A cloud video playback processing method and system based on pupil position recognition

Abstract

This application provides a cloud video playback processing method and system based on pupil position recognition. It can perform optimization processing of first cloud video playback data using at least one sample cloud video playback data. Since the sample cloud video playback data includes key descriptive content of the first cloud video playback data, the obtained optimized cloud video playback data has improved accuracy and reliability compared to the first cloud video playback data. Even if the first cloud video playback data is of poor quality, accurate optimized cloud video playback data can still be obtained by optimizing the sample cloud video playback data. In other words, this application can conveniently perform cloud video playback data optimization based on several sample cloud video playback data, thereby ensuring differentiated video playback upgrade processing based on the optimized cloud video playback data and improving the viewing quality.

Description

Technical Field

[0001] This application relates to the field of data processing technology, and more specifically, to a cloud video playback processing method and system based on pupil position recognition. Background Technology

[0002] Cloud-based video streaming can be understood as "Internet + cloud-based video streaming," based on big data analysis of internet platforms, terminals, user viewing interests, and user needs. In practice, when playing cloud-based videos, the viewing experience may vary depending on the user's seating position. Therefore, a technical solution is urgently needed to address these issues. Summary of the Invention

[0003] To address the technical problems existing in related technologies, this application provides a cloud video playback processing method and system based on pupil position recognition.

[0004] In a first aspect, a cloud video playback processing method based on pupil position recognition is provided. The method includes at least: determining first cloud video playback data; determining at least one sample cloud video playback data of the first cloud video playback data, wherein the sample cloud video playback data covers sample data of reference pupil positions in the first cloud video playback data; and performing targeted optimization on the first cloud video playback data using at least one sample cloud video playback data of the first cloud video playback data to obtain optimized cloud video playback data.

[0005] In one independently implemented embodiment, determining at least one sample cloud video playback data of the first cloud video playback data includes: determining spatial positioning data of the first cloud video playback data, the spatial positioning data covering X types of identification tags for a reference pupil position in the first cloud video playback data; determining sample cloud video playback data associated with at least one reference indicator for the reference pupil position through the spatial positioning data of the first cloud video playback data; wherein, X is a positive number greater than or equal to 1.

[0006] In one independently implemented embodiment, the targeted optimization of the first cloud video playback data using at least one sample of the first cloud video playback data to obtain optimized cloud video playback data includes: optimizing the at least one sample of the first cloud video playback data based on real-time viewing conditions of the reference pupil position in the first cloud video playback data to obtain optimized cloud video playback data bound to the sample of the first cloud video playback data in the real-time viewing conditions; selecting local cloud video playback data of at least one reference indicator associated with the reference pupil position from the optimized cloud video playback data bound to the sample of the first cloud video playback data using at least one reference indicator associated with the reference pupil position in the at least one sample of the first cloud video playback data; and obtaining the optimized cloud video playback data based on the selected local cloud video playback data and the first cloud video playback data.

[0007] In one independently implemented embodiment, obtaining the optimized cloud video playback data based on the selected local cloud video playback data and the first cloud video playback data includes: iterating through the selected local cloud video playback data the indications bound to the reference indications in the first cloud video playback data to obtain the optimized cloud video playback data, or performing feature extraction operations on the local cloud video playback data and the first cloud video playback data to obtain the optimized cloud video playback data.

[0008] In one independently implemented embodiment, the targeted optimization of the first cloud video playback data using at least one sample of the first cloud video playback data to obtain optimized cloud video playback data includes: updating the first cloud video playback data to obtain second cloud video playback data, wherein the quantitative analysis index of the second cloud video playback data exceeds the quantitative analysis index of the first cloud video playback data; optimizing the at least one sample of cloud video playback data based on the real-time viewing situation of the reference pupil position in the second cloud video playback data to obtain optimized cloud video playback data bound to the sample cloud video playback data in the real-time viewing situation; selecting local cloud video playback data of at least one reference indicator associated with the pupil position from the optimized cloud video playback data bound to the sample cloud video playback data based on at least one reference indicator associated with the pupil position in the at least one sample cloud video playback data; and obtaining the optimized cloud video playback data based on the selected local cloud video playback data and the second cloud video playback data.

[0009] In one standalone embodiment, obtaining the optimized cloud video playback data based on the selected local cloud video playback data and the second cloud video playback data includes: iterating through the selected local cloud video playback data the indications bound to the reference indications in the second cloud video playback data to obtain the optimized cloud video playback data, or performing feature extraction operations on the local cloud video playback data and the second cloud video playback data to obtain the optimized cloud video playback data.

[0010] In one standalone embodiment, the method further includes: reading tags using the optimized cloud video playback data to determine tag data associated with the pupil position.

[0011] In one standalone embodiment, the method involves a first AI thread performing the cloud video playback data update process on the first cloud video playback data to obtain the second cloud video playback data. The method further includes a step of training the first AI thread, comprising: determining a first training cloud video playback data cluster, the first training cloud video playback data cluster including several first training cloud video playback data sets and first cloud video description content bound to the first training cloud video playback data sets; loading at least one first training cloud video playback data set from the first training cloud video playback data cluster into the first AI thread to perform the cloud video playback data update process, obtaining prior training cloud video playback data bound to the first training cloud video playback data sets; loading the prior training cloud video playback data into a first comparison thread, a first description filtering thread, and a first cloud video playback data classification thread, respectively, to obtain a distinction result, a description filtering result, and a cloud video playback data classification result for the prior training cloud video playback data; combining the distinction result, description filtering result, and cloud video playback data classification result of the prior training cloud video playback data to obtain a first thread quantitative evaluation result; and updating the calculation vector of the first AI thread using the first thread quantitative evaluation result until a first training condition is met.

[0012] In one standalone embodiment, obtaining the first thread quantitative evaluation result by combining the differentiation results, description filtering results, and cloud video playback data classification results of the prior training cloud video playback data bound to the first training cloud video playback data includes: determining a first key semantic quantitative evaluation result by using the prior training cloud video playback data bound to the first training cloud video playback data and the first example cloud video playback data bound to the first training cloud video playback data in the description content of the first cloud video; obtaining a first comparison quantitative evaluation result by using the differentiation results of the prior training cloud video playback data and the differentiation results of the first comparison thread on the first example cloud video playback data; and obtaining a first comparison quantitative evaluation result by using the prior training cloud video playback data... Based on the artificial intelligence thread operation of the first example cloud video playback data, a first mining quantitative evaluation result is determined; through the description filtering result of the prior training cloud video playback data and the first example description in the first cloud video description content, a first saliency quantitative evaluation result is obtained; through the cloud video playback data classification result of the prior training cloud video playback data and the first example classification result bound to the first training sample in the first cloud video description content, a first differentiation quantitative evaluation result is obtained; through the integrated processing of the first comparison quantitative evaluation result, the first key semantic quantitative evaluation result, the first mining quantitative evaluation result, the first saliency quantitative evaluation result, and the first differentiation quantitative evaluation result, the first thread quantitative evaluation result is obtained.

[0013] In one standalone embodiment, the targeted optimization is performed by a second AI thread to obtain the optimized cloud video playback data. The method further includes a step of training the second AI thread, comprising: determining a second training cloud video playback data cluster, the second training cloud video playback data cluster including second training cloud video playback data, sample training cloud video playback data bound to the second training cloud video playback data, and second cloud video description content; optimizing the sample training cloud video playback data using the second training cloud video playback data to obtain training optimized cloud video playback data; and loading the training optimized cloud video playback data and the second training cloud video playback data into the second AI thread to train the second AI thread. The training cloud video playback data is optimized to obtain optimized pre-trained cloud video playback data for the second training cloud video playback data. The optimized pre-trained cloud video playback data is loaded into the second comparison thread, the second description filtering thread, and the second cloud video playback data classification thread to obtain the discrimination results, description filtering results, and cloud video playback data classification results for the optimized pre-trained cloud video playback data. The second thread quantization evaluation result of the second AI thread is obtained by combining the discrimination results, description filtering results, and cloud video playback data classification results of the optimized pre-trained cloud video playback data. The calculation vector of the second AI thread is updated by the second thread quantization evaluation result until the second training condition is met.

[0014] In one independently implemented embodiment, obtaining the second thread quantitative evaluation result of the second AI thread by combining the differentiation results, description filtering results, and cloud video playback data classification results of the optimized pre-trained cloud video playback data bound to the training cloud video playback data includes: obtaining an overall quantitative evaluation result and a partial quantitative evaluation result through the differentiation results, description filtering results, and cloud video playback data classification results of the optimized pre-trained cloud video playback data bound to the second training cloud video playback data; and obtaining the second thread quantitative evaluation result through the integration processing of the overall quantitative evaluation result and the partial quantitative evaluation result.

[0015] In one independent implementation, an overall quantitative evaluation result is obtained through the differentiation results, description filtering results, and cloud video playback data classification results of the optimized pre-trained cloud video playback data bound to the training cloud video playback data. This includes: determining a second key semantic quantitative evaluation result through the optimized pre-trained cloud video playback data bound to the second training cloud video playback data and the second example cloud video playback data bound to the second training cloud video playback data in the description content of the second cloud video; obtaining a second comparison quantitative evaluation result through the differentiation results of the optimized pre-trained cloud video playback data and the differentiation results of the second comparison thread on the second example cloud video playback data; and obtaining a second comparison quantitative evaluation result through the optimized pre-trained cloud video playback data. The AI ​​thread operates on the video playback data and the second example cloud video playback data to determine the second mining quantitative evaluation result; the second saliency quantitative evaluation result is obtained by optimizing the description filtering result of the pre-trained cloud video playback data and the second example description in the description content of the second cloud video; the second differentiation quantitative evaluation result is obtained by optimizing the cloud video playback data classification result of the pre-trained cloud video playback data and the second example classification result in the description content of the second cloud video; the overall quantitative evaluation result is obtained by integrating the second comparison quantitative evaluation result, the second key semantic quantitative evaluation result, the second mining quantitative evaluation result, the second saliency quantitative evaluation result, and the second differentiation quantitative evaluation result.

[0016] In one independently implemented embodiment, a partial quantitative evaluation result is obtained through the differentiation results, description filtering results, and cloud video playback data classification results of the optimized pre-trained cloud video playback data bound to the training cloud video playback data. This includes: selecting at least one indicator-based local cloud video playback data from the optimized pre-trained cloud video playback data; loading at least one indicator-based local cloud video playback data into the comparison thread, description filtering thread, and cloud video playback data classification thread respectively; and obtaining the differentiation results, description filtering results, and cloud video playback data classification results of the at least one indicator-based local cloud video playback data; and using the differentiation results of the at least one indicator-based local cloud video playback data, and the second comparison thread's evaluation of the second example cloud video playback data bound to the second training cloud video playback data. Based on the differentiation results of the indicator local cloud video playback data of at least one indicator, a third comparison quantitative evaluation result of the at least one indicator is determined; through the description filtering results of the indicator local cloud video playback data of at least one indicator and the example description of the at least one indicator in the second cloud video description content, a third saliency quantitative evaluation result of the at least one indicator is obtained; through the cloud video playback data classification results of the indicator local cloud video playback data of at least one indicator and the example classification results of the at least one indicator in the second cloud video description content, a third differentiation quantitative evaluation result of the at least one indicator is obtained; through the optimization processing of the third comparison quantitative evaluation result, the third saliency quantitative evaluation result, and the third differentiation quantitative evaluation result of the at least one indicator, a partial quantitative evaluation result of the thread is obtained.

[0017] Secondly, a cloud video playback processing system based on pupil position recognition is provided, including a processor and a memory that communicate with each other. The processor is used to retrieve a computer program from the memory and implement the above-mentioned method by running the computer program.

[0018] The cloud video playback processing method and system based on pupil position recognition provided in this application embodiment can perform optimization processing of first cloud video playback data using at least one sample cloud video playback data. Since the sample cloud video playback data includes key descriptive content of the first cloud video playback data, the obtained optimized cloud video playback data has improved accuracy and reliability compared to the first cloud video playback data. Even if the first cloud video playback data has poor performance, accurate optimized cloud video playback data can still be obtained by optimizing the sample cloud video playback data. That is, this application can conveniently perform cloud video playback data optimization based on several sample cloud video playback data, thereby ensuring differentiated video playback upgrade processing based on the optimized cloud video playback data and improving the viewing quality. Attached Figure Description

[0019] To more clearly illustrate the technical solutions of the embodiments of this application, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of this application and should not be regarded as a limitation of the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0020] Figure 1 This is a flowchart illustrating a cloud video playback processing method based on pupil position recognition, provided in an embodiment of this application.

[0021] Figure 2 This is a block diagram of a cloud video playback processing device based on pupil position recognition, provided as an embodiment of this application.

[0022] Figure 3 This is an architecture diagram of a cloud video playback processing system based on pupil position recognition, provided in an embodiment of this application. Detailed Implementation

[0023] To better understand the above technical solutions, the technical solutions of this application will be described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that the embodiments of this application and the specific features in the embodiments are detailed descriptions of the technical solutions of this application, rather than limitations on the technical solutions of this application. In the absence of conflict, the embodiments of this application and the technical features in the embodiments can be combined with each other.

[0024] Please see Figure 1 This paper presents a cloud video playback processing method based on pupil position recognition, which may include the technical solutions described in steps 10-30 below.

[0025] Step 10: Determine the first cloud video playback data.

[0026] In one possible implementation, the pupil position of the cloud video playback data to be processed can first be determined, namely the first cloud video playback data. In this embodiment of the application, the first cloud video playback data can be understood as having relatively low accuracy in quantitative analysis indicators.

[0027] Step 20: Determine at least one sample cloud video playback data from the first cloud video playback data, wherein the sample cloud video playback data covers the sample data of the reference pupil position in the first cloud video playback data.

[0028] In one possible implementation, the first cloud video playback data may have corresponding sample cloud video playback data. The sample cloud video playback data includes sample data of the reference pupil position in the first cloud video playback data, such as sample data of at least one reference indicator of the reference pupil position.

[0029] In one possible implementation, sample cloud video playback data can be associated with the first cloud video playback data, or it can be obtained based on determined spatial positioning data regarding the reference pupil position. The spatial positioning data can include X types of identification tags for the reference pupil position. For example, when the reference pupil position is a local pupil position, the spatial positioning data can include identification tags for X types of reference indicators for the local pupil position. Alternatively, the spatial positioning data can directly include global spatial positioning data of the reference pupil position in the first cloud video playback data, such as spatial positioning data of some determined pupil positions. Through the spatial positioning data, at least one reference indicator of the reference pupil position of the first cloud video playback data can be determined, or cloud video playback data including a pupil position consistent with the pupil position in the first cloud video playback data can be determined. Each obtained identical cloud video playback data or cloud video playback data including a consistent pupil position can be understood as sample cloud video playback data.

[0030] Step 30: Optimize the first cloud video playback data by using at least one sample cloud video playback data to obtain optimized cloud video playback data.

[0031] After obtaining at least one sample cloud video playback data bound to the first cloud video playback data, optimization of the first cloud video playback data can be performed based on the obtained sample cloud video playback data. Since the sample cloud video playback data includes sample data of at least one reference indicator of the pupil position in the first cloud video playback data, the first cloud video playback data can be optimized in a targeted manner based on the sample data. Moreover, even if the first cloud video playback data is severely lacking, more accurate optimized cloud video playback data can still be obtained based on the sample data.

[0032] In one possible implementation, the sample cloud video playback data of the corresponding reference indication can be directly iterated into the first cloud video playback data to obtain optimized cloud video playback data.

[0033] In one possible implementation, optimized cloud video playback data can also be obtained based on feature extraction operations of sample cloud video playback data and first cloud video playback data.

[0034] In one possible implementation, since the pupil position location of the sample cloud video playback data may differ from the reference pupil position location in the first cloud video playback data, it is necessary to compare each sample cloud video playback data with the first cloud video playback data. That is, the pupil position location in the sample cloud video playback data is adjusted to match the reference pupil position location in the first cloud video playback data. Then, the first cloud video playback data is optimized using the adjusted sample cloud video playback data. This step improves the reliability of the optimized cloud video playback data.

[0035] In this embodiment, the first cloud video playback data is optimized based on at least one sample cloud video playback data. The resulting optimized cloud video playback data can optimize the sample data of each sample cloud video playback data, and has relatively high accuracy and reliability.

[0036] A cloud video playback processing method based on pupil position recognition, in conjunction with an embodiment of this application, may specifically include the following steps in determining the content described by at least one sample of the first cloud video playback data.

[0037] Step 21: Determine the spatial location data of the first cloud video playback data.

[0038] Furthermore, the spatial positioning data of the first cloud video playback data may include an identification tag (or key spatial positioning data) of at least one reference indicator of the reference pupil position in the first cloud video playback data.

[0039] Step 22: Determine sample cloud video playback data that are associated with at least one reference indicator of the pupil position using the spatial positioning data of the first cloud video playback data.

[0040] After obtaining the spatial positioning data, it is possible to determine the sample cloud video playback data that is related to the pupil position in the first cloud video playback data based on the spatial positioning data.

[0041] In one possible implementation, the spatial positioning data may also include tag data about the position of the first pupil in the first cloud video playback data. In this case, cloud video playback data that is associated with the tag data can be filtered from the data center as sample cloud video playback data based on the tag data.

[0042] Through the above steps, sample cloud video playback data that are associated with at least one reference indicator of the pupil position in the first cloud video playback data can be determined based on spatial positioning data. Optimizing the cloud video playback data using sample cloud video playback data can improve the accuracy of the determined cloud video playback data.

[0043] After obtaining the sample cloud video playback data, the optimization process of the cloud video playback data can be performed based on the sample cloud video playback data. In addition to directly iterating the sample cloud video playback data to the corresponding reference indication of the first cloud video playback data, the embodiments of this application can also perform iteration or feature extraction after optimizing the sample cloud video playback data to obtain the optimized cloud video playback data.

[0044] A cloud video playback processing method based on pupil position recognition, in conjunction with an embodiment of this application, wherein the first cloud video playback data is optimized by using at least one sample cloud video playback data to obtain optimized cloud video playback data, may specifically include the following steps.

[0045] Step 31: Based on the real-time viewing situation of the reference pupil position in the first cloud video playback data, optimize the cloud video playback data of at least one sample to obtain optimized cloud video playback data that is bound to the sample cloud video playback data in the real-time viewing situation.

[0046] In one possible implementation, since the pupil position of the sample cloud video playback data obtained regarding the pupil position in the first cloud video playback data may differ from the pupil position in the first cloud video playback data, it is necessary to compare each sample cloud video playback data with the first cloud video playback data to ensure that the pupil position in the sample cloud video playback data is consistent with the reference pupil position in the first cloud video playback data.

[0047] The embodiments of this application can optimize the sample cloud video playback data through optimization processing steps, and the positioning of the pupil position in the optimized sample cloud video playback data (i.e., optimized cloud video playback data) is consistent with the positioning of the reference pupil position in the first cloud video playback data.

[0048] Based on the above, at least one optimized cloud video playback data point with the same positioning as the first cloud video playback data point can be obtained (each sample cloud video playback data point is optimized to obtain an optimized cloud video playback data point), thus enabling the comparison between the optimized cloud video playback data point and the first cloud video playback data point.

[0049] Step 32: Select local cloud video playback data of the reference indicator from the optimized cloud video playback data bound to the sample cloud video playback data, based on the reference indicator that is associated with the reference pupil position in the sample cloud video playback data.

[0050] Since the obtained sample cloud video playback data is cloud video playback data that is associated with at least one reference indicator in the first cloud video playback data, after optimizing the cloud video playback data bound to each sample cloud video playback data, the local cloud video playback data of the sample indicator (the reference indicator associated with the pupil position) bound to each sample cloud video playback data can be filtered from the optimized cloud video playback data. That is, the local cloud video playback data of the reference indicator associated with the pupil position in the first cloud video playback data can be distinguished from the optimized cloud video playback data.

[0051] Step 33: Based on the selected local cloud video playback data and the first cloud video playback data, obtain the optimized cloud video playback data.

[0052] After obtaining local cloud video playback data with at least one reference indicator of the reference pupil position, the cloud video playback data can be optimized using the obtained local cloud video playback data and the first cloud video playback data to obtain optimized cloud video playback data.

[0053] In one possible implementation, since each local cloud video playback data can be associated with at least one reference indication in the pupil position of the first cloud video playback data, the cloud video playback data with the associated indication in the local cloud video playback data can be iterated to the corresponding indication in the first cloud video playback data.

[0054] Alternatively, in one possible implementation, the optimized cloud video playback data can be obtained through feature extraction operations on the local cloud video playback data and the first cloud video playback data.

[0055] Specifically, the playback data of each local cloud video and the first cloud video can be loaded into the AI ​​thread, and at least one feature extraction operation can be performed to optimize the features of the cloud video playback data. Finally, the optimized key content can be obtained, and the optimized cloud video playback data bound to the optimized key content can be obtained.

[0056] In one feasible embodiment, to further improve the accuracy and reliability of the optimized cloud video playback data, the first cloud video playback data can be processed to obtain a second cloud video playback data with a larger quantitative analysis index than the first cloud video playback data. The optimized cloud video playback data can then be obtained by performing cloud video playback data optimization using the second cloud video playback data. A cloud video playback processing method based on pupil position recognition in this application embodiment, wherein the targeted optimization of the first cloud video playback data using at least one sample of the first cloud video playback data to obtain the optimized cloud video playback data, specifically includes the following steps.

[0057] Step 301: Perform cloud video playback data update processing on the first cloud video playback data to obtain the second cloud video playback data. The quantitative analysis index of the second cloud video playback data exceeds the quantitative analysis index of the first cloud video playback data.

[0058] In one possible implementation, based on the first cloud video playback data, the cloud video playback data can be updated to obtain second cloud video playback data with improved quantitative analysis indicators. The cloud video playback data update process can update the cloud video playback data with lower quantitative analysis indicators or by sorting the cloud video playback data to obtain cloud video playback data with higher quantitative analysis indicators.

[0059] Step 302: Based on the real-time viewing situation of the reference pupil position in the second cloud video playback data, optimize the at least one sample cloud video playback data to obtain optimized cloud video playback data that is bound to the sample cloud video playback data in the real-time viewing situation.

[0060] Since the second cloud video playback data has improved the quantitative analysis indicators compared to the first cloud video playback data, the positioning of the reference pupil position in the second cloud video playback data may differ from the positioning of the sample cloud video playback data. Before performing optimization, the sample cloud video playback data can be optimized and changed based on the positioning of the reference pupil position in the second cloud video playback data to obtain optimized cloud video playback data that is consistent with the positioning of the reference pupil position in the second cloud video playback data.

[0061] Step 303: Select local cloud video playback data of the not less than one reference indicator that is associated with the pupil position from the optimized cloud video playback data bound to the not less than one sample cloud video playback data.

[0062] Since the obtained sample cloud video playback data is cloud video playback data that is associated with at least one reference indicator in the second cloud video playback data, after optimizing the cloud video playback data bound to each sample cloud video playback data, the local cloud video playback data of the sample indicator (the reference indicator associated with the pupil position) bound to each sample cloud video playback data can be filtered from the optimized cloud video playback data. That is, the local cloud video playback data of the reference indicator associated with the pupil position in the first cloud video playback data can be distinguished from the optimized cloud video playback data.

[0063] Step 304: Obtain the optimized cloud video playback data based on the selected local cloud video playback data and the second cloud video playback data.

[0064] After obtaining local cloud video playback data with at least one reference indicator of the reference pupil position, the cloud video playback data can be optimized using the obtained local cloud video playback data and the second cloud video playback data to obtain optimized cloud video playback data.

[0065] In one possible implementation, since each local cloud video playback data can be associated with at least one reference indicator in the pupil position of the second cloud video playback data, the cloud video playback data with associated indicators in the local cloud video playback data can be iterated to the corresponding indicator in the second cloud video playback data. Alternatively, in one possible implementation, the optimized cloud video playback data can also be obtained through feature extraction operations on the local cloud video playback data and the second cloud video playback data.

[0066] Through the above steps, the accuracy of the quantitative analysis indicators of the first cloud video playback data can be further improved through update processing, and more accurate optimized cloud video playback data can be obtained in parallel.

[0067] This application embodiment trains a first AI thread. The process of training the AI ​​thread may specifically include the following steps.

[0068] Step 51: Determine the first training cloud video playback data cluster. The first training cloud video playback data cluster includes several first training cloud video playback data and first cloud video description content bound to the first training cloud video playback data.

[0069] In one possible implementation, the training cloud video playback data cluster may include several first training cloud video playback data, which can be understood as cloud video playback data with relatively low quantitative analysis indicators.

[0070] Step 52: Load at least one first training cloud video playback data from the first training cloud video playback data cluster into the first AI thread to perform the cloud video playback data update process, and obtain the prior training cloud video playback data bound to the first training cloud video playback data.

[0071] When training the first AI thread, the cloud video playback data in the first training cloud video playback data cluster can be loaded into the first AI thread together, or loaded into the first AI thread multiple times, and the updated and processed prior training cloud video playback data bound to each first training cloud video playback data can be obtained one by one.

[0072] Step 53: Load the prior training cloud video playback data one by one into the first comparison thread, the first description filtering thread, and the first cloud video playback data classification thread to obtain the distinction results, description filtering results, and cloud video playback data classification results of the prior training cloud video playback data bound to the first training cloud video playback data.

[0073] The previously trained cloud video playback data is loaded into the comparison thread, description filtering thread, and cloud video playback data classification thread mentioned above to obtain the distinction results, description filtering results, and cloud video playback data classification results of the previously trained cloud video playback data bound to the trained cloud video playback data.

[0074] Step 54: Combine the differentiation results, description filtering results, and cloud video playback data classification results of the prior training cloud video playback data to obtain the first thread quantitative evaluation result. Update the calculation vector of the first AI thread through the first thread quantitative evaluation result until the first training condition is met.

[0075] In one possible implementation, the comparison quantitative evaluation result can be obtained based on the differentiation result of the prior training cloud video playback data, the differentiation quantitative evaluation result can be obtained based on the classification result of the cloud video playback data, the saliency quantitative evaluation result can be obtained based on the obtained description screening result, and the corresponding key semantic quantitative evaluation result and the processed mining quantitative evaluation result can be obtained based on the obtained prior training cloud video playback data.

[0076] The above describes the training process for the first AI thread. In this embodiment, the cloud video playback data optimization process in step 30 can also be performed using a second AI thread, which can be understood as a convolutional AI thread. Training the second AI thread in conjunction with this embodiment can specifically include the following steps.

[0077] Step 61: Determine the second training cloud video playback data cluster. The second training cloud video playback data cluster includes several second training cloud video playback data, sample training cloud video playback data bound to the second training cloud video playback data, and second cloud video description content.

[0078] In one possible implementation, the second training cloud video playback data in the second training cloud video playback data cluster can be understood as the prior training cloud video playback data formed by the first AI thread in advance, or it can also be understood as cloud video playback data with relatively low accuracy of quantitative analysis indicators obtained through the remaining steps.

[0079] When training the second AI thread, it can be understood that each training cloud video playback data set trains at least one sample training cloud video playback data set. This sample training cloud video playback data set includes sample data bound to the second training cloud video playback data set, such as at least one instruction cloud video playback data set. The sample training cloud video playback data set also consists of highly quantifiable analytical indicators and accurate cloud video playback data. Each second training cloud video playback data set can include a different number of sample training cloud video playback data sets, and the sample instructions bound to each sample training cloud video playback data set can also be different.

[0080] The description content of the second cloud video can also be determined based on the calculation vector of the quantitative evaluation result thread. It can include the second example cloud video playback data (accurate cloud video playback data) bound to the second training cloud video playback data, the second example description of the second example cloud video playback data, the second example classification result (real-time classification result of each indicator), and can also include the distinction result of each indicator in the second example cloud video playback data (distinction result output by the comparison thread), description filtering results, and classification results, etc.

[0081] Step 62: Optimize the sample training cloud video playback data using the second training cloud video playback data to obtain optimized training cloud video playback data. Load the optimized training cloud video playback data and the second training cloud video playback data into the second AI thread. Perform targeted optimization on the second training cloud video playback data to obtain optimized pre-training cloud video playback data of the second training cloud video playback data.

[0082] Furthermore, each second training cloud video playback data set can be bound to at least one sample cloud video playback data set. The pupil position in the second training cloud video playback data set can be used to optimize the sample training cloud video playback data set, resulting in at least one optimized training cloud video playback data set. The at least one optimized training cloud video playback data set bound to the second training cloud video playback data set, along with the second training cloud video playback data set, can be loaded into the second AI thread to obtain the corresponding optimized pre-trained cloud video playback data set.

[0083] Step 63: Load the optimized pre-trained cloud video playback data bound to the training cloud video playback data into the second comparison thread, the second description filtering thread, and the second cloud video playback data classification thread, respectively, to obtain the distinction results, description filtering results, and cloud video playback data classification results for the optimized pre-trained cloud video playback data bound to the second training cloud video playback data.

[0084] Step 64: Combine the differentiation results, description filtering results, and cloud video playback data classification results of the optimized pre-trained cloud video playback data bound to the second training cloud video playback data to obtain the second thread quantitative evaluation result of the second AI thread, and update the calculation vector of the second AI thread through the second thread quantitative evaluation result until the second training condition is met.

[0085] In one possible implementation, the second thread quantitative evaluation result can be understood as the integration of the overall quantitative evaluation result and the partial quantitative evaluation result. That is, the overall quantitative evaluation result and the partial quantitative evaluation result can be obtained by using the differentiation result, description filtering result and cloud video playback data classification result of the optimized pre-trained cloud video playback data bound to the training cloud video playback data, and the second thread quantitative evaluation result is obtained by integrating the overall quantitative evaluation result and the partial quantitative evaluation result.

[0086] Furthermore, the overall quantitative evaluation result can be understood as an integrated processing of the comparative quantitative evaluation result, the key semantic quantitative evaluation result, the mining quantitative evaluation result, the differentiation quantitative evaluation result, and the significance quantitative evaluation result based on the optimized pre-trained cloud video playback data.

[0087] Optionally, consistent with the steps for determining the first comparison and quantitative evaluation result, referring to the comparison and quantitative evaluation result thread, the second comparison and quantitative evaluation result can be obtained through the differentiation results of the optimized pre-trained cloud video playback data and the differentiation results of the second example cloud video playback data in the description content of the second cloud video. Consistent with the steps for determining the first key semantic quantitative evaluation result, referring to the key semantic quantitative evaluation result thread, the second key semantic quantitative evaluation result can be determined through the optimized pre-trained cloud video playback data bound to the second training cloud video playback data and the second example cloud video playback data bound to the second training cloud video playback data. Consistent with the steps for determining the first mining quantitative evaluation result, referring to the mining quantitative evaluation result thread, the second key semantic quantitative evaluation result can be determined through the artificial intelligence thread operation of the optimized pre-trained cloud video playback data and the second example cloud video playback data bound to the second training cloud video playback data. The second mining quantitative evaluation result is determined. Consistent with the steps for determining the first saliency quantitative evaluation result, referring to the saliency quantitative evaluation result thread, the second saliency quantitative evaluation result can be obtained through the description filtering result of the optimized pre-trained cloud video playback data bound to the second training cloud video playback data and the second example description in the second cloud video description content. Consistent with the steps for determining the first differentiation quantitative evaluation result, referring to the differentiation quantitative evaluation result thread, the second differentiation quantitative evaluation result can be obtained through the cloud video playback data classification result of the optimized pre-trained cloud video playback data bound to the second training cloud video playback data and the second example classification result in the second cloud video description content. The overall quantitative evaluation result is obtained through the integrated processing of the second comparison quantitative evaluation result, the second key semantic quantitative evaluation result, the second mining quantitative evaluation result, the second saliency quantitative evaluation result, and the second differentiation quantitative evaluation result.

[0088] In an alternative embodiment, determining the partial quantization evaluation result of the second AI thread may include: selecting at least one indicator-bound indicator local cloud video playback data from the optimized pre-trained cloud video playback data; loading the at least one indicator-bound indicator local cloud video playback data into a comparison thread, a description filtering thread, and a cloud video playback data classification thread, respectively, to obtain the distinction result, description filtering result, and cloud video playback data classification result of the at least one indicator-bound indicator local cloud video playback data; and using the distinction result of the at least one indicator-bound indicator local cloud video playback data, and the second comparison thread's analysis of the at least one indicator-bound indicator local cloud video playback data from the second example cloud video playback data bound to the second training cloud video playback data, the determination may include: selecting indicator local cloud video playback data with at least one indicator bound to the optimized pre-trained cloud video playback data; loading the at least one indicator-bound indicator local cloud video playback data into a comparison thread, a description filtering thread, and a cloud video playback data classification thread, respectively, to obtain the distinction result, description filtering result, and cloud video playback data classification result of the at least one indicator-bound indicator local cloud video playback data; and using the distinction result of the at least one indicator-bound indicator local cloud video playback data, the determination may include: selecting indicator local cloud video playback data with at least one indicator bound to the optimized pre-trained cloud video playback data; and loading the at least one indicator-bound indicator local cloud video playback data into the second example cloud video playback data, the determination may include: selecting indicator local cloud video playback data with at least one indicator bound to the optimized pre-trained cloud video playback data; and loading the at least one indicator-bound indicator local cloud video playback data into the second example cloud video playback data, a ... Based on the data differentiation results, a third comparison quantification evaluation result for at least one indication is determined; a third salience quantification evaluation result for at least one indication is obtained through the description filtering result of the local cloud video playback data of at least one indication and the example description of the corresponding indication in the description content of the second cloud video; a third differentiation quantification evaluation result for at least one indication is obtained through the cloud video playback data classification result of the local cloud video playback data of at least one indication and the example classification result of the at least one indication in the description content of the second cloud video; and a partial quantification evaluation result for the thread is obtained through the optimization processing of the third comparison thread quantification evaluation result, the third salience quantification evaluation result, and the third differentiation quantification evaluation result for at least one indication.

[0089] Consistent with the steps for determining the above quantitative evaluation results, the partial quantitative evaluation results for each indicator can be determined by optimizing the third comparison quantitative evaluation results, the third key semantic quantitative evaluation results, and the third mining quantitative evaluation results of the local cloud video playback data for each indicator in the pre-trained cloud video playback data.

[0090] Based on the above, please refer to the following: Figure 2 A cloud video playback processing device 200 based on pupil position recognition is provided, which is applied to a cloud video playback processing system based on pupil position recognition. The device includes:

[0091] The playback data determination module 210 is used to determine the playback data of the first cloud video;

[0092] The sample data determination module 220 is used to determine at least one sample cloud video playback data of the first cloud video playback data, wherein the sample cloud video playback data covers the sample data of the reference pupil position in the first cloud video playback data;

[0093] The playback data optimization module 230 is used to perform targeted optimization on the first cloud video playback data using at least one sample cloud video playback data to obtain optimized cloud video playback data.

[0094] Based on the above, please refer to the following: Figure 3 The present invention illustrates a cloud video playback processing system 300 based on pupil position recognition, including a processor 310 and a memory 320 that communicate with each other. The processor 310 is used to read computer programs from the memory 320 and execute them to implement the above-described method.

[0095] Based on the above, a computer-readable storage medium is also provided, on which a computer program stored implements the above method during runtime.

[0096] In summary, based on the above scheme, the optimization processing of the first cloud video playback data can be performed using at least one sample cloud video playback data. Since the sample cloud video playback data includes the key descriptive content of the first cloud video playback data, the resulting optimized cloud video playback data has improved accuracy and reliability compared to the first cloud video playback data. Even if the first cloud video playback data has poor performance, accurate optimized cloud video playback data can still be obtained by optimizing the sample cloud video playback data. In other words, this application can conveniently perform cloud video playback data optimization based on several sample cloud video playback data, thereby ensuring differentiated video playback upgrade processing based on the optimized cloud video playback data and improving the viewing quality.

[0097] It should be understood that the systems and modules described above can be implemented in various ways. For example, in some embodiments, the systems and modules can be implemented by hardware, software, or a combination of both. The hardware portion can be implemented using dedicated logic; the software portion can be stored in memory and executed by an appropriate instruction execution system, such as a microprocessor or dedicated-design hardware. Those skilled in the art will understand that the methods and systems described above can be implemented using computer-executable instructions and / or included in processor control code, for example, such code provided on a carrier medium such as a disk, CD, or DVD-ROM, a programmable memory such as read-only memory (firmware), or a data carrier such as an optical or electronic signal carrier. The systems and modules of this application can be implemented not only by hardware circuits such as very large-scale integrated circuits or gate arrays, semiconductors such as logic chips, transistors, or programmable hardware devices such as field-programmable gate arrays, programmable logic devices, etc., but also by software executed by various types of processors, or by a combination of the aforementioned hardware circuits and software (e.g., firmware).

[0098] It should be noted that different embodiments may produce different beneficial effects. In different embodiments, the beneficial effects may be any one or a combination of the above, or any other possible beneficial effects.

[0099] The basic concepts have been described above. Obviously, for those skilled in the art, the detailed disclosure above is merely illustrative and does not constitute a limitation of this application. Although not explicitly stated herein, those skilled in the art may make various modifications, improvements, and corrections to this application. Such modifications, improvements, and corrections are suggested in this application, and therefore remain within the spirit and scope of the exemplary embodiments of this application.

[0100] Furthermore, this application uses specific terms to describe embodiments of the application. For example, "an embodiment," "one embodiment," and / or "some embodiments" refer to a particular feature, structure, or characteristic associated with at least one embodiment of the application. Therefore, it should be emphasized and noted that "an embodiment," "one embodiment," or "an alternative embodiment" mentioned twice or more in different locations in this specification do not necessarily refer to the same embodiment. In addition, certain features, structures, or characteristics in one or more embodiments of the application can be appropriately combined.

[0101] Furthermore, those skilled in the art will understand that aspects of this application can be described and illustrated through several patentable types or situations, including any new and useful combination of processes, machines, products, or substances, or any new and useful improvements thereof. Accordingly, aspects of this application can be implemented entirely by hardware, entirely by software (including firmware, resident software, microcode, etc.), or by a combination of hardware and software. All of the above hardware or software may be referred to as a “data block,” “module,” “engine,” “unit,” “component,” or “system.” Furthermore, aspects of this application may manifest as a computer product located on one or more computer-readable media, the product including computer-readable program code.

[0102] Computer storage media may contain a propagated data signal containing computer program code, for example, on baseband or as part of a carrier wave. This propagated signal may take various forms, including electromagnetic, optical, and suitable combinations thereof. Computer storage media can be any computer-readable medium other than a computer-readable storage medium, which can be connected to an instruction execution system, apparatus, or device to enable communication, propagation, or transmission of a program for use. The program code located on the computer storage medium can be propagated through any suitable medium, including radio, cable, fiber optic cable, RF, or similar media, or any combination of the above media.

[0103] The computer program code required for the operation of each part of this application can be written in any one or more programming languages, including object-oriented programming languages ​​such as Java, Scala, Smalltalk, Eiffel, JADE, Emerald, C++, C#, VB.NET, Python, etc., conventional procedural programming languages ​​such as C, Visual Basic, Fortran 2003, Perl, COBOL 2002, PHP, ABAP, dynamic programming languages ​​such as Python, Ruby, and Groovy, or other programming languages. This program code can run entirely on the user's computer, or as a standalone software package on the user's computer, or partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In the latter case, the remote computer can be connected to the user's computer through any network, such as a local area network (LAN) or wide area network (WAN), or connected to an external computer (e.g., via the Internet), or in a cloud computing environment, or used as a service such as Software as a Service (SaaS).

[0104] Furthermore, unless expressly stated in the claims, the order of processing elements and sequences, the use of numbers and letters, or other names described in this application are not intended to limit the order of the processes and methods of this application. Although the foregoing disclosure has discussed some currently considered useful embodiments of the invention through various examples, it should be understood that such details are for illustrative purposes only, and the appended claims are not limited to the disclosed embodiments; rather, the claims are intended to cover all modifications and equivalent combinations that conform to the substance and scope of the embodiments of this application. For example, while the system components described above can be implemented using hardware devices, they can also be implemented solely through software solutions, such as installing the described system on existing servers or mobile devices.

[0105] Similarly, it should be noted that, in order to simplify the description of the present application and thus aid in the understanding of one or more embodiments of the invention, the foregoing description of the embodiments of the present application sometimes combines multiple features into a single embodiment, drawing, or description thereof. However, this disclosure method does not imply that the subject matter of the application requires more features than those mentioned in the claims. In fact, the embodiments contain fewer features than all the features of the single embodiments disclosed above.

[0106] In some embodiments, numbers describing the quantity of components and attributes are used. It should be understood that such numbers used in the description of embodiments are modified in some examples with the terms "approximately," "approximately," or "generally." Unless otherwise stated, "approximately," "approximately," or "generally" indicates that the numbers are open to adaptive variation. Accordingly, in some embodiments, the numerical parameters used in the specification and claims are approximate values, which may be changed depending on the characteristics required by individual embodiments. In some embodiments, numerical parameters are taken into account a specified number of significant digits and employ a general method of digit reservation. Although the numerical ranges and parameters used to confirm their breadth of application in some embodiments of this application are approximate values, in specific embodiments, such values ​​are set as precisely as feasible.

[0107] For each patent, patent application, patent application publication, and other material such as articles, books, specifications, publications, and documents referenced in this application, the entire contents of that patent are incorporated herein by reference. This excludes historical application documents that are inconsistent with or conflict with the content of this application, as well as documents that limit the broadest scope of the claims in this application (currently or subsequently appended to this application). It should be noted that if there are any inconsistencies or conflicts between the descriptions, definitions, and / or terminology used in the supplementary materials of this application and the content of this application, the descriptions, definitions, and / or terminology used in this application shall prevail.

[0108] Finally, it should be understood that the embodiments described in this application are merely illustrative of the principles of the embodiments of this application. Other modifications may also fall within the scope of this application. Therefore, alternative configurations of the embodiments of this application are considered as examples and not limitations, and are regarded as consistent with the teachings of this application. Accordingly, the embodiments of this application are not limited to the embodiments explicitly described and illustrated in this application.

[0109] The above are merely embodiments of this application and are not intended to limit the scope of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of the claims of this application.

Claims

1. A cloud video playback processing method based on pupil position recognition, characterized in that, The method includes at least: Determine the first cloud video playback data; Determine at least one sample cloud video playback data of the first cloud video playback data, wherein the sample cloud video playback data covers sample data of the reference pupil position in the first cloud video playback data; The first cloud video playback data is optimized by using at least one sample cloud video playback data to obtain optimized cloud video playback data. The step of determining at least one sample cloud video playback data of the first cloud video playback data includes: determining spatial positioning data of the first cloud video playback data, wherein the spatial positioning data covers X kinds of identification tags of the reference pupil position in the first cloud video playback data; determining sample cloud video playback data that are associated with at least one reference indicator of the reference pupil position through the spatial positioning data of the first cloud video playback data; wherein X is a positive number greater than or equal to 1; The step of optimizing the first cloud video playback data by using at least one sample of cloud video playback data to obtain optimized cloud video playback data includes: The first cloud video playback data is updated to obtain the second cloud video playback data. The quantitative analysis index of the second cloud video playback data exceeds the quantitative analysis index of the first cloud video playback data. By using the real-time viewing situation of the reference pupil position in the second cloud video playback data, the not less than one sample cloud video playback data is optimized to obtain optimized cloud video playback data that is bound to the sample cloud video playback data in the real-time viewing situation. Using at least one reference indicator that is associated with the pupil position in at least one sample cloud video playback data, select local cloud video playback data of at least one reference indicator from the optimized cloud video playback data bound to the sample cloud video playback data; The optimized cloud video playback data is obtained based on the selected local cloud video playback data and the second cloud video playback data; The step of obtaining the optimized cloud video playback data based on the selected local cloud video playback data and the second cloud video playback data includes: iterating through the selected local cloud video playback data the indications bound to the reference indications in the second cloud video playback data to obtain the optimized cloud video playback data, or performing feature extraction operations through the local cloud video playback data and the second cloud video playback data to obtain the optimized cloud video playback data; The method further includes a step of training the first AI thread to update the cloud video playback data, thereby obtaining the second cloud video playback data, by performing the cloud video playback data update process on the first cloud video playback data through the first AI thread. A first training cloud video playback data cluster is determined. The first training cloud video playback data cluster includes several first training cloud video playback data and first cloud video description content bound to the first training cloud video playback data. Load at least one first training cloud video playback data from the first training cloud video playback data cluster into the first AI thread to perform the cloud video playback data update process, and obtain the prior training cloud video playback data bound to the first training cloud video playback data. The prior training cloud video playback data is loaded into the first comparison thread, the first description filtering thread, and the first cloud video playback data classification thread respectively to obtain the distinction results, description filtering results, and cloud video playback data classification results for the prior training cloud video playback data. The first thread quantitative evaluation result is obtained by combining the differentiation result, description filtering result, and cloud video playback data classification result of the prior training cloud video playback data. The calculation vector of the first AI thread is updated by the first thread quantitative evaluation result until the first training condition is met. The step of obtaining the first thread quantitative evaluation result by combining the differentiation results, description filtering results, and cloud video playback data classification results of the prior training cloud video playback data bound to the first training cloud video playback data includes: A first key semantic quantitative evaluation result is determined by combining the prior training cloud video playback data bound to the first training cloud video playback data with the first example cloud video playback data bound to the first cloud video description content. A first comparison quantitative evaluation result is obtained by combining the differentiation results of the prior training cloud video playback data with the differentiation results of the first comparison thread on the first example cloud video playback data. A first mining quantitative evaluation result is determined by combining the artificial intelligence thread operation of the prior training cloud video playback data and the first example cloud video playback data. A first saliency quantitative evaluation result is obtained by combining the description filtering results of the prior training cloud video playback data with the first example description in the first cloud video description content. The first discrimination quantification evaluation result is obtained by combining the cloud video playback data classification result of the prior training cloud video playback data and the first example classification result of the first cloud video description content bound to the first training sample. The first thread quantitative evaluation result is obtained by integrating the first comparison quantitative evaluation result, the first key semantic quantitative evaluation result, the first mining quantitative evaluation result, the first saliency quantitative evaluation result, and the first differentiation quantitative evaluation result. The optimized cloud video playback data is obtained by performing the targeted optimization through a second AI thread. The method further includes a step of training the second AI thread, including: A second training cloud video playback data cluster is determined, which includes second training cloud video playback data, sample training cloud video playback data bound to the second training cloud video playback data, and second cloud video description content. The sample training cloud video playback data is optimized using the second training cloud video playback data to obtain training optimized cloud video playback data. The training optimized cloud video playback data and the second training cloud video playback data are loaded into the second AI thread, and targeted optimization is performed on the second training cloud video playback data to obtain optimized pre-training cloud video playback data of the second training cloud video playback data. The optimized pre-trained cloud video playback data is loaded into the second comparison thread, the second description filtering thread, and the second cloud video playback data classification thread to obtain the distinction results, description filtering results, and cloud video playback data classification results for the optimized pre-trained cloud video playback data. The second AI thread's second thread quantitative evaluation result is obtained by combining the differentiation result, description filtering result, and cloud video playback data classification result of the optimized pre-trained cloud video playback data. The second AI thread's calculation vector is then updated using the second thread quantitative evaluation result until the second training condition is met. Specifically, obtaining the second thread quantitative evaluation result of the second AI thread by combining the discrimination results, description filtering results, and cloud video playback data classification results of the optimized pre-trained cloud video playback data bound to the training cloud video playback data includes: obtaining an overall quantitative evaluation result and a partial quantitative evaluation result through the discrimination results, description filtering results, and cloud video playback data classification results of the optimized pre-trained cloud video playback data bound to the second training cloud video playback data; and obtaining the second thread quantitative evaluation result through the integration processing of the overall quantitative evaluation result and the partial quantitative evaluation result. The overall quantitative evaluation result is obtained by binding the optimized pre-trained cloud video playback data with the training cloud video playback data, including the differentiation results, description filtering results, and cloud video playback data classification results. The second key semantic quantitative evaluation result is determined by binding the optimized pre-trained cloud video playback data to the second training cloud video playback data and the second example cloud video playback data bound to the second training cloud video playback data in the description content of the second cloud video. The second comparison quantitative evaluation result is obtained by using the differentiation result of the optimized pre-trained cloud video playback data and the differentiation result of the second comparison thread on the second example cloud video playback data. The second mining quantitative evaluation result is determined by the artificial intelligence thread operation of the optimized pre-trained cloud video playback data and the second example cloud video playback data; the second significance quantitative evaluation result is obtained by the description filtering result of the optimized pre-trained cloud video playback data and the second example description in the description content of the second cloud video. The second differentiation quantification evaluation result is obtained by combining the cloud video playback data classification result of the optimized pre-trained cloud video playback data with the second example classification result in the second cloud video description content. The overall quantitative evaluation result is obtained by integrating the second comparison quantitative evaluation result, the second key semantic quantitative evaluation result, the second mining quantitative evaluation result, the second saliency quantitative evaluation result, and the second differentiation quantitative evaluation result. The partial quantitative evaluation results are obtained through the differentiation results, description filtering results, and cloud video playback data classification results of the optimized pre-trained cloud video playback data bound to the training cloud video playback data. This includes: selecting at least one indicator-based local cloud video playback data from the optimized pre-trained cloud video playback data; loading at least one indicator-based local cloud video playback data into the comparison thread, description filtering thread, and cloud video playback data classification thread respectively; and obtaining the differentiation results, description filtering results, and cloud video playback data classification results of the at least one indicator-based local cloud video playback data. The evaluation results are then obtained through the differentiation results of the at least one indicator-based local cloud video playback data, and the results obtained by the second comparison thread on the second example cloud video playback data bound to the second training cloud video playback data. The differentiation results of less than one indicator in the local cloud video playback data are used to determine the third comparison and quantification evaluation result of at least one indicator; the description filtering results of the less than one indicator in the local cloud video playback data and the example descriptions of at least one indicator in the second cloud video description content are used to obtain the third saliency quantification evaluation result of at least one indicator; the cloud video playback data classification results of the less than one indicator in the local cloud video playback data and the example classification results of at least one indicator in the second cloud video description content are used to obtain the third differentiation quantification evaluation result of at least one indicator; through the optimization processing of the third comparison and quantification evaluation result, the third saliency quantification evaluation result, and the third differentiation quantification evaluation result of at least one indicator, the partial quantification evaluation result of the thread is obtained.

2. The method as described in claim 1, characterized in that, The step of optimizing the first cloud video playback data by using at least one sample of cloud video playback data to obtain optimized cloud video playback data includes: Based on the real-time viewing situation of the reference pupil position in the first cloud video playback data, the not less than one sample cloud video playback data is optimized to obtain optimized cloud video playback data that is bound to the sample cloud video playback data in the real-time viewing situation. Using at least one reference indicator associated with the reference pupil position from at least one sample cloud video playback data, select local cloud video playback data of the at least one reference indicator from the optimized cloud video playback data bound to the sample cloud video playback data; obtain the optimized cloud video playback data based on the selected local cloud video playback data and the first cloud video playback data.

3. The method as described in claim 2, characterized in that, The process of obtaining the optimized cloud video playback data based on the selected local cloud video playback data and the first cloud video playback data includes: The optimized cloud video playback data is obtained by iterating over the selected local cloud video playback data and the indications bound to the reference indications in the first cloud video playback data and the local cloud video playback data, or by performing feature extraction operations on the local cloud video playback data and the first cloud video playback data.

4. The method as described in claim 1, characterized in that, The method further includes: reading tags using the optimized cloud video playback data to determine tag data associated with the pupil position.

5. A cloud video playback processing system based on pupil position recognition, characterized in that, The method includes a processor and a memory that communicate with each other, the processor being configured to retrieve a computer program from the memory and to implement the method of any one of claims 1-4 by running the computer program.

Images