Cloud video processing method, system, and storage medium

Abstract

The application provides a cloud video processing method and system and a storage medium, and the method comprises the following steps: acquiring cloud video data uploaded by a user, determining a clip element and a clip template according to a clip requirement of the user, performing clip processing on the cloud video data according to the clip element and the clip template, and obtaining a clip cloud video; acquiring annotation information for the clip cloud video, performing annotation on the clip cloud video according to the annotation information, and obtaining an annotated cloud video; acquiring a modified cloud video of the annotated cloud video, performing video classification on the modified cloud video when the modified cloud video meets a film review requirement, and obtaining a video type; determining a video push plan according to the video type, and performing video push on the modified cloud video according to the video push plan. According to the application, the annotation information for the clip cloud video is acquired, the clip cloud video can be effectively annotated based on the annotation information, and the safety of subsequent cloud video push is improved.

Description

Technical Field

[0001] This invention relates to the field of cloud video technology, and in particular to a cloud video processing method, system and storage medium. Background Technology

[0002] Cloud video refers to video network platform services based on cloud computing business models. On the cloud platform, all video providers, agents, planning service providers, producers, industry associations, management agencies, industry media, legal structures, etc., are centrally integrated into a resource pool. These resources can be displayed and interact with each other, communicate on demand, and reach agreements, thereby reducing costs and improving efficiency.

[0003] In the current cloud video processing process, after users complete the editing of the cloud video, they usually push the video directly for playback. When there are illegal scenes in the cloud video, the pushed cloud video will be taken down directly, which reduces the security of cloud video push. Summary of the Invention

[0004] The purpose of this invention is to provide a cloud video processing method, system, and storage medium to solve the problem of low security in cloud video push in the prior art.

[0005] This invention is implemented as follows: a cloud video processing method, the method comprising:

[0006] Obtain cloud video data uploaded by the user, and obtain the user's editing requirements;

[0007] Based on the editing requirements, determine the editing elements and editing templates, and then edit the cloud video data according to the editing elements and editing templates to obtain the edited cloud video;

[0008] Obtain annotation information for the edited cloud video, and annotate the edited cloud video according to the annotation information to obtain an annotated cloud video;

[0009] Obtain the modified cloud video of the annotated cloud video, and when the modified cloud video meets the review requirements, classify the modified cloud video to obtain the video type;

[0010] A video push plan is determined based on the video type, and the modified cloud video is pushed according to the video push plan.

[0011] Preferably, after pushing the modified cloud video according to the video push plan, the method further includes:

[0012] If an encryption instruction is received for the modified cloud video, the modified cloud video is set as the cloud video to be encrypted, and keyframe images are determined based on the video frame comments in the cloud video to be encrypted.

[0013] The cloud video to be encrypted is segmented based on the keyframe image to obtain the segmented video, and the user gaze point corresponding to the keyframe image is obtained.

[0014] A gaze point sequence is generated based on the user gaze point, and the corresponding segmented video is scrambled based on the gaze point sequence to obtain a scrambled video;

[0015] Obtain the number of likes and copies of the video frame comments, and determine the most popular comments in the video frame comments based on the number of likes and copies;

[0016] The keyframe image and the popular comments are input into a pre-trained large model for feature extraction to obtain image features and text features, and a feature key is generated based on the keyframe image, the image features, and the text features.

[0017] The scrambled video is encrypted using the specified feature key to obtain a cloud-encrypted video.

[0018] Preferably, generating a feature key based on the keyframe image, the image features, and the text features includes:

[0019] Connect the user gaze points in adjacent keyframe images to obtain gaze point line segments, and obtain the angle between the gaze point line segments and the horizontal direction to obtain the motion angle;

[0020] The image features and text features are vectorized to obtain image feature vectors and text feature vectors, and the length of the gaze point line segment is obtained to obtain the motion distance;

[0021] The feature key is generated based on the motion distance, the motion angle, the image feature vector, and the text feature vector.

[0022] Preferably, the formula used to generate the feature key based on the motion distance, the motion angle, the image feature vector, and the text feature vector includes:

[0023]

[0024] in, K The key is defined as a feature key, where SHA stands for cryptographic hash function, which is used to convert the input into a 256-bit binary key. I Represents the image feature vector, TThe text feature vector is represented by Hash, which represents the feature hash function. This indicates rounding down. ⊕ represents the modulo operation, ⊕ represents the XOR operation, Sd represents the cosine of the motion angle, ES represents the number of entities in the trending comment, Hh represents the sine of the motion angle, Sa represents the tangent of the motion angle, A represents the number of likes on the trending comment, It represents the motion distance, St represents the number of copies of the trending comment, Sim( I , T The similarity between the image feature vector and the text feature vector is represented by ).

[0025] Preferably, after encrypting the scrambled video according to the feature key to obtain the cloud-encrypted video, the method further includes:

[0026] If a data transmission instruction is received, the keyframe image is marked with a gaze point based on the user's gaze point to obtain a gaze point marked image.

[0027] The cloud-encrypted video, the viewpoint marker image, and the trending comments are combined to obtain cloud-encrypted video data, and the cloud-encrypted video data is transmitted according to the output address in the data transmission instruction.

[0028] Preferably, obtaining the user gaze point corresponding to the keyframe image includes:

[0029] Obtain the user's eye image corresponding to the keyframe image, and perform pupil localization on the user's eye image to obtain the pupil coordinates;

[0030] The gaze point coordinates are determined based on the pupil coordinates, and the keyframe image is divided into a grid to obtain an image grid.

[0031] The number of gaze point coordinates in the image grid is obtained to get the number of gazes, and the image grid corresponding to the maximum number of gazes is determined as the key grid;

[0032] The user's gaze point is obtained by averaging the coordinates of the gaze points in the key grid.

[0033] Another objective of this invention is to provide a cloud video processing system, the system comprising:

[0034] The demand acquisition module is used to acquire cloud video data uploaded by users and to acquire the editing requirements of the users.

[0035] The editing module is used to determine editing elements and editing templates according to the editing requirements, and to perform editing processing on the cloud video data according to the editing elements and the editing templates to obtain the edited cloud video;

[0036] The video annotation module is used to obtain annotation information for the edited cloud video, and to annotate the edited cloud video according to the annotation information to obtain an annotated cloud video.

[0037] The video classification module is used to obtain the modified cloud video of the annotation cloud video, and when the modified cloud video meets the review requirements, to classify the modified cloud video to obtain the video type.

[0038] The video push module is used to determine a video push plan based on the video type, and to push the modified cloud video according to the video push plan.

[0039] In this embodiment of the invention, by acquiring the user's editing requirements, editing elements and editing templates can be automatically determined based on the editing requirements. Based on the editing elements and editing templates, cloud video data can be effectively edited to obtain edited cloud videos. By acquiring annotation information for the edited cloud videos, annotations can be effectively added to the edited cloud videos, preventing the low security of cloud video push caused by direct cloud video push and improving the accuracy of cloud video data processing. Attached Figure Description

[0040] Figure 1 This is a flowchart of the cloud video processing method provided in the first embodiment of the present invention;

[0041] Figure 2 This is a flowchart of the cloud video processing method provided in the second embodiment of the present invention;

[0042] Figure 3 This is a schematic diagram of the cloud video processing system provided in the third embodiment of the present invention;

[0043] Figure 4 This is a schematic diagram of the structure of the terminal device provided in the fourth embodiment of the present invention. Detailed Implementation

[0044] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0045] To illustrate the technical solution described in this invention, specific embodiments are described below.

[0046] Example 1

[0047] Please see Figure 1 This is a flowchart of a cloud video processing method provided in the second embodiment of the present invention. This cloud video processing method can be applied to any device or system, and includes the following steps:

[0048] Step S1: Obtain cloud video data uploaded by the user and obtain the user's editing requirements;

[0049] Step S2: Determine the editing elements and editing templates according to the editing requirements, and perform editing processing on the cloud video data according to the editing elements and editing templates to obtain the edited cloud video;

[0050] Step S3: Obtain annotation information for the edited cloud video, and annotate the edited cloud video according to the annotation information to obtain an annotated cloud video;

[0051] Step S4: Obtain the modified cloud video of the annotated cloud video, and when the modified cloud video meets the review requirements, classify the modified cloud video to obtain the video type;

[0052] Step S5: Determine a video push plan based on the video type, and push the modified cloud video according to the video push plan.

[0053] In this embodiment, by acquiring the user's editing requirements, editing elements and editing templates can be automatically determined based on the editing requirements. Based on the editing elements and editing templates, cloud video data can be effectively edited to obtain edited cloud videos. By acquiring annotation information for the edited cloud videos, annotations can be effectively added to the edited cloud videos, preventing the low security of cloud video push caused by direct cloud video push and improving the accuracy of cloud video data processing.

[0054] Example 2

[0055] Please see Figure 2 This is a flowchart of a cloud video processing method provided in the second embodiment of the present invention. This cloud video processing method can be applied to any device or system, and includes the following steps:

[0056] Step S10: If an encryption instruction for the modified cloud video is received, the modified cloud video is set as the cloud video to be encrypted, and key frame images are determined based on the video frame comments in the cloud video to be encrypted.

[0057] Specifically, the number of comments for each video frame in the cloud video to be encrypted is obtained, and the number of comments for each frame is compared with a number threshold. If the number of comments is greater than the number threshold, the frame image corresponding to the number of comments is determined as the key frame image. The number threshold can be set according to the requirements.

[0058] Preferably, in this step, the duration interval between adjacent keyframe images is obtained. If the duration interval is less than the duration threshold, the frame images with fewer comments in the adjacent keyframe images are deleted until the duration interval between adjacent keyframe images is greater than the duration threshold. The duration threshold can be set according to the requirements.

[0059] Optionally, if there are no video frame comments in the encrypted cloud video, an asymmetric encryption algorithm is used to encrypt the encrypted cloud video.

[0060] Step S20: Perform video segmentation on the cloud video to be encrypted based on the keyframe image to obtain the segmented video, and obtain the user gaze point corresponding to the keyframe image.

[0061] Optionally, obtaining the user gaze point corresponding to the keyframe image includes:

[0062] Obtain the user's eye image corresponding to the keyframe image, and perform pupil localization on the user's eye image to obtain pupil coordinates; wherein, the user's eye image is the captured image of the corresponding user's eyes when viewing the keyframe image;

[0063] The gaze point coordinates are determined based on the pupil coordinates, and the keyframe image is divided into a grid to obtain an image grid; wherein, the keyframe image is divided into a grid according to a preset grid shape to obtain an image grid, and the preset grid image can be set according to requirements;

[0064] The number of gaze coordinates in the image grid is obtained to get the number of gazes, and the image grid corresponding to the maximum number of gazes is determined as the key grid; wherein, by obtaining the number of gaze coordinates in the image grid, the number of gaze points falling in the image grid is obtained;

[0065] The user's gaze point is obtained by averaging the coordinates of the gaze points in the key grid.

[0066] Step S30: Generate a gaze point sequence based on the user gaze points, and scramble the corresponding segmented video according to the gaze point sequence to obtain a scrambled video;

[0067] Specifically, the horizontal and vertical coordinates of the user's gaze point are converted into binary to obtain a horizontal sequence and a vertical sequence. The horizontal and vertical sequences are then combined to obtain a gaze point sequence. The corresponding segmented video is then scrambled using the gaze point sequence to achieve the effect of encrypting the image based on the gaze point.

[0068] Step S40: Obtain the number of likes and copies of the video frame comments, and determine the most popular comments in the video frame comments based on the number of likes and copies.

[0069] The number of likes refers to the number of times other users like the video frame comment, and the number of copies refers to the number of identical video frame comments in the cloud video to be encrypted. The number of likes and copies are weighted to obtain a weighted value, and the video frame comment corresponding to the maximum weighted value is determined as the popular comment. The weighting coefficients for the number of likes and copies during the weighting process can be set according to requirements. Preferably, if the number of characters in the video frame comment is less than the character threshold, the popular comment is re-determined. The character threshold can be set according to requirements.

[0070] Optionally, in this step, if the video frame comments do not have any likes or copies, an asymmetric encryption algorithm is used to encrypt the cloud video to be encrypted.

[0071] Step S50: Input the keyframe image and the popular comments into the pre-trained large model for feature extraction to obtain image features and text features, and generate a feature key based on the keyframe image, the image features and the text features;

[0072] Before inputting the keyframe images and trending comments into the pre-trained large model for feature extraction, the process also includes:

[0073] The model samples are input into the large model for feature extraction to obtain image sample features and text sample features. The model loss is calculated based on the image sample features and text sample features. The parameters of the large model are updated based on the model loss until the large model converges, resulting in the pre-trained large model.

[0074] Optionally, generating a feature key based on the keyframe image, the image features, and the text features includes:

[0075] Connect the user gaze points in adjacent keyframe images to obtain gaze point line segments, and obtain the angle between the gaze point line segments and the horizontal direction to obtain the motion angle;

[0076] The image features and text features are vectorized to obtain image feature vectors and text feature vectors, and the length of the gaze point line segment is obtained to obtain the motion distance;

[0077] The feature key is generated based on the motion distance, the motion angle, the image feature vector, and the text feature vector.

[0078] Furthermore, the formula used to generate the feature key based on the motion distance, the motion angle, the image feature vector, and the text feature vector includes:

[0079]

[0080] in, K The key is defined as a feature key, where SHA stands for cryptographic hash function, which is used to convert the input into a 256-bit binary key. I Represents the image feature vector, T The text feature vector is represented by Hash, which represents the feature hash function. This indicates rounding down. ⊕ represents the modulo operation, ⊕ represents the XOR operation, Sd represents the cosine of the motion angle, ES represents the number of entities in the trending comment, Hh represents the sine of the motion angle, Sa represents the tangent of the motion angle, A represents the number of likes on the trending comment, It represents the motion distance, St represents the number of copies of the trending comment, Sim( I , T The similarity between the image feature vector and the text feature vector is represented by ).

[0081] Step S60: Encrypt the scrambled video according to the feature key to obtain a cloud-encrypted video;

[0082] Specifically, the scrambled video is automatically encrypted using a feature key, achieving the effect of encrypting video data based on image and text features.

[0083] Optionally, after encrypting the scrambled video according to the feature key to obtain the cloud-encrypted video, the method further includes:

[0084] If a data transmission instruction is received, the keyframe image is marked with a gaze point based on the user's gaze point to obtain a gaze point marked image.

[0085] The cloud encrypted video, the viewpoint marker image, and the trending comments are combined to obtain cloud video encrypted data, and the cloud video encrypted data is transmitted according to the output address in the data transmission instruction;

[0086] The receiving end inputs viewpoint-marked images and trending comments into a pre-trained large model for feature extraction, obtaining image features and text features. Based on keyframe images, image features, and text features, a feature key is generated. The cloud-encrypted video is decoded based on the feature key to obtain a scrambled video. A gaze point sequence is generated based on the viewpoint-marked images. The scrambled video is then de-scrambled based on the gaze point sequence to obtain the cloud video to be encrypted.

[0087] Furthermore, before encrypting the scrambled video based on the feature key, the method further includes:

[0088] The image feature vector and the text feature vector are combined to obtain a combined vector, and a feature matrix is ​​constructed based on the combined vector;

[0089] Calculate the sum of adjacent elements in the feature matrix, and calculate the difference between the sums.

[0090] Calculate the product of the differences between elements along the diagonal direction to obtain the element-wise product, and calculate the sum of the element-wise products to obtain the product sum;

[0091] The diffusion coefficient is obtained by performing a modulo operation on the product sum, and adjacent pixels within the scrambled image in the scrambled video are combined to obtain a pixel combination image;

[0092] The diffusion coefficient is multiplied by each pixel combination in the pixel combination image to obtain the diffusion image, and the diffusion images are combined to obtain the diffusion video;

[0093] The distributed video is encrypted using the feature key to obtain the cloud-encrypted video;

[0094] The receiving end inputs viewpoint-marked images and trending comments into a pre-trained large model for feature extraction, obtaining image and text features. Based on keyframe images, image features, and text features, a feature key is generated. The cloud-encrypted video is decoded based on the feature key to obtain a diffused video. A diffusion coefficient is generated based on image and text feature vectors. The diffused video is then de-diffused based on the diffusion coefficient to obtain a scrambled video. A gaze point sequence is generated based on the viewpoint-marked images. The scrambled video is then de-scrambled based on the gaze point sequence to obtain the cloud video to be encrypted.

[0095] In this embodiment, keyframe images in the cloud video to be encrypted can be effectively determined by video frame comments. By obtaining the user gaze points corresponding to the keyframe images, a gaze point sequence can be effectively generated. Based on the gaze point sequence, the segmented video can be effectively scrambled to achieve the effect of encrypting the image based on gaze points. By inputting the keyframe images and popular comments into a pre-trained large model for feature extraction, image features and text features can be effectively obtained. Based on the keyframe images, image features, and text features, a feature key can be effectively generated. The scrambled video is automatically encrypted using the feature key to achieve the effect of encrypting video data based on image and text features. In this embodiment, the user does not need to manually input the key. Cloud video encryption is automatically performed based on gaze points and image and text features, improving the user experience.

[0096] Example 3

[0097] Please see Figure 3 This is a schematic diagram of the structure of the cloud video processing system 100 provided in the third embodiment of the present invention, including:

[0098] The requirement acquisition module 10 is used to acquire cloud video data uploaded by users and to acquire the editing requirements of the users.

[0099] The editing module 11 is used to determine the editing elements and editing templates according to the editing requirements, and to perform editing processing on the cloud video data according to the editing elements and the editing templates to obtain the edited cloud video.

[0100] The video annotation module 12 is used to obtain annotation information for the edited cloud video, and to annotate the edited cloud video according to the annotation information to obtain an annotated cloud video.

[0101] The video classification module 13 is used to obtain the modified cloud video of the annotation cloud video, and when the modified cloud video meets the review requirements, to classify the modified cloud video to obtain the video type.

[0102] The video push module 14 is used to determine a video push plan based on the video type, and to push the modified cloud video according to the video push plan.

[0103] Preferably, in this embodiment, the cloud video processing system 100 further includes:

[0104] The keyframe module is used to, if an encryption instruction for the modified cloud video is received, set the modified cloud video as a cloud video to be encrypted, and determine keyframe images based on video frame comments in the cloud video to be encrypted.

[0105] The video segmentation module is used to segment the cloud video to be encrypted based on the keyframe images, obtain the segmented video, and acquire the user gaze points corresponding to the keyframe images.

[0106] Optionally, the video segmentation module is further configured to: acquire the user's eye image corresponding to the keyframe image, and perform pupil localization on the user's eye image to obtain pupil coordinates;

[0107] The gaze point coordinates are determined based on the pupil coordinates, and the keyframe image is divided into a grid to obtain an image grid.

[0108] The number of gaze point coordinates in the image grid is obtained to get the number of gazes, and the image grid corresponding to the maximum number of gazes is determined as the key grid;

[0109] The user's gaze point is obtained by averaging the coordinates of the gaze points in the key grid.

[0110] The image scrambling module is used to generate a gaze point sequence based on the user's gaze point, and to scramble the corresponding segmented video based on the gaze point sequence to obtain a scrambled video.

[0111] The popularity determination module is used to obtain the number of likes and copies of the comments in the video frame, and determine the popular comments in the video frame comments based on the number of likes and copies.

[0112] The key generation module is used to input the keyframe image and the popular comments into a pre-trained large model for feature extraction, obtain image features and text features, and generate a feature key based on the keyframe image, the image features and the text features.

[0113] Optionally, the key generation module is further configured to: connect user gaze points in adjacent keyframe images to obtain gaze point line segments, and obtain the angle between the gaze point line segments and the horizontal direction to obtain the motion angle;

[0114] The image features and text features are vectorized to obtain image feature vectors and text feature vectors, and the length of the gaze point line segment is obtained to obtain the motion distance;

[0115] The feature key is generated based on the motion distance, the motion angle, the image feature vector, and the text feature vector.

[0116] Furthermore, the formula used to generate the feature key based on the motion distance, the motion angle, the image feature vector, and the text feature vector includes:

[0117]

[0118] in, K The key is defined as a feature key, where SHA stands for cryptographic hash function, which is used to convert the input into a 256-bit binary key. I Represents the image feature vector, T The text feature vector is represented by Hash, which represents the feature hash function. This indicates rounding down. ⊕ represents the modulo operation, ⊕ represents the XOR operation, Sd represents the cosine of the motion angle, ES represents the number of entities in the trending comment, Hh represents the sine of the motion angle, Sa represents the tangent of the motion angle, A represents the number of likes on the trending comment, It represents the motion distance, St represents the number of copies of the trending comment, Sim( I , T The similarity between the image feature vector and the text feature vector is represented by ).

[0119] The video encryption module is used to encrypt the scrambled video according to the feature key to obtain a cloud-encrypted video.

[0120] Optionally, the video encryption module is further configured to: if a data transmission instruction is received, mark the keyframe image with gaze points according to the user's gaze points to obtain a gaze point marked image;

[0121] The cloud-encrypted video, the viewpoint marker image, and the trending comments are combined to obtain cloud-encrypted video data, and the cloud-encrypted video data is transmitted according to the output address in the data transmission instruction.

[0122] In this embodiment, keyframe images in the cloud video to be encrypted can be effectively determined by video frame comments. By obtaining the user gaze points corresponding to the keyframe images, a gaze point sequence can be effectively generated. Based on the gaze point sequence, the segmented video can be effectively scrambled to achieve the effect of encrypting the image based on gaze points. By inputting the keyframe images and popular comments into a pre-trained large model for feature extraction, image features and text features can be effectively obtained. Based on the keyframe images, image features, and text features, a feature key can be effectively generated. The scrambled video is automatically encrypted using the feature key to achieve the effect of encrypting video data based on image and text features. In this embodiment, the user does not need to manually input the key. Cloud video encryption is automatically performed based on gaze points and image and text features, improving the user experience.

[0123] Example 4

[0124] Figure 4This is a structural block diagram of a terminal device 2 provided in the fourth embodiment of this application. For example... Figure 4 As shown, the terminal device 2 in this embodiment includes: a processor 20, a memory 21, and a computer program 22 stored in the memory 21 and executable on the processor 20, such as a program for a cloud video processing method. When the processor 20 executes the computer program 22, it implements the steps in the various embodiments of the above-described cloud video processing methods.

[0125] For example, the computer program 22 may be divided into one or more modules, which are stored in the memory 21 and executed by the processor 20 to complete this application. The one or more modules may be a series of computer program instruction segments capable of performing specific functions, which describe the execution process of the computer program 22 in the terminal device 2. The terminal device may include, but is not limited to, the processor 20 and the memory 21.

[0126] The processor 20 may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or any conventional processor.

[0127] The memory 21 can be an internal storage unit of the terminal device 2, such as a hard drive or memory of the terminal device 2. The memory 21 can also be an external storage device of the terminal device 2, such as a plug-in hard drive, Smart Media Card (SMC), Secure Digital (SD) card, or Flash Card equipped on the terminal device 2. Furthermore, the memory 21 can include both internal and external storage units of the terminal device 2. The memory 21 is used to store the computer program and other programs and data required by the terminal device. The memory 21 can also be used to temporarily store data that has been output or will be output.

[0128] Furthermore, the functional modules in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0129] If an integrated module is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. This computer-readable storage medium can be non-volatile or volatile. Based on this understanding, all or part of the processes in the methods of the above embodiments can also be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. The computer-readable storage medium can include: any entity or device capable of carrying computer program code, recording media, USB flash drives, portable hard drives, magnetic disks, optical disks, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signals, telecommunication signals, and software distribution media, etc. It should be noted that the contents of a computer-readable storage medium may be appropriately added to or subtracted from the contents as required by the legislation and patent practice in a jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, a computer-readable storage medium may not include electrical carrier signals and telecommunication signals.

[0130] The above-described embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application, and should all be included within the protection scope of this application.

Claims

1. A cloud video processing method, characterized in that, The method includes: Obtain cloud video data uploaded by the user, and obtain the user's editing requirements; Based on the editing requirements, determine the editing elements and editing templates, and then edit the cloud video data according to the editing elements and editing templates to obtain the edited cloud video; Obtain annotation information for the edited cloud video, and annotate the edited cloud video according to the annotation information to obtain an annotated cloud video; Obtain the modified cloud video of the annotated cloud video, and when the modified cloud video meets the review requirements, classify the modified cloud video to obtain the video type; A video push plan is determined based on the video type, and the modified cloud video is pushed according to the video push plan; After pushing the modified cloud video according to the video push plan, the method further includes: If an encryption instruction is received for the modified cloud video, the modified cloud video is set as the cloud video to be encrypted, and keyframe images are determined based on the video frame comments in the cloud video to be encrypted. The cloud video to be encrypted is segmented based on the keyframe image to obtain the segmented video, and the user gaze point corresponding to the keyframe image is obtained. A gaze point sequence is generated based on the user gaze point, and the corresponding segmented video is scrambled based on the gaze point sequence to obtain a scrambled video; Obtain the number of likes and copies of the video frame comments, and determine the most popular comments in the video frame comments based on the number of likes and copies; The keyframe image and the popular comments are input into a pre-trained large model for feature extraction to obtain image features and text features, and a feature key is generated based on the keyframe image, the image features, and the text features. The scrambled video is encrypted using the specified feature key to obtain a cloud-encrypted video. 2.The cloud video processing method of claim 1, wherein, Generating a feature key based on the keyframe image, the image features, and the text features includes: Connect the user gaze points in adjacent keyframe images to obtain gaze point line segments, and obtain the angle between the gaze point line segments and the horizontal direction to obtain the motion angle; The image features and text features are vectorized to obtain image feature vectors and text feature vectors, and the length of the gaze point line segment is obtained to obtain the motion distance; The feature key is generated based on the motion distance, the motion angle, the image feature vector, and the text feature vector. 3.The cloud video processing method of claim 2, wherein, The formula used to generate the feature key based on the motion distance, the motion angle, the image feature vector, and the text feature vector includes: in, K The key is defined as a feature key, where SHA stands for cryptographic hash function, which is used to convert the input into a 256-bit binary key. I Represents the image feature vector, T The text feature vector is represented by Hash, which represents the feature hash function. This indicates rounding down. ⊕ represents the modulo operation, ⊕ represents the XOR operation, Sd represents the cosine of the motion angle, ES represents the number of entities in the trending comment, Hh represents the sine of the motion angle, Sa represents the tangent of the motion angle, A represents the number of likes on the trending comment, It represents the motion distance, St represents the number of copies of the trending comment, Sim( I , T The similarity between the image feature vector and the text feature vector is represented by ).

4. The cloud video processing method as described in claim 1, characterized in that, After encrypting the scrambled video using the specified feature key to obtain the cloud-encrypted video, the process further includes: If a data transmission instruction is received, the keyframe image is marked with a gaze point based on the user's gaze point to obtain a gaze point marked image. The cloud-encrypted video, the viewpoint marker image, and the trending comments are combined to obtain cloud-encrypted video data, and the cloud-encrypted video data is transmitted according to the output address in the data transmission instruction.

5. The cloud video processing method as described in claim 1, characterized in that, Obtaining the user gaze point corresponding to the keyframe image includes: Obtain the user's eye image corresponding to the keyframe image, and perform pupil localization on the user's eye image to obtain the pupil coordinates; The gaze point coordinates are determined based on the pupil coordinates, and the keyframe image is divided into a grid to obtain an image grid. The number of gaze point coordinates in the image grid is obtained to get the number of gazes, and the image grid corresponding to the maximum number of gazes is determined as the key grid; The user's gaze point is obtained by averaging the coordinates of the gaze points in the key grid.

6. A cloud video processing system, characterized in that, The system includes: The demand acquisition module is used to acquire cloud video data uploaded by users and to acquire the editing requirements of the users. The editing module is used to determine editing elements and editing templates according to the editing requirements, and to perform editing processing on the cloud video data according to the editing elements and the editing templates to obtain the edited cloud video; The video annotation module is used to obtain annotation information for the edited cloud video, and to annotate the edited cloud video according to the annotation information to obtain an annotated cloud video. The video classification module is used to obtain the modified cloud video of the annotation cloud video, and when the modified cloud video meets the review requirements, to classify the modified cloud video to obtain the video type. The video push module is used to determine a video push plan based on the video type, and to push the modified cloud video according to the video push plan; The keyframe module is used to acquire the cloud video to be encrypted and determine the keyframe image based on the video frame comments in the cloud video to be encrypted. The video segmentation module is used to segment the cloud video to be encrypted based on the keyframe image to obtain the segmented video and to obtain the user gaze point corresponding to the keyframe image. The image scrambling module is used to generate a gaze point sequence based on the user's gaze point, and to scramble the corresponding segmented video based on the gaze point sequence to obtain a scrambled video. A popularity determination module is used to obtain the number of likes and copies of the comments in the video frame, and to determine the popular comments in the video frame comments based on the number of likes and copies. The key generation module is used to input the keyframe image and the popular comments into a pre-trained large model for feature extraction, to obtain image features and text features, and to generate a feature key based on the keyframe image, the image features and the text features; The video encryption module is used to encrypt the scrambled video according to the feature key to obtain a cloud-encrypted video.

7. The cloud video processing system as described in claim 6, characterized in that, The key generation module is also used for: Connect the user gaze points in adjacent keyframe images to obtain gaze point line segments, and obtain the angle between the gaze point line segments and the horizontal direction to obtain the motion angle; The image features and text features are vectorized to obtain image feature vectors and text feature vectors, and the length of the gaze point line segment is obtained to obtain the motion distance; The feature key is generated based on the motion distance, the motion angle, the image feature vector, and the text feature vector.

8. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by a processor, it implements the steps of the method as described in any one of claims 1 to 5.

Images