal; The INVITE request is sent to the S-CSCF route to which the calling terminal belongs through the P-CSCF route to which the calling terminal belongs; The INVITE request is sent to the S-CSCF route to which the intermediate number belongs through the I-CSCF to which the intermediate number belongs, and the INVITE request is sent to the P-CSCF route to which the called party belongs through the S-CSCF route to which the intermediate number belongs; The INVITE request is sent to the called terminal through the P-CSCF to which the called party belongs; If the intermediate number and the called terminal belong to different domains; the data feature extraction module is also used to: The calling terminal initiates an INVITE request to the called terminal; The INVITE request is sent to the S-CSCF route to which the calling terminal belongs through the P-CSCF route to which the calling terminal belongs; The INVITE request is sent to the S-CSCF route to which the intermediate number belongs through the I-CSCF route to which the intermediate number belongs, and the INVITE request is sent to the I-CSC route to which the called terminal belongs through the S-CSCF route to which the intermediate number belongs; The INVITE request is sent to the S-CSCF route of the called terminal through the I-CSCF route of the called terminal, and the S-CSCF route of the called terminal sends the INVITE request to the P-CSCF route of the called terminal, and the INVITE request is sent to the called terminal through the P-CSCF route of the called terminal; The data feature extraction module is also used for: The intermediate number application server initiates a media negotiation request to the calling terminal; Based on the media negotiation request, when receiving the media negotiation request returned by the calling terminal, it is recorded as a request response process, and the corresponding time difference is recorded; The time difference corresponding to all request and response processes during the entire call process is recorded to record SIP_delay_time as the delay characteristic parameter of the SIP signaling process.

8. A computer device, characterized in that: including a processor and a memory; Among them, the processor runs the program corresponding to the executable program code by reading the executable program code stored in the memory, so as to implement the video playback quality evaluation and early warning method based on the intermediate number platform as described in any one of claims 1-6.

9. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, an evaluation and early warning method for video playback quality based on an intermediate number platform as described in any one of claims 1-6 is implemented.

Images