CN110677701A - Video stream recommendation method, electronic device and storage medium - Google Patents
Video stream recommendation method, electronic device and storage medium Download PDFInfo
- Publication number
- CN110677701A CN110677701A CN201911142687.1A CN201911142687A CN110677701A CN 110677701 A CN110677701 A CN 110677701A CN 201911142687 A CN201911142687 A CN 201911142687A CN 110677701 A CN110677701 A CN 110677701A
- Authority
- CN
- China
- Prior art keywords
- video stream
- user
- candidate
- candidate video
- preference
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N21/00—Selective content distribution, e.g. interactive television or video on demand [VOD]
- H04N21/20—Servers specifically adapted for the distribution of content, e.g. VOD servers; Operations thereof
- H04N21/25—Management operations performed by the server for facilitating the content distribution or administrating data related to end-users or client devices, e.g. end-user or client device authentication, learning user preferences for recommending movies
- H04N21/251—Learning process for intelligent management, e.g. learning user preferences for recommending movies
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N21/00—Selective content distribution, e.g. interactive television or video on demand [VOD]
- H04N21/20—Servers specifically adapted for the distribution of content, e.g. VOD servers; Operations thereof
- H04N21/25—Management operations performed by the server for facilitating the content distribution or administrating data related to end-users or client devices, e.g. end-user or client device authentication, learning user preferences for recommending movies
- H04N21/258—Client or end-user data management, e.g. managing client capabilities, user preferences or demographics, processing of multiple end-users preferences to derive collaborative data
- H04N21/25866—Management of end-user data
- H04N21/25891—Management of end-user data being end-user preferences
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N21/00—Selective content distribution, e.g. interactive television or video on demand [VOD]
- H04N21/40—Client devices specifically adapted for the reception of or interaction with content, e.g. set-top-box [STB]; Operations thereof
- H04N21/45—Management operations performed by the client for facilitating the reception of or the interaction with the content or administrating data related to the end-user or to the client device itself, e.g. learning user preferences for recommending movies, resolving scheduling conflicts
- H04N21/4508—Management of client data or end-user data
- H04N21/4532—Management of client data or end-user data involving end-user characteristics, e.g. viewer profile, preferences
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N21/00—Selective content distribution, e.g. interactive television or video on demand [VOD]
- H04N21/40—Client devices specifically adapted for the reception of or interaction with content, e.g. set-top-box [STB]; Operations thereof
- H04N21/45—Management operations performed by the client for facilitating the reception of or the interaction with the content or administrating data related to the end-user or to the client device itself, e.g. learning user preferences for recommending movies, resolving scheduling conflicts
- H04N21/466—Learning process for intelligent management, e.g. learning user preferences for recommending movies
- H04N21/4662—Learning process for intelligent management, e.g. learning user preferences for recommending movies characterized by learning algorithms
- H04N21/4666—Learning process for intelligent management, e.g. learning user preferences for recommending movies characterized by learning algorithms using neural networks, e.g. processing the feedback provided by the user
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N21/00—Selective content distribution, e.g. interactive television or video on demand [VOD]
- H04N21/40—Client devices specifically adapted for the reception of or interaction with content, e.g. set-top-box [STB]; Operations thereof
- H04N21/45—Management operations performed by the client for facilitating the reception of or the interaction with the content or administrating data related to the end-user or to the client device itself, e.g. learning user preferences for recommending movies, resolving scheduling conflicts
- H04N21/466—Learning process for intelligent management, e.g. learning user preferences for recommending movies
- H04N21/4667—Processing of monitored end-user data, e.g. trend analysis based on the log file of viewer selections
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N21/00—Selective content distribution, e.g. interactive television or video on demand [VOD]
- H04N21/40—Client devices specifically adapted for the reception of or interaction with content, e.g. set-top-box [STB]; Operations thereof
- H04N21/45—Management operations performed by the client for facilitating the reception of or the interaction with the content or administrating data related to the end-user or to the client device itself, e.g. learning user preferences for recommending movies, resolving scheduling conflicts
- H04N21/466—Learning process for intelligent management, e.g. learning user preferences for recommending movies
- H04N21/4668—Learning process for intelligent management, e.g. learning user preferences for recommending movies for recommending content, e.g. movies
Landscapes
- Engineering & Computer Science (AREA)
- Databases & Information Systems (AREA)
- Multimedia (AREA)
- Signal Processing (AREA)
- Computer Graphics (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Computation (AREA)
- Computing Systems (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The embodiment of the invention provides a video stream recommendation method, electronic equipment and a storage medium, wherein the method comprises the following steps: determining a plurality of candidate video streams; determining a user preference for each of the candidate video streams based on the user profile; and recommending the video streams based on the user preference of each candidate video stream. According to the method, the electronic device and the storage medium provided by the embodiment of the invention, personalized video stream recommendation is carried out based on the user preference of each candidate video stream, so that the video stream meeting the personal preference can be improved for the user, and the user experience is optimized. In addition, the video stream can be pushed to the user without the synthesis of the director, different video streams can be pushed for different users, and the resource waste is effectively avoided.
Description
Technical Field
The present invention relates to the field of video processing technologies, and in particular, to a video stream recommendation method, an electronic device, and a storage medium.
Background
With the increasing of internet network bandwidth, video on demand and live broadcast technologies are rapidly developed, and network video programs are more and more favored by people.
At present, when live scenes such as concerts, sports competitions, remote teaching, commercial propaganda, teleconferencing and the like are subjected to video live broadcasting, videos are usually recorded by a plurality of cameras, then a plurality of paths of video streams recorded by the plurality of cameras are combined into one path of video live broadcasting stream by a broadcasting guide platform, and then the video live broadcasting stream is uniformly transmitted to a client.
However, the video live stream synthesized by the director according to the director experience cannot meet the watching requirements of different users. In addition, although multiple video streams are recorded on site, the user can only view one of the video streams at the same time, which results in resource waste.
Disclosure of Invention
The embodiment of the invention provides a video stream recommendation method, electronic equipment and a storage medium, which are used for solving the problems that a video live stream synthesized by a director cannot meet the watching requirements of different users and resource waste is caused.
In a first aspect, an embodiment of the present invention provides a video stream recommendation method, including:
determining a plurality of candidate video streams;
determining a user preference for each of the candidate video streams based on the user profile;
and recommending the video streams based on the user preference of each candidate video stream.
Preferably, the user representation comprises a number of favorite video streams of the user;
correspondingly, the determining the user preference of each candidate video stream based on the user profile specifically includes:
determining, for any favorite video stream and any candidate video stream, a first number of users that favorite the any candidate video stream, and a second number of users that favorite the any favorite video stream and the any candidate video stream at the same time;
determining a similarity of the any favorite video stream and the any candidate video stream based on the first number of users and the second number of users;
determining a user preference for each of the candidate video streams based on the similarity of the candidate video stream to the preferred video stream.
Preferably, said user representation further comprises a user preference for each of said preferred video streams;
correspondingly, the determining the user preference of any candidate video stream based on the similarity between each preferred video stream and any candidate video stream specifically includes:
if the similarity between the favorite video stream and the candidate video stream is greater than a preset similarity threshold, determining that the favorite video stream is a reference video stream of the candidate video stream;
and weighting the user preference of each reference video stream by taking the similarity of each reference video stream and any candidate video stream as a weight to obtain the user preference of any candidate video stream.
Preferably, the determining the user preference of each candidate video stream based on the user profile specifically includes:
for any candidate video stream, inputting the user portrait and any candidate video stream into a preference prediction model, and outputting user preference;
the preference prediction model is obtained by training user preference of a sample user for a sample video stream based on a user image of the sample user, the sample video stream and the sample user.
Preferably, the preference prediction model is composed of a sparse feature layer, a dense feature layer and a sparse dense feature interaction layer;
wherein the sparse feature layer is used to extract user features from the user representation;
the dense feature layer is used for extracting video stream features from any candidate video stream;
the sparse dense feature interaction layer is configured to output the user preference based on the user feature and the video stream feature.
Preferably, the sparse dense feature interaction layer is represented by the following formula:
in the formula (I), the compound is shown in the specification,
for user preference, b and w are parameters obtained by pre-training, x is the video stream characteristic, v is the user characteristic, and upper represents rounding up.
Preferably, the recommending the video stream based on the user preference of each candidate video stream specifically includes:
performing video stream recommendation based on the universal score and the user preference of each candidate video stream;
wherein the universal score is determined based on recording parameters of the candidate video streams; the recording parameter includes at least one of a lens type, a scene type, a sound element type, a transition type, and a special effect type.
Preferably, the recommending the video stream based on the general score and the user preference of each candidate video stream specifically includes:
and recommending the video stream based on the first preset number of paths of candidate video streams with the highest user preference and the second preset number of paths of candidate video streams with the highest general score.
Preferably, the first preset number is greater than the second preset number;
in a second aspect, an embodiment of the present invention provides an electronic device, which includes a processor, a communication interface, a memory, and a bus, where the processor and the communication interface communicate with each other through the bus, and the processor may call logic instructions in the memory to perform the steps of the method as provided in the first aspect.
In a third aspect, an embodiment of the present invention provides a non-transitory computer readable storage medium, on which a computer program is stored, which when executed by a processor implements the steps of the method as provided in the first aspect.
According to the video stream recommendation method, the electronic device and the storage medium provided by the embodiment of the invention, personalized video stream recommendation is carried out based on the user preference of each candidate video stream, so that the video streams meeting the personal preference can be improved for the user, and the user experience is optimized. In addition, the video stream can be pushed to the user without the synthesis of the director, different video streams can be pushed for different users, and the resource waste is effectively avoided.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, and it is obvious that the drawings in the following description are some embodiments of the present invention, and those skilled in the art can also obtain other drawings according to the drawings without creative efforts.
Fig. 1 is a schematic flowchart of a video stream recommendation method according to an embodiment of the present invention;
fig. 2 is a flowchart illustrating a video stream recommendation method according to another embodiment of the present invention;
fig. 3 is a schematic structural diagram of a video stream recommendation apparatus according to an embodiment of the present invention;
fig. 4 is a schematic structural diagram of an electronic device according to an embodiment of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the embodiments of the present invention clearer, the drawings in the embodiments of the present invention will be combined to clearly and completely describe the technical solutions in the embodiments of the present invention, and it is obvious that the described embodiments are a part of the embodiments of the present invention, but not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
When video live broadcasting is performed, video recording is generally performed through a plurality of cameras, and then a director combines a plurality of video streams recorded by the plurality of cameras into one video live stream. However, the video live stream synthesized according to the director experience cannot meet the viewing requirements of different users, and causes resource waste. In contrast, the embodiment of the invention provides a video stream recommendation method, which can be applied to a live video scene to realize recommendation of multiple paths of video streams, and can also be applied to a video on demand scene to recommend video streams to a user. Fig. 1 is a schematic flowchart of a video stream recommendation method according to an embodiment of the present invention, and as shown in fig. 1, the video stream recommendation method includes:
in step 110, multiple candidate video streams are determined.
Here, the candidate video stream is a video stream to be recommended to the user. In a live video scene, when a user is watching a live program, the candidate video streams may be video streams shot by different cameras for the live program. Likewise, in a video-on-demand scenario, when a user characterizes a recorded video of a certain program, the candidate video streams may be video streams captured for that program by different cameras.
Specifically, the user portrait is used for outlining relevant information of the user, so that the purpose of associating the user requirement with product design is achieved, and the relevant information of the user can be user background, characteristics, character labels, behavior scenes and the like. In an embodiment of the invention, the user profile is predetermined, and the user profile is used for representing user preferences for different video streams. For example, the user profile may include a video stream preferred by the user, or may include the type or specific characteristics of the video stream preferred by the user, and the user profile may be obtained by analyzing the historical operation behavior of the user.
The user preference is used to measure the preference of the user to the candidate video stream, for example, the preference of the user to the video stream may be divided into 3 grades, which are general preference, comparative preference and extreme preference in turn. Where a user profile is known, a user preference for each candidate video stream may be determined based on the user profile. For example, the user profile includes user preferences of the user for different types of video streams, where the user preferences corresponding to the long-shot type video stream, the medium-shot type video stream, and the close-up type video stream are general preferences, comparative preferences, and extreme preferences, respectively. Assuming that, among the 4 candidate video streams, the candidate video stream A, B is a medium-scene type video stream, the candidate video stream C is a close-up type video stream, and the candidate video stream D is an original type video stream, the user preference of the candidate video stream A, B is relatively preferred, the user preference of the candidate video stream C is very preferred, and the user preference of the candidate video stream D is general preferred.
Specifically, after determining the user preference of each candidate video stream, the video stream may be pushed to the user according to the user preference of each candidate video stream. For example, only the candidate video streams with the user preference being very preferred are pushed to the user, and for example, each candidate video stream is arranged in the order from high to low of the user preference, and the first 3 candidate video streams are pushed to the user.
The method provided by the embodiment of the invention carries out personalized video stream recommendation based on the user preference of each candidate video stream, can improve the video stream meeting the personal preference for the user, and optimizes the user experience. In addition, the video stream can be pushed to the user without the synthesis of the director, different video streams can be pushed for different users, and the resource waste is effectively avoided.
Based on the above embodiment, in the method, the user representation includes several favorite video streams of the user.
Here, the favorite video stream is a video stream preferred by the user, and the favorite video stream may be a video stream collected in a favorite by the user, or a video stream whose viewing time exceeds a preset time threshold, and the like.
Correspondingly, step 120 specifically includes:
in step 121, for any candidate video stream, the similarity between each favorite video stream and the candidate video stream is determined.
Specifically, for any favorite video stream and any candidate video stream, the similarity of the two can be determined. Here, the similarity may be determined by the type or specific feature of the favorite video stream and the candidate video stream, and the more similar the type or specific feature is, the more similar the favorite video stream and the candidate video stream is; it can also be determined by a user population that likes the favorite video stream and the candidate video stream, the more similar the user population, the more similar the favorite video stream and the candidate video stream.
Step 122, determining the user preference of the candidate video stream based on the similarity of each preferred video stream and the candidate video stream.
For example, the average value of the similarity between any candidate video stream and each favorite video stream can be taken to measure whether the candidate video stream meets the user's favorite, and then the user's favorite of the candidate video stream is determined; for another example, a similarity threshold may be preset, and a similarity ratio exceeding the similarity threshold is calculated to measure whether the candidate video stream meets the user's preference, so as to determine the user's preference of the candidate video stream.
Based on any of the above embodiments, in the method, step 121 specifically includes: determining a first number of users of any favorite video stream and any candidate video stream, and a second number of users of both favorite video stream and the candidate video stream; the similarity of the favorite video stream and the candidate video stream is determined based on the first number of users and the second number of users.
Specifically, the similarity w for any candidate video stream i and any favorite video stream jijThe calculation can be obtained by the following formula:
where n (i) is the first number of users, which is used to characterize the preference i, and n (ij) is the second number of users, which is used to characterize the number of users who prefer i and j simultaneously.
Further, the first number of users n (i) may be obtained by obtaining a user set with the preference i, and counting the number of users included in the user set; the second number of users n (ij) may be obtained by obtaining a user set with a preference i and a user set with a preference j, determining an intersection of the two user sets, and further counting the number of users included in the intersection.
In any of the above embodiments, the method wherein the user representation further comprises user preferences for each preferred video stream. Here, the user preference of the favorite video stream is used to measure the user preference of the favorite video stream.
Correspondingly, step 122 specifically includes: if the similarity between any favorite video stream and the candidate video stream is greater than a preset similarity threshold, determining the favorite video stream as a reference video stream of the candidate video stream; and weighting the user preference of each reference video stream by taking the similarity of each reference video stream and the candidate video stream as a weight to obtain the user preference of the candidate video stream.
Specifically, the preset similarity threshold is a preset minimum similarity of the reference video stream corresponding to the favorite video stream as any candidate video stream. And aiming at any candidate video stream i, calculating the similarity between the candidate video stream and each favorite video stream of the user, and screening the favorite video stream with the similarity larger than a preset similarity threshold value from each favorite video stream of the user as a reference video stream of the candidate video stream i. Here, the reference video stream is a video stream used to measure the user preference of the candidate video stream.
User preference P for any one of the candidate video streams iuiCan be calculated by the following formula:
in the formula, PuiThe user preference of the user u for the candidate video stream i, K is the reference video stream, K belongs to S (u) ∩ S (i, K), S (u) is the preferred video set of the user u, S (i, K) is the similar video set of the candidate video stream i, wikIs the similarity of i and k, PukThe user preference of user u for reference video stream k. Further, S (i, K) is a video set with a similarity to the candidate video stream i greater than a preset similarity threshold.
Based on any of the above embodiments, in the method, step 120 specifically includes: inputting the user portrait and the candidate video stream into a preference prediction model aiming at any candidate video stream, and outputting user preference; the preference prediction model is obtained by training user preference of a sample user for a sample video stream based on a user image of the sample user, the sample video stream and the sample user.
Here, the preference prediction model is a model trained in advance, and is used to evaluate and output the user preference of the user for the candidate video stream based on the input user image and the candidate video stream. Before the prediction of the user preference is executed, a preference prediction model can be obtained through pre-training, and specifically, the preference prediction model can be obtained through training in the following way: first, a user representation of a large number of sample users, a large number of sample video streams are collected, and user preferences of the sample users for the sample video streams are determined. And then, training the initial model based on the user image of the sample user, the sample video stream and the user preference of the sample user to the sample video stream, thereby obtaining a preference prediction model. The initial model may be a single neural network model or a combination of multiple neural network models, and parameters in the initial model may be obtained by initialization or may be preset.
Further, let w be the feature vector of the sample video stream iiThe feature vector of the user portrait of sample user j is vjThe user preference of the sample user j to the sample video stream i is rijThen, the objective function for training the initial model is:
in the formula (I), the compound is shown in the specification,
for representing the output results obtained after inputting the sample video stream i and the user representation of the sample user j into the initial model,
the deviation of the user preference from the predicted result.
Based on any one of the above embodiments, in the method, the preference prediction model is composed of a sparse feature layer, a dense feature layer, and a sparse dense feature interaction layer; wherein the sparse feature layer is used for extracting user features from the user portrait; the dense feature layer is used for extracting video stream features from the candidate video stream; and the sparse dense feature interaction layer is used for outputting the user preference degree based on the user features and the video stream features.
Specifically, in the preference prediction model, it is necessary to perform feature extraction on the input user portrait and candidate video stream, respectively. And in consideration of the sparsity of the corresponding features of the user image and the density of the corresponding features of the candidate video stream, setting different feature layers for the user image and the candidate video stream to perform feature extraction during feature extraction. Here, the user image corresponds to a sparse feature layer, the sparse feature layer may be any network structure for performing sparse feature processing, the candidate video stream corresponds to a dense feature layer, and the dense feature layer may be a network structure for performing dense feature processing by any user, such as a multi-layer perceptron MLP.
For example, the sparse feature layer may be the following structure:
(a) EmbeddingBag (4,2, mode) first class may have 4 one-hot lengths
(b) EmbeddingBag (3,2, mode) second class # may have 3 one-hot lengths
(c) EmbeddingBag (2,2, mode) third class may have 2 one-hot lengths
The dense feature layer may be of the structure:
(a):Linear(in_features＝4,out_features＝3,bias＝True)
(b):ReLU()
(c):Linear(in_features＝3,out_features＝2,bias＝True)
(d) the output at the bottom of ReLU () # is the Relu function
And after the extraction of the user characteristics and the video stream characteristics is finished, the output of the user preference degree is realized through the sparse dense characteristic interaction layer. Further, the sparse dense feature interaction layer may be implemented by a prediction function as follows:
in the formula, user preference
Wherein +1 predicts that the user clicks on the candidate video stream, i.e., the user likes the candidate video stream, and-1 predicts that the user does not click on the candidate video stream, i.e., the user does not like the candidate video stream. b and w are parameters obtained by pre-training, x is the video stream characteristic of the candidate video stream, v is the user characteristic of the user portrait, and upper represents rounding-up.
The sparse dense feature interaction layer may be of the structure:
(a):Linear(in_features＝8,out_features＝4,bias＝True)
(b):ReLU()
(c):Linear(in_features＝4,out_features＝2,bias＝True)
(d):ReLU()
(e):Linear(in_features＝2,out_features＝1,bias＝True)
(f) sigmoid () top output is for Sigmoid function
If the user preference is judged only based on the historical behavior of the user and then the video stream is recommended based on the user preference, the user does not have the opportunity to select the video stream of a different type from the historical behavior, and although the user preference is met, the willingness of the user to explore a new type of video stream is also hindered. Based on any of the above embodiments, in the method, step 120 specifically includes: performing video stream recommendation based on the general score and the user preference of each candidate video stream; wherein the universal score is determined based on recording parameters of the candidate video streams; the recording parameter includes at least one of a shot type, a scene type, a sound element type, a transition type, and a special effect type.
Here, the lens type is used to characterize a type of an apparatus physical lens used for recording the candidate video stream, and the lens type may be at least one of a wide-angle lens, a standard lens and a narrow-angle lens, wherein the wide-angle lens is used to generate an image with a large foreground object, a small middle scene object and a smaller background object, the standard lens is used to generate an image with a field of view very close to a scene seen by human eyes, and the narrow-angle lens is used to reduce the field of view and enlarge a shot object in the background.
The scene type is a difference of a range size of the object in the camera finder caused by a difference of a distance between the camera and the object, and the scene type may be at least one of a long-distance scene, a full-view scene, a medium-distance scene, a short-distance scene and a close-up scene, and may specifically include at least one of a medium-distance scene, a double-shot, a triple-shot, a shoulder-crossing shot and a cross-shot. The medium-close shot refers to an image from the upper half of the subject to the head of the subject, the medium panorama refers to an image from above the knees of the subject, the double shot refers to placing two persons or objects in one picture, the triple shot refers to placing three persons or objects in one picture, the over-the-shoulder shot refers to shooting over the shoulders of the person close to the camera, and the cross shot refers to shooting by the camera among a plurality of persons in turn.
The sound element type is used to characterize sound aesthetic elements in the candidate video stream, and the sound element type may be at least one of environment, subject-background relationship, perspective, consistency, and energy. The environment refers to a sound element that is heard in the background of the primary sound source, can indicate the place of the occurrence of time or can suggest important clues of event feeling, and is helpful for creating the overall environment of the event. The subject-background relationship is to highlight the subject sound source over the overall background sound. Perspective is the matching of close-up sound with close-up view and far-out sound with long-range view. Consistency is to maintain the quality of sound when combining sounds shot multiple times, for example, in a news video stream, the recorded sound is always maintained. The energy refers to the energy of the picture matched with the sound with the same strength.
The transition type may include at least one of shot switching, eclipsing, wipe, and black, wherein shot switching refers to directly switching from one shot to another shot, eclipsing refers to gradual switching between shots, there is a brief overlap between two images, wipe refers to a transition gradual change of two shots, black and black refers to a gradual black change of images, or a gradual display from a black screen.
The special effect types may include at least one of overlay, embedding, matting, scribing, shrinking and enlarging, stretching, position and viewpoint, perspective, mosaic, poster art, and negative effects. Wherein, the superposition means that a picture from one video source is superposed on a picture from another video source; embedding means that a part of an image is cut out and filled with another image or a part of another image; matting refers to the application of a special color (usually blue or green) as a background to contrast a person or object from a foreground; the marking image means that the new image gradually replaces part or all of the image being played by a certain geometric shape; the zooming-out and zooming-in means that one picture is zoomed out or zoomed in, and the integrity and fixed aspect ratio of the picture are ensured; stretching refers to horizontally or vertically stretching the image; the position and viewpoint refer to placing the reduced image at any position of the picture; perspective refers to distortion operation on an image, so that a visual 3D effect is achieved; mosaic means that an image is fragmented into a plurality of discontinuous fragments; the poster drawing type is that the image is simplified into a plurality of single colors and brightness levels; the negative effect is to gradually change the brightness value of the image to a negative value.
The universal score is a score obtained by evaluating the candidate video stream based on the recording parameters of the candidate video stream. The generic score reflects candidate video stream evaluations unaffected by the particular user's preferences. Further, the corresponding relation between the recording parameters and the general scores can be preset, and after the recording parameters of the candidate video streams are determined, the general scores of the candidate video streams are determined based on the corresponding relation between the recording parameters and the general scores; the general scoring model for evaluating the general score may also be trained in advance, the recording parameters of the candidate video stream are input into the general scoring model, and the general score is output by the general scoring model, which is not specifically limited in the embodiment of the present invention.
After the general score of each candidate video stream is determined, a plurality of candidate video streams can be selected based on user preference, a plurality of candidate video streams are selected based on the general score, and the selected candidate video streams are all pushed to the user, so that conditions are provided for the user to explore different types of video streams while the user preference is met.
The method provided by the embodiment of the invention carries out video stream recommendation based on the general score on the basis of ensuring that the video stream recommendation meets the user preference, provides conditions for the user to explore different types of video streams, and further optimizes the user experience.
Based on any of the above embodiments, in the method, step 120 specifically includes: and recommending the video streams based on the first preset number of candidate video streams with the highest user preference and the second preset number of candidate video streams with the highest general score.
Here, a first preset number and a second preset number are preset, wherein the first preset number is used for selecting candidate video streams based on user preference, and the second preset number is used for selecting candidate video streams based on general scores. Assuming that the first preset number is 3 and the second preset number is 2, the candidate video streams can be sorted in the sequence from high to low according to the preference of the user, the first 3 candidate video streams are selected and pushed to the user, meanwhile, the candidate video streams are sorted in the sequence from high to low according to the comprehensive score, and the first two candidate video streams are selected and pushed to the user.
Based on any of the above embodiments, in the method, the first preset number is greater than the second preset number; correspondingly, step 120 specifically includes: constructing a user recommendation video stream set based on the candidate video streams with the highest user preference degrees on the first preset number of roads; taking the candidate video stream with the highest general score of the second preset number of paths as a general recommended video stream; replacing a second preset number of candidate video streams with the lowest user preference in the user recommended video stream set with the general recommended video stream; and pushing the user recommended video stream set to the user side.
Suppose that the user preferences and generic scores for candidate video streams are as set forth in the following table:
| candidate video streams | A | B | C | D | E | F | G | H | I |
| User preference | 60 | 79 | 92 | 30 | 10 | 88 | 50 | 22 | 13 |
| General score | 75 | 56 | 78 | 93 | 66 | 84 | 52 | 30 | 55 |
Assume that the first predetermined number is 3 and the second predetermined number is 1. The candidate video streams with the highest user preference for the first preset number of channels are C, F, B, respectively, and the user recommended video stream set includes C, F, B. And D is the candidate video stream with the highest score used by the paths with the second preset number, and D is the general recommended video stream. In the user recommended video stream set, the second preset number of candidate video streams with the lowest user preference degree is B, so that B in the user recommended video stream set is replaced by a general recommended video stream D, the updated user recommended video stream set comprises C, F, D, and C, F, D is pushed to the user side.
Based on any embodiment, in the method, the method for determining the universal score includes: inputting the recording parameters of the candidate video streams into a universal scoring model to obtain the universal scores of the candidate video streams output by the universal scoring model; the universal scoring model is obtained by training based on sample recording parameters of the sample video stream and the sample universal score.
Here, the general scoring model is a pre-trained model for evaluating and outputting a general score of the corresponding candidate video stream based on the input recording parameters. Before the method for determining the universal score is executed, the universal score model can be obtained through pre-training, and specifically, the universal score model can be obtained through training in the following way: firstly, a large number of sample video streams are collected, sample recording parameters in the sample video streams are extracted, meanwhile, manual scoring is carried out on the sample video streams, and sample general scores of the sample video streams are determined. And then, training the initial model based on the sample recording parameters of the sample video stream and the sample universal score to obtain a universal score model. The initial model may be a single neural network model or a combination of multiple neural network models, and parameters in the initial model may be obtained by initialization or may be preset.
Further, in the initial model used for training the general scoring model, the weights corresponding to the recording parameters are preset. By presetting the weight corresponding to each recording parameter in the initial model, the convergence of the initial model can be accelerated, and the training speed of the initial model is improved.
Based on any of the above embodiments, the initial weights set for the respective recording parameters in the initial model are as follows:
based on any of the above embodiments, fig. 2 is a schematic flow chart of a video stream recommendation method according to another embodiment of the present invention, as shown in fig. 2, in the field of live video broadcasting, when live video broadcasting is performed for live events such as concerts, sports matches, remote education, commercial promotions, and teleconferencing, there are 6 cameras A, B, C, D, E and F that record video simultaneously. In the embodiment of the present invention, the cameras are all configured with HDMI data lines and built-in network cards, and transmit the recorded candidate video streams A, B, C, D, E and F to the recommendation system module in a wired or wireless manner.
And after the 6 paths of candidate video streams are obtained, the recommending system module interacts with the user portrait model to obtain the favorite video streams of the user and the user preference degree of the user for the favorite video streams.
Then, for any candidate video stream i and any favorite video stream j, the recommender module calculates the similarity w between the candidate video stream i and the favorite video stream j based on the following equationij：
Where N (i) is the number of users who like i, and N (ij) is the number of users who like i and j simultaneously.
Then, the recommendation system module compares the similarity between any candidate video stream i and each favorite video stream j with a preset similarity threshold, and takes the favorite video stream j corresponding to the similarity larger than the preset similarity threshold as a reference video stream k of the candidate video stream i, where the reference video stream k is not only the favorite video stream of the user, but also similar to the candidate video stream i.
Next, the recommender module calculates the user preference P for the candidate video stream i based onui：
In the formula, PuiUsers for candidate video stream i for user uPreference, K being the reference video stream, K ∈ S (u) ∩ S (i, K), S (u) being the set of preferred videos of user u, S (i, K) being the set of similar videos of candidate video stream iikIs the similarity of i and k, PukThe user preference of user u for reference video stream k.
After determining the user preference of each path of candidate video stream, the recommendation system module sorts the 6 paths of candidate video streams according to the sequence of the user preference from high to low, and transmits the sorted 6 paths of candidate video streams to the recommendation result module. Assume here that the ordering results are A > B > C > F > E > D.
Then, the recommendation result module sends the recording parameters of each path of candidate video stream to the general strategy module, the general strategy module inputs the recording parameters into a general scoring model obtained by pre-training, and returns the general score of each path of candidate video stream output by the general scoring model to the recommendation result module.
And the recommendation result module returns the general scores of each candidate video stream to the recommendation result module after receiving the general scores. Assuming that the first preset number is 3 and the second preset number is 2, the first three candidate video streams ranked according to the user preference degree are A > B > C, the first candidate video stream obtained by ranking the 6 candidate video streams according to the sequence of the general scores from high to low is D, the recommendation result module replaces the last candidate video stream C in the A > B > C with D to obtain A > B > D, and the candidate video streams are recommended to the user side according to the sequence of A, B, D.
The method provided by the embodiment of the invention carries out personalized video stream recommendation based on the user preference of each candidate video stream, can improve the video stream meeting the personal preference for the user, and optimizes the user experience. In addition, the video stream can be pushed to the user without the synthesis of the director, different video streams can be pushed for different users, and the resource waste is effectively avoided. In addition, on the basis of ensuring that the video stream recommendation meets the user preferences, the video stream recommendation is carried out based on the general scores, conditions are provided for the user to explore different types of video streams, and the user experience is further optimized.
Based on any of the above embodiments, fig. 3 is a schematic structural diagram of a video stream recommendation apparatus according to an embodiment of the present invention, as shown in fig. 3, the apparatus includes a video stream determining unit 310, a user preference determining unit 320, and a recommending unit 330;
the video stream determining unit 310 is configured to determine multiple candidate video streams;
the user preference determining unit 320 is configured to determine a user preference for each of the candidate video streams based on the user profile;
the recommending unit 330 is configured to recommend video streams based on the user preference of each of the candidate video streams.
The device provided by the embodiment of the invention carries out personalized video stream recommendation based on the user preference of each candidate video stream, can improve the video stream meeting the personal preference for the user, and optimizes the user experience. In addition, the video stream can be pushed to the user without the synthesis of the director, different video streams can be pushed for different users, and the resource waste is effectively avoided.
According to any of the above embodiments, in the apparatus, the user representation includes several favorite video streams of the user;
correspondingly, the user preference determining unit 320 includes:
a similarity determining subunit, configured to determine, for any candidate video stream and any favorite video stream, a first number of users that favorite the any candidate video stream, and a second number of users that favorite the any favorite video stream and the any candidate video stream at the same time;
determining a similarity of the any favorite video stream and the any candidate video stream based on the first number of users and the second number of users;
and a user preference determining subunit, configured to determine a user preference of any one of the candidate video streams based on a similarity between each of the favorite video streams and the any one of the candidate video streams.
In the apparatus according to any of the above embodiments, the user representation further includes a user preference for each of the preferred video streams;
correspondingly, the user preference determination subunit is specifically configured to:
if the similarity between the favorite video stream and the candidate video stream is greater than a preset similarity threshold, determining that the favorite video stream is a reference video stream of the candidate video stream;
and weighting the user preference of each reference video stream by taking the similarity of each reference video stream and any candidate video stream as a weight to obtain the user preference of any candidate video stream.
Based on any of the above embodiments, in the apparatus, the user preference determining unit 320 is specifically configured to:
for any candidate video stream, inputting the user portrait and any candidate video stream into a preference prediction model, and outputting user preference;
the preference prediction model is obtained by training user preference of a sample user for a sample video stream based on a user image of the sample user, the sample video stream and the sample user.
Based on any one of the above embodiments, in the apparatus, the preference prediction model is composed of a sparse feature layer, a dense feature layer, and a sparse dense feature interaction layer;
wherein the sparse feature layer is used to extract user features from the user representation;
the dense feature layer is used for extracting video stream features from any candidate video stream;
the sparse dense feature interaction layer is configured to output the user preference based on the user feature and the video stream feature.
In the apparatus according to any of the above embodiments, the sparse dense feature interaction layer is represented by the following formula:
in the formula (I), the compound is shown in the specification,
for user preference, b and w are parameters obtained by pre-training, x is the video stream characteristic, v is the user characteristic, and upper represents rounding up.
Based on any of the above embodiments, in the apparatus, the recommending unit 330 is specifically configured to:
performing video stream recommendation based on the universal score and the user preference of each candidate video stream;
wherein the universal score is determined based on recording parameters of the candidate video streams; the recording parameter includes at least one of a lens type, a scene type, a sound element type, a transition type, and a special effect type.
Based on any of the above embodiments, in the apparatus, the recommending unit 330 is specifically configured to:
and recommending the video streams based on a first preset number of candidate video streams with the highest user preference and a second preset number of candidate video streams with the highest general score.
Based on any of the above embodiments, in the apparatus, the first preset number is greater than the second preset number;
correspondingly, the recommending unit 330 is specifically configured to:
constructing a user recommendation video stream set based on the first preset number of candidate video streams with the highest user preference;
taking the second preset number of candidate video streams with the highest general score as general recommended video streams;
replacing the second preset number of candidate video streams with the lowest user preference in the user recommended video stream set with the general recommended video stream;
and pushing the user recommended video stream set to a user side.
Based on any embodiment above, the apparatus further comprises:
the scoring unit is used for inputting the recording parameters of the candidate video streams into a general scoring model to obtain general scores of the candidate video streams output by the general scoring model;
the universal scoring model is obtained by training based on sample recording parameters of the sample video stream and the sample universal score.
Fig. 4 is a schematic structural diagram of an electronic device according to an embodiment of the present invention, and as shown in fig. 4, the electronic device may include: a processor (processor)401, a communication Interface (communication Interface)402, a memory (memory)403 and a communication bus 404, wherein the processor 401, the communication Interface 402 and the memory 403 complete communication with each other through the communication bus 404. The processor 401 may call a computer program stored in the memory 403 and executable on the processor 401 to execute the video stream recommendation method provided by the above embodiments, for example, including: determining a plurality of candidate video streams; determining a user preference for each of the candidate video streams based on the user profile; and recommending the video streams based on the user preference of each candidate video stream.
In addition, the logic instructions in the memory 403 may be implemented in the form of software functional units and stored in a computer readable storage medium when the software functional units are sold or used as independent products. Based on such understanding, the technical solutions of the embodiments of the present invention may be essentially implemented or make a contribution to the prior art, or may be implemented in the form of a software product stored in a storage medium and including instructions for causing a computer device (which may be a personal computer, a server, or a network device) to execute all or part of the steps of the methods described in the embodiments of the present invention. And the aforementioned storage medium includes: a U-disk, a removable hard disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a magnetic disk or an optical disk, and other various media capable of storing program codes.
Embodiments of the present invention further provide a non-transitory computer-readable storage medium, on which a computer program is stored, where the computer program is implemented to perform the video stream recommendation method provided in the foregoing embodiments when executed by a processor, for example, the method includes: determining a plurality of candidate video streams; determining a user preference for each of the candidate video streams based on the user profile; and recommending the video streams based on the user preference of each candidate video stream.
The above-described embodiments of the apparatus are merely illustrative, and the units described as separate parts may or may not be physically separate, and parts displayed as units may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of the present embodiment. One of ordinary skill in the art can understand and implement it without inventive effort.
Through the above description of the embodiments, those skilled in the art will clearly understand that each embodiment can be implemented by software plus a necessary general hardware platform, and certainly can also be implemented by hardware. With this understanding in mind, the above-described technical solutions may be embodied in the form of a software product, which can be stored in a computer-readable storage medium such as ROM/RAM, magnetic disk, optical disk, etc., and includes instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute the methods described in the embodiments or some parts of the embodiments.
Finally, it should be noted that: the above examples are only intended to illustrate the technical solution of the present invention, but not to limit it; although the present invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some technical features may be equivalently replaced; and such modifications or substitutions do not depart from the spirit and scope of the corresponding technical solutions of the embodiments of the present invention.
Claims (10)
1. A video stream recommendation method, comprising:
determining a plurality of candidate video streams;
determining a user preference for each of the candidate video streams based on the user profile;
and recommending the video streams based on the user preference of each candidate video stream.
2. The video stream recommendation method of claim 1, wherein said user representation comprises a number of favorite video streams of the user;
correspondingly, the determining the user preference of each candidate video stream based on the user profile specifically includes:
determining, for any favorite video stream and any candidate video stream, a first number of users that favorite the any candidate video stream, and a second number of users that favorite the any favorite video stream and the any candidate video stream at the same time;
determining a similarity of the any favorite video stream and the any candidate video stream based on the first number of users and the second number of users;
determining a user preference for each of the candidate video streams based on the similarity of the candidate video stream to the preferred video stream.
3. The video stream recommendation method of claim 2, wherein said user representation further comprises user preferences for each of said preferred video streams;
correspondingly, the determining the user preference of any candidate video stream based on the similarity between each preferred video stream and any candidate video stream specifically includes:
if the similarity between the favorite video stream and the candidate video stream is greater than a preset similarity threshold, determining that the favorite video stream is a reference video stream of the candidate video stream;
and weighting the user preference of each reference video stream by taking the similarity of each reference video stream and any candidate video stream as a weight to obtain the user preference of any candidate video stream.
4. The method of claim 1, wherein the determining the user preference for each of the candidate video streams based on the user profile comprises:
for any candidate video stream, inputting the user portrait and any candidate video stream into a preference prediction model, and outputting user preference;
the preference prediction model is obtained by training user preference of a sample user for a sample video stream based on a user image of the sample user, the sample video stream and the sample user.
5. The video stream recommendation method of claim 4, wherein said preference prediction model is composed of a sparse feature layer, a dense feature layer, and a sparse dense feature interaction layer;
wherein the sparse feature layer is used to extract user features from the user representation;
the dense feature layer is used for extracting video stream features from any candidate video stream;
the sparse dense feature interaction layer is configured to output the user preference based on the user feature and the video stream feature.
6. The video stream recommendation method of claim 5, wherein said sparse dense feature interaction layer is represented by the following formula:
in the formula (I), the compound is shown in the specification,
for user preference, b and w are parameters obtained by pre-training, x is the video stream characteristic, v is the user characteristic, and upper represents rounding up.
7. The method of claim 1, wherein the recommending a video stream based on the user preference of each candidate video stream comprises:
performing video stream recommendation based on the universal score and the user preference of each candidate video stream;
wherein the universal score is determined based on recording parameters of the candidate video streams; the recording parameter includes at least one of a lens type, a scene type, a sound element type, a transition type, and a special effect type.
8. The method according to claim 7, wherein the recommending a video stream based on the general score and the user preference of each of the candidate video streams comprises:
and recommending the video stream based on the first preset number of paths of candidate video streams with the highest user preference and the second preset number of paths of candidate video streams with the highest general score.
9. An electronic device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, characterized in that the processor implements the steps of the video stream recommendation method according to any one of claims 1 to 8 when executing the program.
10. A non-transitory computer readable storage medium, having stored thereon a computer program, characterized in that the computer program, when being executed by a processor, implements the steps of the video stream recommendation method according to any one of claims 1 to 8.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911142687.1A CN110677701A (en) | 2019-11-20 | 2019-11-20 | Video stream recommendation method, electronic device and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911142687.1A CN110677701A (en) | 2019-11-20 | 2019-11-20 | Video stream recommendation method, electronic device and storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110677701A true CN110677701A (en) | 2020-01-10 |
Family
ID=69087954
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201911142687.1A Pending CN110677701A (en) | 2019-11-20 | 2019-11-20 | Video stream recommendation method, electronic device and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110677701A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111726649A (en) * | 2020-06-28 | 2020-09-29 | 百度在线网络技术（北京）有限公司 | Video stream processing method, device, computer equipment and medium |
| CN111881358A (en) * | 2020-07-31 | 2020-11-03 | 北京达佳互联信息技术有限公司 | Object recommendation system, method and device, electronic equipment and storage medium |
| CN113569668A (en) * | 2021-07-12 | 2021-10-29 | 杭州网易云音乐科技有限公司 | Method, medium, apparatus and computing device for determining highlight segments in video |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103488764A (en) * | 2013-09-26 | 2014-01-01 | 天脉聚源(北京)传媒科技有限公司 | Personalized video content recommendation method and system |
| WO2018118972A1 (en) * | 2016-12-19 | 2018-06-28 | Qualcomm Incorporated | Preferred rendering of signalled regions-of-interest or viewports in virtual reality video |
| CN108733842A (en) * | 2018-05-29 | 2018-11-02 | 北京奇艺世纪科技有限公司 | Video recommendation method and device |
| CN108763314A (en) * | 2018-04-26 | 2018-11-06 | 深圳市腾讯计算机系统有限公司 | A kind of interest recommends method, apparatus, server and storage medium |
| CN110035302A (en) * | 2018-01-11 | 2019-07-19 | 阿里巴巴集团控股有限公司 | Information recommendation and model training method and device calculate equipment, storage medium |
| CN110446056A (en) * | 2019-07-30 | 2019-11-12 | 咪咕文化科技有限公司 | A kind of method for processing video frequency, equipment, system and computer readable storage medium |
-
2019
- 2019-11-20 CN CN201911142687.1A patent/CN110677701A/en active Pending
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103488764A (en) * | 2013-09-26 | 2014-01-01 | 天脉聚源(北京)传媒科技有限公司 | Personalized video content recommendation method and system |
| WO2018118972A1 (en) * | 2016-12-19 | 2018-06-28 | Qualcomm Incorporated | Preferred rendering of signalled regions-of-interest or viewports in virtual reality video |
| CN110035302A (en) * | 2018-01-11 | 2019-07-19 | 阿里巴巴集团控股有限公司 | Information recommendation and model training method and device calculate equipment, storage medium |
| CN108763314A (en) * | 2018-04-26 | 2018-11-06 | 深圳市腾讯计算机系统有限公司 | A kind of interest recommends method, apparatus, server and storage medium |
| CN108733842A (en) * | 2018-05-29 | 2018-11-02 | 北京奇艺世纪科技有限公司 | Video recommendation method and device |
| CN110446056A (en) * | 2019-07-30 | 2019-11-12 | 咪咕文化科技有限公司 | A kind of method for processing video frequency, equipment, system and computer readable storage medium |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111726649A (en) * | 2020-06-28 | 2020-09-29 | 百度在线网络技术（北京）有限公司 | Video stream processing method, device, computer equipment and medium |
| CN111726649B (en) * | 2020-06-28 | 2021-12-28 | 百度在线网络技术（北京）有限公司 | Video stream processing method, device, computer equipment and medium |
| WO2022000826A1 (en) * | 2020-06-28 | 2022-01-06 | 百度在线网络技术（北京）有限公司 | Video stream processing method and apparatus, and computer device and medium |
| EP4149115A4 (en) * | 2020-06-28 | 2023-06-21 | Baidu Online Network Technology (Beijing) Co., Ltd. | Video stream processing method and apparatus, and computer device and medium |
| CN111881358A (en) * | 2020-07-31 | 2020-11-03 | 北京达佳互联信息技术有限公司 | Object recommendation system, method and device, electronic equipment and storage medium |
| CN111881358B (en) * | 2020-07-31 | 2021-08-03 | 北京达佳互联信息技术有限公司 | Object recommendation system, method and device, electronic equipment and storage medium |
| CN113569668A (en) * | 2021-07-12 | 2021-10-29 | 杭州网易云音乐科技有限公司 | Method, medium, apparatus and computing device for determining highlight segments in video |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11601699B2 (en) | Predictive content delivery for video streaming services | |
| Lin et al. | MCL-V: A streaming video quality assessment database | |
| KR102027670B1 (en) | Spectator relational video production device and production method | |
| CN110677701A (en) | Video stream recommendation method, electronic device and storage medium | |
| WO2021129435A1 (en) | Method for training video definition evaluation model, video recommendation method, and related device | |
| CN112954312B (en) | Non-reference video quality assessment method integrating space-time characteristics | |
| US20120287233A1 (en) | Personalizing 3dtv viewing experience | |
| Yang et al. | An objective assessment method based on multi-level factors for panoramic videos | |
| US20230113222A1 (en) | Automatic selection of viewpoint characteristics and trajectories in volumetric video presentations | |
| Ma et al. | No-reference retargeted image quality assessment based on pairwise rank learning | |
| EP3761187A1 (en) | Method and apparatus for matching multimedia resource, and storage medium and electronic device | |
| US11663725B2 (en) | Selecting viewpoints for rendering in volumetric video presentations | |
| CN110866563B (en) | Similar video detection and recommendation method, electronic device and storage medium | |
| CN109472764B (en) | Method, apparatus, device and medium for image synthesis and image synthesis model training | |
| CN107103070B (en) | Cover setting method, device and system | |
| Jang et al. | On evaluating perceptual quality of online user-generated videos | |
| CN110493609B (en) | Live broadcast method, terminal and computer readable storage medium | |
| CN114079777B (en) | Video processing method and device | |
| US10867586B1 (en) | Virtual reality streaming media system and method of use | |
| CN110113595B (en) | Method and device for converting 2D video into 3D video and electronic equipment | |
| CN117714712B (en) | Data steganography method, equipment and storage medium for video conference | |
| JP6623905B2 (en) | Server device, information processing method and program | |
| CN117640877B (en) | Picture reconstruction method for online conference and electronic equipment | |
| CN116596752B (en) | Face image replacement method, device, equipment and storage medium | |
| US20240119637A1 (en) | Image synthesis method and system |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20200110 |