A method and system for evaluating content quality of OTT platforms based on data analysis

Abstract

The present invention discloses a method and system for evaluating the content quality of an OTT platform based on data analysis. The original video data of the target video and the video data received by the user, the performance data of the user device and the network transmission performance data are obtained, and feature extraction is performed to obtain the original video features and the video features received by the user; the original video features and the video features received by the user are compared by a fingerprint algorithm to evaluate the quality of the video received by the user and obtain a first quality evaluation result; motion estimation is performed to evaluate the stability and motion smoothness of the video and obtain a second quality evaluation result; the above evaluation results are combined to obtain a comprehensive quality evaluation result; the impact of the device and network performance on the video quality is analyzed to obtain the impact data; and the pre-caching operation is performed according to the impact data to optimize the video quality. The present invention can accurately evaluate and optimize the quality of the video received by the user and improve the user's viewing experience.

Description

Technical Field

[0001] The present invention relates to the technical field of data analysis, and in particular to an OTT platform content quality assessment method and system based on data analysis. Background Art

[0002] With the rapid development of Internet technology, OTT (Over-The-Top) platforms have become one of the main ways for modern people to obtain video content. OTT platforms provide video services to users through the Internet, and users can watch the required video content at any time and any place through various terminal devices (such as smart TVs, mobile phones, tablets, etc.). However, in some public places, such as subways, shopping malls, museums, etc., it is necessary to obtain fixed types of video content on the OTT platform for playback. At this time, the user device is an information bulletin screen, publicity or popular science display screen, and due to differences in user device performance and network transmission quality in public places, the quality of the video received by users is often uneven, affecting the user's viewing experience.

[0003] Existing OTT platform content quality assessment methods mainly rely on user subjective ratings and simple technical indicators (such as resolution, frame rate, etc.), which cannot fully reflect the true quality of the video received by users. In addition, existing methods lack consideration of user device performance and network transmission performance, making it difficult to accurately identify and analyze key factors affecting video quality, thus failing to effectively improve users' viewing experience.

[0004] In order to overcome the shortcomings of the prior art, a method and system for evaluating the content quality of an OTT platform based on data analysis is proposed. By comprehensively collecting and analyzing the video content pushed to users in the OTT platform, that is, the original video data, and the video data received by the user, the user device performance data, and the network transmission performance data, the video quality is evaluated from multiple dimensions, aiming to provide a more scientific and accurate content quality evaluation method. The present invention can not only accurately evaluate the overall quality of the video, but also identify the key factors affecting the video quality, thereby providing users with an effective video quality optimization solution when receiving the video content pushed by the OTT platform, and improving the user's viewing experience. Summary of the invention

[0005] In order to solve at least one of the above technical problems, the present invention proposes a method and system for evaluating content quality of an OTT platform based on data analysis.

[0006] A first aspect of the present invention provides an OTT platform content quality assessment method based on data analysis, comprising:

[0007] Obtaining original video data of a target video in an OTT platform, user received video data, user device performance data, and network transmission performance data, and performing feature extraction on the original video data and user received video data to obtain original video features and user received video features;

[0008] Comparing the original video features with the user received video features based on a fingerprint algorithm, determining a feature similarity index between the user received video data and the original video data, and performing a quality assessment on the user received video according to the feature similarity index to obtain a first quality assessment result;

[0009] Performing motion estimation on the video data received by the user, determining the stability and motion smoothness of the video received by the user, and performing a secondary quality assessment on the video received by the user according to the stability and motion smoothness to obtain a second quality assessment result;

[0010] Performing a comprehensive quality assessment on the video received by the user on the OTT platform according to the first quality assessment result and the second quality assessment result to obtain a comprehensive quality assessment result of the video received by the user;

[0011] Performing correlation analysis on the user equipment performance data and the network transmission performance data and the comprehensive quality evaluation result to determine the impact of the user equipment performance and the network transmission performance on the quality of the video received by the user, and obtaining impact data;

[0012] According to the impact data, the user-pushed videos in the OTT platform are cached in advance to obtain a solution for optimizing the quality of the videos received by the users.

[0013] In this solution, the original video data of the target video in the OTT platform, the user received video data, the user device performance data and the network transmission performance data are obtained, and the feature extraction of the original video data and the user received video data is performed to obtain the original video features and the user received video features, which are specifically:

[0014] Extracting a preset number of original files of the target video from the content library of the OTT platform to obtain original video data of the target video;

[0015] Collecting user received video data of the target video in real time through the user terminal, and recording user device performance data and network transmission performance data of the user in the process of receiving the target video, wherein the user device performance data includes processor utilization and memory occupancy, and the network transmission performance data includes bandwidth, delay, and jitter parameters of the user device network connection;

[0016] Extracting the video frame data of the original video data and the video frame data of the video data received by the user based on the OpenCV video processing library, and aligning the original video data and the video frame data of the video data received by the user according to a time series;

[0017] The brightness features, contrast features, color features, texture features, and shape features of the video frame data of the original video data and the video data received by the user are extracted based on an image processing algorithm to obtain the original video features and the user received video features.

[0018] In this solution, the fingerprint algorithm is used to compare the original video features with the user received video features, determine the feature similarity index between the user received video data and the original video data, and perform quality assessment on the user received video according to the feature similarity index to obtain a first quality assessment result, which is specifically:

[0019] A fingerprint algorithm is introduced to generate an original video feature vector and a user received video feature vector from each feature item in the original video feature and the user received video feature;

[0020] Dividing the original video feature vector and the user received video feature vector into N feature vector blocks of fixed size based on a fingerprint algorithm;

[0021] For the brightness feature vector block, the contrast feature vector block, and the color feature vector block, a hash value is generated for the brightness feature vector block, the contrast feature vector block, and the color feature vector block based on the MD5 hash function to obtain a numerical feature hash code of the original video data and the video data received by the user;

[0022] For the texture feature vector block and the shape feature vector block, a discrete cosine transform is applied to each texture feature vector block and the shape feature vector block based on a perceptual hash function to generate a frequency domain feature of each feature vector block represented in the frequency domain, a low-frequency component is selected as a feature representation of the texture feature vector block and the shape feature vector block according to the frequency domain feature, an average value of the low-frequency component is calculated, each value in the frequency domain feature is compared with the average value, a binary hash code is generated, and an image-type feature hash code of the original video data and the video data received by the user is obtained;

[0023] The hash codes of each feature vector block are connected in series to generate the original video feature fingerprint and the user received video feature fingerprint;

[0024] Compare the original video feature fingerprint with the user received video feature fingerprint by hash value, calculate the Hamming distance between the hash values, evaluate the similarity score between the user received video feature and each feature of the original video feature according to the Hamming distance, perform weighted average on the similarity scores of each feature, and obtain the feature similarity index between the user received video data and the original video data;

[0025] The distortion degree of the video received by the user is evaluated according to the feature similarity index to obtain a first quality evaluation result.

[0026] In this solution, the motion estimation is performed on the video data received by the user to determine the stability and motion smoothness of the video received by the user, and the secondary quality evaluation is performed on the video received by the user according to the stability and motion smoothness to obtain a second quality evaluation result, which is specifically:

[0027] The video frame data of the video data received by the user is converted into a grayscale image, and feature point detection is performed on each frame of the video frame data based on the SIFT feature point detection algorithm to obtain feature points of the video frame data;

[0028] Tracking feature points between adjacent frames of video data received by the user based on the optical flow method, and calculating motion vectors of the feature points;

[0029] Calculating the standard deviation of the motion vectors of the feature points between adjacent frames according to the transport vector, evaluating the stability between the video frames of the video data received by the user according to the standard deviation, and obtaining the stability evaluation result;

[0030] Calculate the displacement and speed between feature points of adjacent frames according to the motion vector, and draw the motion trajectory of the feature points according to the displacement and speed between feature points of adjacent frames;

[0031] Determining the motion continuity of the feature point according to the motion trajectory, and evaluating the motion smoothness of the video data received by the user according to the motion continuity to obtain a motion smoothness evaluation result;

[0032] A secondary quality evaluation is performed on the video received by the user according to the stability evaluation result and the motion smoothness evaluation result to obtain a second quality evaluation result.

[0033] In this solution, the comprehensive quality evaluation of the video received by the user in the OTT platform is performed according to the first quality evaluation result and the second quality evaluation result to obtain the comprehensive quality evaluation result of the video received by the user, which is specifically:

[0034] Performing a weight analysis on the first quality assessment result and the second quality assessment result based on an information entropy weight method to determine weight information of the first quality assessment result and the second quality assessment result for a comprehensive quality assessment;

[0035] The first quality assessment result and the second quality assessment result are weighted averaged according to the weight information to obtain a comprehensive quality assessment result of the video received by the user.

[0036] In this solution, the user equipment performance data and the network transmission performance data are subjected to correlation analysis with the comprehensive quality evaluation result to determine the impact of the user equipment performance and the network transmission performance on the quality of the video received by the user, and to obtain the impact data, specifically:

[0037] Obtaining timestamp information of the video data received by the user, the performance data of the user's device, and the performance data of the network transmission, and performing a time alignment operation on the comprehensive quality evaluation result corresponding to the video received by the user with the performance data of the user's device and the performance data of the network transmission according to the timestamp information, to generate a time corresponding data sequence table;

[0038] Calculate the Pearson correlation coefficient between each data feature of the user performance data and the network transmission performance data feature and the comprehensive quality evaluation result according to the time corresponding data sequence table;

[0039] According to the Pearson correlation coefficient between each data feature and the comprehensive quality evaluation result, the influencing factors and influencing weights of the user equipment performance data and the network transmission performance data on the comprehensive quality of the user received video are determined to obtain the influencing data.

[0040] In this solution, the user pushed video in the OTT platform is cached in advance according to the impact data to obtain a solution for optimizing the quality of the video received by the user, which is specifically:

[0041] Obtaining minimum quality standard data of the video received by the user, comparing the comprehensive quality evaluation result with the minimum quality standard data, and determining whether the comprehensive quality of the video received by the user meets the minimum quality standard;

[0042] If not satisfied, monitor the performance changes of the factors affecting the comprehensive quality of the video received by the user in each preset time period according to the influencing data according to the preset time period to obtain a periodic performance fluctuation curve, and perform comprehensive performance calculation on the periodic performance fluctuation curve based on the influence weights of the influencing factors to obtain a periodic comprehensive performance change curve;

[0043] Analyzing the periodic comprehensive performance change law of the video received by the user according to the periodic comprehensive performance change curve, and determining the high-performance time period for the video received by the user according to the periodic comprehensive performance change law;

[0044] Obtain a video list pushed by the OTT platform to the user, perform a video caching operation on the pushed video list in the user device according to the high-performance time period, and obtain a solution for optimizing the quality of the video received by the user.

[0045] The second aspect of the present invention further provides an OTT platform content quality assessment system based on data analysis, the system comprising: a memory, a processor, the memory comprising an OTT platform content quality assessment method program based on data analysis, and when the OTT platform content quality assessment method program based on data analysis is executed by the processor, the following steps are implemented:

[0046] Obtaining original video data of a target video in an OTT platform, user received video data, user device performance data, and network transmission performance data, and performing feature extraction on the original video data and user received video data to obtain original video features and user received video features;

[0047] Comparing the original video features with the user received video features based on a fingerprint algorithm, determining a feature similarity index between the user received video data and the original video data, and performing a quality assessment on the user received video according to the feature similarity index to obtain a first quality assessment result;

[0048] Performing motion estimation on the video data received by the user, determining the stability and motion smoothness of the video received by the user, and performing a secondary quality assessment on the video received by the user according to the stability and motion smoothness to obtain a second quality assessment result;

[0049] Performing a comprehensive quality assessment on the video received by the user on the OTT platform according to the first quality assessment result and the second quality assessment result to obtain a comprehensive quality assessment result of the video received by the user;

[0050] Performing correlation analysis on the user equipment performance data and the network transmission performance data and the comprehensive quality evaluation result to determine the impact of the user equipment performance and the network transmission performance on the quality of the video received by the user, and obtaining impact data;

[0051] According to the impact data, the user-pushed videos in the OTT platform are cached in advance to obtain a solution for optimizing the quality of the videos received by the users.

[0052] In this solution, the user equipment performance data and the network transmission performance data are subjected to correlation analysis with the comprehensive quality evaluation result to determine the impact of the user equipment performance and the network transmission performance on the quality of the video received by the user, and to obtain the impact data, specifically:

[0053] Obtaining timestamp information of the video data received by the user, the performance data of the user's device, and the performance data of the network transmission, and performing a time alignment operation on the comprehensive quality evaluation result corresponding to the video received by the user with the performance data of the user's device and the performance data of the network transmission according to the timestamp information, to generate a time corresponding data sequence table;

[0054] Calculate the Pearson correlation coefficient between each data feature of the user performance data and the network transmission performance data feature and the comprehensive quality evaluation result according to the time corresponding data sequence table;

[0055] According to the Pearson correlation coefficient between each data feature and the comprehensive quality evaluation result, the influencing factors and influencing weights of the user equipment performance data and the network transmission performance data on the comprehensive quality of the user received video are determined to obtain the influencing data.

[0056] In this solution, the user pushed video in the OTT platform is cached in advance according to the impact data to obtain a solution for optimizing the quality of the video received by the user, which is specifically:

[0057] Obtaining minimum quality standard data of the video received by the user, comparing the comprehensive quality evaluation result with the minimum quality standard data, and determining whether the comprehensive quality of the video received by the user meets the minimum quality standard;

[0058] If not satisfied, monitor the performance changes of the factors affecting the comprehensive quality of the video received by the user in each preset time period according to the influencing data according to the preset time period to obtain a periodic performance fluctuation curve, and perform comprehensive performance calculation on the periodic performance fluctuation curve based on the influence weights of the influencing factors to obtain a periodic comprehensive performance change curve;

[0059] Analyzing the periodic comprehensive performance change law of the video received by the user according to the periodic comprehensive performance change curve, and determining the high-performance time period for the video received by the user according to the periodic comprehensive performance change law;

[0060] Obtain a video list pushed by the OTT platform to the user, perform a video caching operation on the pushed video list in the user device according to the high-performance time period, and obtain a solution for optimizing the quality of the video received by the user.

[0061] The present invention discloses a method and system for evaluating the content quality of an OTT platform based on data analysis. The original video data of the target video and the video data received by the user, the performance data of the user device and the network transmission performance data are obtained, and feature extraction is performed to obtain the original video features and the video features received by the user; the original video features and the video features received by the user are compared by a fingerprint algorithm to evaluate the quality of the video received by the user and obtain a first quality evaluation result; motion estimation is performed to evaluate the stability and motion smoothness of the video and obtain a second quality evaluation result; the above evaluation results are combined to obtain a comprehensive quality evaluation result; the impact of the device and network performance on the video quality is analyzed to obtain the impact data; and the pre-caching operation is performed according to the impact data to optimize the video quality. The present invention can accurately evaluate and optimize the quality of the video received by the user and improve the user's viewing experience. BRIEF DESCRIPTION OF THE DRAWINGS

[0062] Figure 1 A flow chart of a method for evaluating content quality of an OTT platform based on data analysis is shown in the present invention;

[0063] Figure 2 A flow chart of obtaining influence data according to the present invention is shown;

[0064] Figure 3 A flow chart of the scheme for optimizing the quality of video received by a user is shown in the present invention;

[0065] Figure 4 A block diagram of an OTT platform content quality assessment system based on data analysis of the present invention is shown. DETAILED DESCRIPTION

[0066] In order to more clearly understand the above-mentioned purpose, features and advantages of the present invention, the present invention is further described in detail below in conjunction with the accompanying drawings and specific embodiments. It should be noted that the embodiments of the present application and the features in the embodiments can be combined with each other without conflict.

[0067] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways different from those described herein. Therefore, the protection scope of the present invention is not limited to the specific embodiments disclosed below.

[0068] Figure 1 A flow chart of a method for evaluating content quality of an OTT platform based on data analysis is shown in the present invention.

[0069] like Figure 1 As shown, the first aspect of the present invention provides an OTT platform content quality assessment method based on data analysis, comprising:

[0070] S102, obtaining original video data of a target video in the OTT platform, user received video data, user device performance data, and network transmission performance data, and performing feature extraction on the original video data and user received video data to obtain original video features and user received video features;

[0071] S104, comparing the original video feature with the user received video feature based on a fingerprint algorithm, determining a feature similarity index between the user received video data and the original video data, and performing a quality assessment on the user received video according to the feature similarity index to obtain a first quality assessment result;

[0072] S106, performing motion estimation on the video data received by the user, determining the stability and motion smoothness of the video received by the user, and performing a secondary quality assessment on the video received by the user according to the stability and motion smoothness to obtain a second quality assessment result;

[0073] S108, performing a comprehensive quality assessment on the video received by the user in the OTT platform according to the first quality assessment result and the second quality assessment result, to obtain a comprehensive quality assessment result of the video received by the user;

[0074] S110, performing correlation analysis on the user equipment performance data and the network transmission performance data and the comprehensive quality evaluation result to determine the impact of the user equipment performance and the network transmission performance on the quality of the video received by the user, and obtain impact data;

[0075] S112, pre-caching the user-pushed video in the OTT platform according to the impact data, and obtaining a solution for optimizing the quality of the video received by the user.

[0076] It should be noted that, by extracting the original video data of the target video in the OTT platform and the original video features and user-received video features of the video data received by the user, and then comparing the original video features with the user-received video features based on the fingerprint algorithm, the similarity between the user-received video and the original video is determined, and a first quality assessment result is obtained; further motion estimation is performed on the user video data to determine the stability and motion smoothness of the user-received video, and a secondary quality assessment is performed on the user-received video based on the stability and motion smoothness to obtain a second quality assessment result; in the existing video quality assessment means, when there is a reference video, the video quality assessment is only performed by comparing with the reference video, and when there is no reference video, the quality of the video data itself is assessed; in the traditional reference video quality assessment method, it relies on direct comparison with the reference video. Although this method can accurately reflect the degree of distortion of the video during transmission, it cannot comprehensively evaluate the actual viewing experience of the video at the playback end, such as stability and fluency. The non-reference video quality assessment focuses on the feature analysis of the video content itself, ignoring the comparison with the original video, and it is difficult to accurately capture the quality degradation caused by transmission problems. The present invention performs dual quality assessment on the user-received video by simultaneously performing a reference video, that is, the target video and the video data itself, on the video received by the user, thereby greatly improving the accuracy of the video quality assessment, and being more in line with the reality of quality assessment of user-received video data for the OTT platform; by performing a correlation analysis on the user device performance data and the network transmission performance data and the comprehensive quality assessment result, the influence of the user device performance and the network transmission performance on the user-received video quality is determined, and then the user-pushed video in the OTT platform is cached in advance. The advance caching can reduce buffering and stuttering during playback, especially when the user device performance is poor or the network transmission is unstable, thereby improving the overall viewing experience. The advance caching can play the target video after high-quality caching, avoiding the situation where the target video reaches the user device and the user receives the video due to the influence of the device performance and the network performance. Video distortion occurs after the user receives the video.

[0077] According to an embodiment of the present invention, the original video data of the target video in the OTT platform and the user received video data, the user device performance data and the network transmission performance data are obtained, and the feature extraction of the original video data and the user received video data is performed to obtain the original video features and the user received video features, specifically:

[0078] Extracting a preset number of original files of the target video from the content library of the OTT platform to obtain original video data of the target video;

[0079] Collecting user received video data of the target video in real time through the user terminal, and recording user device performance data and network transmission performance data of the user in the process of receiving the target video, wherein the user device performance data includes processor utilization and memory occupancy, and the network transmission performance data includes bandwidth, delay, and jitter parameters of the user device network connection;

[0080] Extracting the video frame data of the original video data and the video frame data of the video data received by the user based on the OpenCV video processing library, and aligning the original video data and the video frame data of the video data received by the user according to a time series;

[0081] The brightness features, contrast features, color features, texture features, and shape features of the video frame data of the original video data and the video data received by the user are extracted based on an image processing algorithm to obtain the original video features and the user received video features.

[0082] It should be noted that OpenCV is an open source computer vision and machine learning software library. OpenCV provides hundreds of functions and tools for image and video processing, which are widely used in various computer vision tasks; the image processing algorithms include gray-level co-occurrence matrix, wavelet transform, and edge detection algorithm.

[0083] According to an embodiment of the present invention, the fingerprint algorithm is used to compare the original video features with the user received video features, determine the feature similarity index between the user received video data and the original video data, and perform quality assessment on the user received video according to the feature similarity index to obtain a first quality assessment result, which is specifically:

[0084] A fingerprint algorithm is introduced to generate an original video feature vector and a user received video feature vector from each feature item in the original video feature and the user received video feature;

[0085] Dividing the original video feature vector and the user received video feature vector into N feature vector blocks of fixed size based on a fingerprint algorithm;

[0086] For the brightness feature vector block, the contrast feature vector block, and the color feature vector block, a hash value is generated for the brightness feature vector block, the contrast feature vector block, and the color feature vector block based on the MD5 hash function to obtain a numerical feature hash code of the original video data and the video data received by the user;

[0087] For the texture feature vector block and the shape feature vector block, a discrete cosine transform is applied to each texture feature vector block and the shape feature vector block based on a perceptual hash function to generate a frequency domain feature of each feature vector block represented in the frequency domain, a low-frequency component is selected as a feature representation of the texture feature vector block and the shape feature vector block according to the frequency domain feature, an average value of the low-frequency component is calculated, each value in the frequency domain feature is compared with the average value, a binary hash code is generated, and an image-type feature hash code of the original video data and the video data received by the user is obtained;

[0088] The hash codes of each feature vector block are connected in series to generate the original video feature fingerprint and the user received video feature fingerprint;

[0089] Compare the original video feature fingerprint with the user received video feature fingerprint by hash value, calculate the Hamming distance between the hash values, evaluate the similarity score between the user received video feature and each feature of the original video feature according to the Hamming distance, perform weighted average on the similarity scores of each feature, and obtain the feature similarity index between the user received video data and the original video data;

[0090] The distortion degree of the video received by the user is evaluated according to the feature similarity index to obtain a first quality evaluation result.

[0091] It should be noted that since the brightness feature, contrast feature and color feature in the original video features are represented by numerical values, the MD5 hash function is used to generate hash values ​​for the brightness feature vector block, contrast feature vector block and color feature vector block. Since the texture feature and shape feature are image format features, the MD5 hash function cannot encode the image format features, so the image features must be converted into hash values ​​through the perceptual hash function. The perceptual hash function is a hash function that specifically converts image data into hash values. The fingerprint algorithm extracts the features of the video and generates feature vectors, so that the content of the video can be digitally represented. The feature vector of each video consists of multiple feature blocks (such as brightness, contrast, color, texture and shape), which remain relatively stable under different viewing angles and operations. In this way, even if the video is processed or compressed to a certain extent, the main features can still be retained and effectively compared; for texture and shape features, discrete cosine transform (DCT) is used to convert time domain features into frequency domain features, and low-frequency components are selected as feature representations. The low-frequency component retains the main information of the image, reduces the impact of high-frequency noise, and improves the robustness and stability of feature representation; the fingerprint algorithm converts complex video features into concise hash codes, making the feature comparison process fast and efficient, and greatly reducing the amount of data that needs to be stored and processed. The fixed length and simplified representation of the hash code can significantly reduce the consumption of computing resources; by calculating the Hamming distance to quantify the difference between the feature fingerprints of the original video and the video received by the user, the similarity between the two can be accurately evaluated. The calculation result of the Hamming distance directly reflects the degree of feature matching and provides an intuitive and quantifiable similarity index. The smaller the Hamming distance, the higher the similarity, and the larger the Hamming distance, the lower the similarity; the frequency domain feature is to transform the image from the spatial domain (pixel value representation) to the frequency domain through transformation, and describe the image features in the form of frequency.

[0092] According to an embodiment of the present invention, the motion estimation is performed on the video data received by the user, the stability and motion smoothness of the video received by the user are determined, and the secondary quality evaluation is performed on the video received by the user according to the stability and motion smoothness to obtain a second quality evaluation result, which is specifically:

[0093] The video frame data of the video data received by the user is converted into a grayscale image, and feature point detection is performed on each frame of the video frame data based on the SIFT feature point detection algorithm to obtain feature points of the video frame data;

[0094] Tracking feature points between adjacent frames of video data received by the user based on the optical flow method, and calculating motion vectors of the feature points;

[0095] Calculating the standard deviation of the motion vectors of the feature points between adjacent frames according to the transport vector, evaluating the stability between the video frames of the video data received by the user according to the standard deviation, and obtaining the stability evaluation result;

[0096] Calculate the displacement and speed between feature points of adjacent frames according to the motion vector, and draw the motion trajectory of the feature points according to the displacement and speed between feature points of adjacent frames;

[0097] Determining the motion continuity of the feature point according to the motion trajectory, and evaluating the motion smoothness of the video data received by the user according to the motion continuity to obtain a motion smoothness evaluation result;

[0098] A secondary quality evaluation is performed on the video received by the user according to the stability evaluation result and the motion smoothness evaluation result to obtain a second quality evaluation result.

[0099] It should be noted that by converting the data of the video received by the user into a grayscale image and using the SIFT (Scale-Invariant Feature Transform) feature point detection algorithm to detect the video feature points of each frame, the important feature points in the video can be accurately captured. Based on the optical flow method, the motion of these feature points is tracked, the motion vector between adjacent frames is calculated, and the standard deviation of the motion vector is further calculated to evaluate the stability between video frames. The stability evaluation result can reflect whether the video has obvious jitter or instability; by tracking the motion trajectory of the feature points and calculating the displacement and speed between adjacent frames, the motion smoothness of the video can be accurately evaluated. The continuity of the motion trajectory reflects the smoothness and fluency of the moving object in the video. In this way, it is possible to identify whether there is discontinuous or sudden motion in the video, thereby affecting the viewing experience; the present invention not only performs a preliminary quality evaluation through the feature similarity index, but also performs a secondary quality evaluation on the stability and motion smoothness of the video to ensure the comprehensiveness and accuracy of the video quality evaluation; the feature points include corner points, edge points, and spots; the motion vector describes the displacement and speed of the feature points between video frames.

[0100] According to an embodiment of the present invention, the comprehensive quality evaluation of the video received by the user in the OTT platform is performed according to the first quality evaluation result and the second quality evaluation result to obtain the comprehensive quality evaluation result of the video received by the user, which is specifically:

[0101] Performing a weight analysis on the first quality assessment result and the second quality assessment result based on an information entropy weight method to determine weight information of the first quality assessment result and the second quality assessment result for a comprehensive quality assessment;

[0102] The first quality assessment result and the second quality assessment result are weighted averaged according to the weight information to obtain a comprehensive quality assessment result of the video received by the user.

[0103] Figure 2 A flow chart of obtaining influence data according to the present invention is shown.

[0104] According to an embodiment of the present invention, the user equipment performance data and the network transmission performance data are subjected to correlation analysis with the comprehensive quality evaluation result to determine the impact of the user equipment performance and the network transmission performance on the quality of the video received by the user, and to obtain the impact data, specifically:

[0105] S202, obtaining timestamp information of the video data received by the user, the performance data of the user device, and the network transmission performance data, and performing a time alignment operation on the comprehensive quality evaluation result corresponding to the video received by the user with the performance data of the user device and the network transmission performance data according to the timestamp information, to generate a time corresponding data sequence table;

[0106] S204, calculating the Pearson correlation coefficient between each data feature of the user performance data and the network transmission performance data feature and the comprehensive quality evaluation result according to the time corresponding data sequence table;

[0107] S206, determining the influencing factors and influencing weights of the user equipment performance data and the network transmission performance data on the comprehensive quality of the user received video according to the Pearson correlation coefficient between each data feature and the comprehensive quality evaluation result, and obtaining influencing data.

[0108] It should be noted that in the video streaming transmission of OTT platforms, video quality is affected by many factors, including the performance of user devices and the quality of network transmission. In order to determine the specific impact of device performance and network transmission on video quality, it is necessary to correlate these data with the video quality assessment results; the Pearson correlation coefficient is a statistic used to measure the strength and direction of the linear relationship between two variables. In this step, by calculating the Pearson correlation coefficient between the device performance data and network transmission performance data characteristics and the video quality assessment results, the specific impact of each factor on video quality can be determined. This statistical analysis helps understand which factors affect the video quality to what extent, and whether these impacts are positive or negative; the larger the Pearson correlation coefficient, the stronger the positive linear correlation between the two variables. When the correlation coefficient is close to 1, the two variables are highly positively correlated, that is, when one variable increases, the other variable also increases. When the correlation coefficient is close to -1, the two variables are highly negatively correlated, that is, when one variable increases, the other variable decreases. When the correlation coefficient is close to 0, there is almost no linear relationship between the two variables.

[0109] Figure 3 The flowchart of the present invention for obtaining a solution for optimizing the quality of video received by a user is shown.

[0110] According to an embodiment of the present invention, the user pushed video in the OTT platform is cached in advance according to the impact data to obtain a solution for optimizing the quality of the video received by the user, which is specifically:

[0111] S302, obtaining minimum quality standard data of the video received by the user, comparing the comprehensive quality evaluation result with the minimum quality standard data, and determining whether the comprehensive quality of the video received by the user meets the minimum quality standard;

[0112] S304: If not satisfied, monitor the performance change of the factors affecting the comprehensive quality of the video received by the user in each preset time period according to the influencing data to obtain a periodic performance fluctuation curve, and perform comprehensive performance calculation on the periodic performance fluctuation curve based on the influence weights of the influencing factors to obtain a periodic comprehensive performance change curve;

[0113] S306, analyzing the periodic comprehensive performance change rule of the video received by the user according to the periodic comprehensive performance change curve, and determining the high-performance time period for the video received by the user according to the periodic comprehensive performance change rule;

[0114] S308, obtaining a video list pushed by the OTT platform to the user, performing a video caching operation on the pushed video list in the user device according to the high-performance time period, and obtaining a solution for optimizing the quality of the video received by the user.

[0115] It should be noted that caching video files in advance can reduce the need for real-time buffering when users are watching, reduce playback interruptions or buffering caused by network fluctuations or delays, select the time period when the user's device and network performance are optimal for caching, ensure the video quality during playback is at the best level, and reduce the degradation of video quality due to fluctuations in device or network performance; by caching during network peak hours or time periods with low bandwidth load, avoid pressure on network bandwidth during high-traffic periods, improve overall bandwidth utilization efficiency, and avoid situations in which videos pushed by OTT platforms cannot be played normally due to network quality problems in some public places, or the quality of received videos is degraded due to device performance and network performance.

[0116] According to an embodiment of the present invention, it also includes:

[0117] Obtaining background record data of the user device, and periodically obtaining viewing behavior data of the user device for the video pushed by the OTT platform within a preset time period according to the background record data, wherein the viewing behavior data includes video viewing time, video viewing type, number of views, video skipping, pausing, and fast forwarding;

[0118] Based on the K-means clustering algorithm, the viewing behavior data of each period is clustered, and the data with similar viewing behaviors are classified into one category to obtain the clustering results of the viewing behavior data of each period;

[0119] evaluating the viewing interest level of each type of clustering result according to the viewing behavior data in the viewing behavior data clustering result to obtain a viewing interest evaluation result, and assigning an interest level label to each type of clustering result according to the viewing interest evaluation result;

[0120] Obtaining the video viewing types included in each clustering result, counting the number of video viewing types included in each clustering result, determining the distribution of different video viewing types in each interest tag according to the interest level tag of each clustering result and the number of video viewing types included in each clustering result, and scoring the interest of each video viewing type according to the distribution to obtain the user's interest value for different video types in each cycle;

[0121] Sorting the interest values ​​of each video type in each period within a preset time period according to time to obtain video type interest value change data within the preset time period;

[0122] A video type interest value change prediction model is constructed based on the LSTM algorithm, and the video type interest value change data is divided into training data and prediction basic data according to a preset ratio.

[0123] Importing the training data into the video type interest value change prediction model for learning and training, and importing the prediction basic data into the trained video type interest value change prediction model, predicting the interest value change of each video type of the user in a preset time period in the future, and obtaining the interest value change prediction results of each video type;

[0124] According to the prediction results of the interest value changes of each video type, the OTT platform adjusts the pushed videos to users within a future preset time period.

[0125] It should be noted that, since the user's interest in the video types pushed by the OTT platform will change over time, the user's viewing behavior data in each time period within the preset time period is clustered through a clustering algorithm, and an interest level label is assigned to each type of clustering result, so as to determine the root of the cluster of each type in different interest level labels, and the user's viewing interest value for different video types in each time period can be determined; a video type interest value change prediction model is constructed through LSTM (Long Short-Term Memory Network), and the user's interest value for different video types in the future preset time period is predicted based on the video type interest value change data in the preset time period, and then the OTT platform adjusts the pushed videos to the user in the future preset time period, and prioritizes the push of video types with high interest values, and determines the push ratio of different types of videos based on the interest value of each video type; by accurately grasping the user's interest and recommending the content that the user is more interested in, it can effectively increase the user's viewing time and improve the platform traffic. Through the interest value change prediction model, the platform can understand the changing trend of user interests in advance, thereby optimizing video push and reducing user loss or dissatisfaction caused by improper recommendation.

[0126] According to an embodiment of the present invention, it also includes:

[0127] Performing dimension reduction on the user equipment performance data and the network transmission performance data based on the PCA algorithm, and extracting user performance characteristics from the dimension-reduced user equipment performance data and the network transmission performance data to obtain user performance characteristic data;

[0128] Performing a clustering operation on the user performance characteristic data based on a clustering algorithm to obtain a user performance characteristic clustering result;

[0129] According to the user performance characteristic clustering result, the users in the OTT platform are divided into performance levels according to the user performance characteristics, and the user performance characteristic level division result is obtained, wherein the user performance characteristic level division result includes a high performance characteristic level, a medium performance characteristic level, and a low performance characteristic level;

[0130] Obtain the evaluation cycle and the amount of video data for evaluating the quality of videos received by the OTT platform for the users, adjust the video quality evaluation cycle and the amount of video data for the users of each user performance characteristic level according to the user performance characteristic level classification result, when the user is at a high performance characteristic level, extend the evaluation cycle for the quality evaluation of the videos received by the users and reduce the amount of video data for evaluation, when the user is at a medium performance characteristic level, maintain the evaluation cycle and the amount of video data for the quality evaluation of the videos received by the current users, when the user is at a low performance characteristic level, shorten the evaluation cycle for the quality evaluation of the videos received by the users and increase the amount of video data for evaluation, and obtain the real-time quality evaluation strategy for the videos received by the users by the OTT platform, wherein the real-time quality evaluation strategy includes adjusting the evaluation cycle and the amount of video data for evaluation in the primary quality evaluation and the secondary quality evaluation of the videos received by the users;

[0131] The quality of the videos received by users in the OTT platform is evaluated according to the real-time quality evaluation strategy, and the user performance characteristic level of each user in the OTT platform is updated in real time according to the evaluation result, so as to obtain a dynamic user received video quality evaluation scheme of the OTT platform.

[0132] It should be noted that the data processing resources of the OTT platform are also limited. When evaluating the quality of the video received by the users of the OTT platform, since video quality evaluation of a large user group will greatly consume the data processing resources of the OTT platform, resulting in a decrease in the data processing capacity of the OTT platform, the performance characteristics of the users are extracted and then clustered to divide the users into high, medium and low performance characteristic levels. Users with high performance characteristics have a high receiving video performance and have a low impact on the change of the received video quality. Therefore, the evaluation of the received video quality of the users with high performance characteristics can be extended, and the amount of video data for quality evaluation can be reduced. Users with low performance characteristics have a large impact on the change of the received video quality due to their low receiving video performance. Therefore, it is necessary to shorten the video quality evaluation cycle to avoid the video quality received by the users being too low, so as to form a real-time quality evaluation strategy for the OTT platform to evaluate the quality of the videos received by the users; finally, the quality of the videos received by the users is evaluated according to the real-time quality evaluation strategy, and the performance characteristic level of the users is adjusted to form a dynamic user receiving video quality evaluation scheme; according to the different levels of user performance characteristics, the evaluation resources are reasonably allocated to avoid unnecessary computing overhead, reduce the data processing pressure of the OTT platform, improve the platform stability, and reduce the resource consumption of the OTT platform for large-scale user video quality evaluation.

[0133] Figure 4 A block diagram of an OTT platform content quality assessment system based on data analysis of the present invention is shown.

[0134] The second aspect of the present invention further provides an OTT platform content quality assessment system 4 based on data analysis, the system comprising: a memory 41, a processor 42, the memory comprising an OTT platform content quality assessment method program based on data analysis, and when the OTT platform content quality assessment method program based on data analysis is executed by the processor, the following steps are implemented:

[0135] Obtaining original video data of a target video in an OTT platform, user received video data, user device performance data, and network transmission performance data, and performing feature extraction on the original video data and user received video data to obtain original video features and user received video features;

[0136] Comparing the original video features with the user received video features based on a fingerprint algorithm, determining a feature similarity index between the user received video data and the original video data, and performing a quality assessment on the user received video according to the feature similarity index to obtain a first quality assessment result;

[0137] Performing motion estimation on the video data received by the user, determining the stability and motion smoothness of the video received by the user, and performing a secondary quality assessment on the video received by the user according to the stability and motion smoothness to obtain a second quality assessment result;

[0138] Performing a comprehensive quality assessment on the video received by the user on the OTT platform according to the first quality assessment result and the second quality assessment result to obtain a comprehensive quality assessment result of the video received by the user;

[0139] Performing correlation analysis on the user equipment performance data and the network transmission performance data and the comprehensive quality evaluation result to determine the impact of the user equipment performance and the network transmission performance on the quality of the video received by the user, and obtaining impact data;

[0140] According to the impact data, the user-pushed videos in the OTT platform are cached in advance to obtain a solution for optimizing the quality of the videos received by the users.

[0141] The present invention discloses a method and system for evaluating the content quality of an OTT platform based on data analysis. The original video data of the target video and the video data received by the user, the performance data of the user device and the network transmission performance data are obtained, and feature extraction is performed to obtain the original video features and the video features received by the user; the original video features and the video features received by the user are compared by a fingerprint algorithm to evaluate the quality of the video received by the user and obtain a first quality evaluation result; motion estimation is performed to evaluate the stability and motion smoothness of the video and obtain a second quality evaluation result; the above evaluation results are combined to obtain a comprehensive quality evaluation result; the impact of the device and network performance on the video quality is analyzed to obtain the impact data; and the pre-caching operation is performed according to the impact data to optimize the video quality. The present invention can accurately evaluate and optimize the quality of the video received by the user and improve the user's viewing experience.

[0142] In the several embodiments provided in the present application, it should be understood that the disclosed devices and methods can be implemented in other ways. The device embodiments described above are only schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation, such as: multiple units or components can be combined, or can be integrated into another system, or some features can be ignored or not executed. In addition, the coupling, direct coupling, or communication connection between the components shown or discussed can be through some interfaces, and the indirect coupling or communication connection of the devices or units can be electrical, mechanical or other forms.

[0143] The units described above as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units; they may be located in one place or distributed on multiple network units; some or all of the units may be selected according to actual needs to achieve the purpose of the present embodiment.

[0144] In addition, all functional units in the embodiments of the present invention may be integrated into one processing unit, or each unit may be separately used as a unit, or two or more units may be integrated into one unit; the above-mentioned integrated units may be implemented in the form of hardware or in the form of hardware plus software functional units.

[0145] A person of ordinary skill in the art can understand that: all or part of the steps of implementing the above method embodiment can be completed by hardware related to program instructions, and the aforementioned program can be stored in a computer-readable storage medium. When the program is executed, it executes the steps of the above method embodiment; and the aforementioned storage medium includes: a mobile storage device, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, and other media that can store program codes.

[0146] Alternatively, if the above-mentioned integrated unit of the present invention is implemented in the form of a software function module and sold or used as an independent product, it can also be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the embodiment of the present invention can be essentially or partly reflected in the form of a software product that contributes to the prior art. The computer software product is stored in a storage medium and includes several instructions for a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the methods described in each embodiment of the present invention. The aforementioned storage medium includes: various media that can store program codes, such as mobile storage devices, ROM, RAM, magnetic disks or optical disks.

[0147] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art who is familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.

Claims

1. A method for evaluating the content quality of an OTT platform based on data analysis, characterized in that: The following steps are involved: Obtaining original video data of a target video in an OTT platform, user received video data, user device performance data, and network transmission performance data, and performing feature extraction on the original video data and user received video data to obtain original video features and user received video features; Comparing the original video features with the user received video features based on a fingerprint algorithm, determining a feature similarity index between the user received video data and the original video data, and performing a quality assessment on the user received video according to the feature similarity index to obtain a first quality assessment result; Performing motion estimation on the video data received by the user, determining the stability and motion smoothness of the video received by the user, and performing a secondary quality assessment on the video received by the user according to the stability and motion smoothness to obtain a second quality assessment result; Performing a comprehensive quality assessment on the video received by the user on the OTT platform according to the first quality assessment result and the second quality assessment result to obtain a comprehensive quality assessment result of the video received by the user; Performing correlation analysis on the user equipment performance data and the network transmission performance data and the comprehensive quality evaluation result to determine the impact of the user equipment performance and the network transmission performance on the quality of the video received by the user, and obtaining impact data; According to the impact data, the user-pushed videos in the OTT platform are cached in advance to obtain a solution for optimizing the quality of the videos received by the users.

2. According to claim 1, a method for evaluating the content quality of an OTT platform based on data analysis is characterized in that: The original video data of the target video in the OTT platform, the video data received by the user, the performance data of the user device, and the network transmission performance data are obtained, and the feature extraction is performed on the original video data and the video data received by the user to obtain the original video features and the video features received by the user, specifically: Extracting a preset number of original files of the target video from the content library of the OTT platform to obtain original video data of the target video; Collecting user received video data of the target video in real time through the user terminal, and recording user device performance data and network transmission performance data of the user in the process of receiving the target video, wherein the user device performance data includes processor utilization and memory occupancy, and the network transmission performance data includes bandwidth, delay, and jitter parameters of the user device network connection; Extracting the video frame data of the original video data and the video frame data of the video data received by the user based on the OpenCV video processing library, and aligning the original video data and the video frame data of the video data received by the user according to a time series; The brightness features, contrast features, color features, texture features, and shape features of the video frame data of the original video data and the video data received by the user are extracted based on an image processing algorithm to obtain the original video features and the user received video features.

3. According to claim 1, a method for evaluating the content quality of an OTT platform based on data analysis is characterized in that: The fingerprint algorithm is used to compare the original video features with the user received video features, determine the feature similarity index between the user received video data and the original video data, and perform quality assessment on the user received video according to the feature similarity index to obtain a first quality assessment result, which is specifically: A fingerprint algorithm is introduced to generate an original video feature vector and a user received video feature vector from each feature item in the original video feature and the user received video feature; Dividing the original video feature vector and the user received video feature vector into N feature vector blocks of fixed size based on a fingerprint algorithm; For the brightness feature vector block, the contrast feature vector block, and the color feature vector block, a hash value is generated for the brightness feature vector block, the contrast feature vector block, and the color feature vector block based on the MD5 hash function to obtain a numerical feature hash code of the original video data and the video data received by the user; For the texture feature vector block and the shape feature vector block, a discrete cosine transform is applied to each texture feature vector block and the shape feature vector block based on a perceptual hash function to generate a frequency domain feature of each feature vector block represented in the frequency domain, a low-frequency component is selected as a feature representation of the texture feature vector block and the shape feature vector block according to the frequency domain feature, an average value of the low-frequency component is calculated, each value in the frequency domain feature is compared with the average value, a binary hash code is generated, and an image-type feature hash code of the original video data and the video data received by the user is obtained; The hash codes of each feature vector block are connected in series to generate the original video feature fingerprint and the user received video feature fingerprint; Compare the original video feature fingerprint with the user received video feature fingerprint by hash value, calculate the Hamming distance between the hash values, evaluate the similarity score between the user received video feature and each feature of the original video feature according to the Hamming distance, perform weighted average on the similarity scores of each feature, and obtain the feature similarity index between the user received video data and the original video data; The distortion degree of the video received by the user is evaluated according to the feature similarity index to obtain a first quality evaluation result.

4. The method for evaluating the content quality of an OTT platform based on data analysis according to claim 1, characterized in that: The motion estimation is performed on the video data received by the user, the stability and motion smoothness of the video received by the user are determined, and a secondary quality evaluation is performed on the video received by the user according to the stability and motion smoothness to obtain a second quality evaluation result, which is specifically: The video frame data of the video data received by the user is converted into a grayscale image, and feature point detection is performed on each frame of the video frame data based on the SIFT feature point detection algorithm to obtain feature points of the video frame data; Tracking feature points between adjacent frames of video data received by the user based on the optical flow method, and calculating motion vectors of the feature points; Calculating a standard deviation of a motion vector of a feature point between adjacent frames according to the motion vector, evaluating the stability between video frames of video data received by a user according to the standard deviation, and obtaining a stability evaluation result; Calculate the displacement and speed between feature points of adjacent frames according to the motion vector, and draw the motion trajectory of the feature points according to the displacement and speed between feature points of adjacent frames; Determining the motion continuity of the feature point according to the motion trajectory, and evaluating the motion smoothness of the video data received by the user according to the motion continuity to obtain a motion smoothness evaluation result; A secondary quality evaluation is performed on the video received by the user according to the stability evaluation result and the motion smoothness evaluation result to obtain a second quality evaluation result.

5. The method for evaluating the content quality of an OTT platform based on data analysis according to claim 1, characterized in that: The comprehensive quality evaluation of the video received by the user in the OTT platform is performed according to the first quality evaluation result and the second quality evaluation result to obtain the comprehensive quality evaluation result of the video received by the user, specifically: Performing a weight analysis on the first quality assessment result and the second quality assessment result based on an information entropy weight method to determine weight information of the first quality assessment result and the second quality assessment result for a comprehensive quality assessment; The first quality assessment result and the second quality assessment result are weighted averaged according to the weight information to obtain a comprehensive quality assessment result of the video received by the user.

6. The method for evaluating the content quality of an OTT platform based on data analysis according to claim 1, characterized in that: The correlation analysis of the user equipment performance data and the network transmission performance data with the comprehensive quality evaluation result is performed to determine the impact of the user equipment performance and the network transmission performance on the quality of the video received by the user, and to obtain the impact data, specifically: Obtaining timestamp information of the video data received by the user, the performance data of the user's device, and the performance data of the network transmission, and performing a time alignment operation on the comprehensive quality evaluation result corresponding to the video received by the user with the performance data of the user's device and the performance data of the network transmission according to the timestamp information, to generate a time corresponding data sequence table; Calculate the Pearson correlation coefficient between each data feature of the user performance data and the network transmission performance data feature and the comprehensive quality evaluation result according to the time corresponding data sequence table; According to the Pearson correlation coefficient between each data feature and the comprehensive quality evaluation result, the influencing factors and influencing weights of the user equipment performance data and the network transmission performance data on the comprehensive quality of the user received video are determined to obtain the influencing data.

7. The method for evaluating the content quality of an OTT platform based on data analysis according to claim 6, characterized in that: The user-pushed video in the OTT platform is cached in advance according to the impact data to obtain a solution for optimizing the quality of the video received by the user, specifically: Obtaining minimum quality standard data of the video received by the user, comparing the comprehensive quality evaluation result with the minimum quality standard data, and determining whether the comprehensive quality of the video received by the user meets the minimum quality standard; If not satisfied, monitor the performance changes of the factors affecting the comprehensive quality of the video received by the user in each preset time period according to the influencing data according to the preset time period to obtain a periodic performance fluctuation curve, and perform comprehensive performance calculation on the periodic performance fluctuation curve based on the influence weights of the influencing factors to obtain a periodic comprehensive performance change curve; Analyzing the periodic comprehensive performance change law of the video received by the user according to the periodic comprehensive performance change curve, and determining the high-performance time period for the video received by the user according to the periodic comprehensive performance change law; Obtain a video list pushed by the OTT platform to the user, perform a video caching operation on the pushed video list in the user device according to the high-performance time period, and obtain a solution for optimizing the quality of the video received by the user.

8. An OTT platform content quality assessment system based on data analysis, characterized in that: The OTT platform content quality assessment system based on data analysis includes a storage device and a processor. The storage device includes an OTT platform content quality assessment method program based on data analysis. When the OTT platform content quality assessment method program based on data analysis is executed by the processor, the following steps are implemented: Obtaining original video data of a target video in an OTT platform, user received video data, user device performance data, and network transmission performance data, and performing feature extraction on the original video data and user received video data to obtain original video features and user received video features; Comparing the original video features with the user received video features based on a fingerprint algorithm, determining a feature similarity index between the user received video data and the original video data, and performing a quality assessment on the user received video according to the feature similarity index to obtain a first quality assessment result; Performing motion estimation on the video data received by the user, determining the stability and motion smoothness of the video received by the user, and performing a secondary quality assessment on the video received by the user according to the stability and motion smoothness to obtain a second quality assessment result; Performing a comprehensive quality assessment on the video received by the user on the OTT platform according to the first quality assessment result and the second quality assessment result to obtain a comprehensive quality assessment result of the video received by the user; Performing correlation analysis on the user equipment performance data and the network transmission performance data and the comprehensive quality evaluation result to determine the impact of the user equipment performance and the network transmission performance on the quality of the video received by the user, and obtaining impact data; According to the impact data, the user-pushed videos in the OTT platform are cached in advance to obtain a solution for optimizing the quality of the videos received by the users.

9. The OTT platform content quality assessment system based on data analysis according to claim 8, characterized in that: The correlation analysis of the user equipment performance data and the network transmission performance data with the comprehensive quality evaluation result is performed to determine the impact of the user equipment performance and the network transmission performance on the quality of the video received by the user, and to obtain the impact data, specifically: Obtaining timestamp information of the video data received by the user, the performance data of the user's device, and the performance data of the network transmission, and performing a time alignment operation on the comprehensive quality evaluation result corresponding to the video received by the user with the performance data of the user's device and the performance data of the network transmission according to the timestamp information, to generate a time corresponding data sequence table; Calculate the Pearson correlation coefficient between each data feature of the user performance data and the network transmission performance data feature and the comprehensive quality evaluation result according to the time corresponding data sequence table; According to the Pearson correlation coefficient between each data feature and the comprehensive quality evaluation result, the influencing factors and influencing weights of the user equipment performance data and the network transmission performance data on the comprehensive quality of the user received video are determined to obtain the influencing data.

10. The OTT platform content quality assessment system based on data analysis according to claim 9, characterized in that: The user-pushed video in the OTT platform is cached in advance according to the impact data to obtain a solution for optimizing the quality of the video received by the user, specifically: Obtaining minimum quality standard data of the video received by the user, comparing the comprehensive quality evaluation result with the minimum quality standard data, and determining whether the comprehensive quality of the video received by the user meets the minimum quality standard; If not satisfied, monitor the performance changes of the factors affecting the comprehensive quality of the video received by the user in each preset time period according to the influencing data according to the preset time period to obtain a periodic performance fluctuation curve, and perform comprehensive performance calculation on the periodic performance fluctuation curve based on the influence weights of the influencing factors to obtain a periodic comprehensive performance change curve; Analyzing the periodic comprehensive performance change law of the video received by the user according to the periodic comprehensive performance change curve, and determining the high-performance time period for the video received by the user according to the periodic comprehensive performance change law; Obtain a video list pushed by the OTT platform to the user, perform a video caching operation on the pushed video list in the user device according to the high-performance time period, and obtain a solution for optimizing the quality of the video received by the user.

Images