CN115767113A - Cloud rebroadcasting method, device, medium and system - Google Patents
Cloud rebroadcasting method, device, medium and system Download PDFInfo
- Publication number
- CN115767113A CN115767113A CN202211169502.8A CN202211169502A CN115767113A CN 115767113 A CN115767113 A CN 115767113A CN 202211169502 A CN202211169502 A CN 202211169502A CN 115767113 A CN115767113 A CN 115767113A
- Authority
- CN
- China
- Prior art keywords
- video stream
- reconstructed
- original
- picture
- frame
- 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.)
- Granted
Links
Images
Landscapes
- Two-Way Televisions, Distribution Of Moving Picture Or The Like (AREA)
Abstract
The application discloses a cloud rebroadcasting method, a device, a medium and a system, wherein the cloud rebroadcasting method firstly receives an original video stream related to a target area, and then continuously identifies a target object on a picture of the original video stream, the identification result comprises the position of the target object in the picture, then reconstructs the original video stream in real time according to the identification result to obtain a reconstructed video stream so as to simulate movement and zooming in the frame of the original video stream through the reconstructed video stream, wherein the reconstructed frame of the reconstructed video stream is smaller than the original frame of the original video stream, the reconstructed resolution of the reconstructed video stream is smaller than the original resolution of the original video stream, and finally the reconstructed video stream is sent to a playing end, so that the picture effect of manually tracking the target object to shoot is simulated through a single camera at a fixed machine position, the functions of shooting, reconstructing and rebroadcasting are automatically carried out, the cost investment of on-site shooting, broadcasting and post-stage production is reduced, and the rebroadcasting efficiency is improved.
Description
Technical Field
The present application relates to the field of video processing technologies, and in particular, to a cloud relay method, apparatus, medium, and system.
Background
With the increasing abundance of human activities, more and more entertainment and stylistic activity sites need to be filmed and viewed live remotely, such as small and medium-sized sporting events, lecture games, dance activities, and the like. In the course of implementing the present application, the inventors have found that the relaying of entertainment activities often requires the employment of professional shooting teams, which results in higher costs for relaying and therefore currently lacks an efficient and low cost solution.
It is noted that the information disclosed in this background section is only for background understanding of the concepts of the application and, therefore, it may contain information that does not form the prior art.
Disclosure of Invention
A first objective of the present application is to provide a cloud relay method, which reduces relay cost and improves relay efficiency.
A second object of the present application is to propose a computer-readable storage medium.
A third objective of the present application is to provide a cloud relay device.
A fourth objective of the present application is to provide a cloud relay system.
To achieve the above object, an embodiment of a first aspect of the present application provides a cloud relay method, including: receiving an original video stream regarding a target area; continuously carrying out target object identification on the picture of the original video stream, wherein the identification result comprises the position of a target object in the picture; reconstructing the original video stream in real time according to the identification result to obtain a reconstructed video stream so as to facilitate moving and zooming in the frame of the original video stream through simulation of the reconstructed video stream, wherein the reconstructed frame of the reconstructed video stream is smaller than the original frame of the original video stream, and the reconstruction resolution of the reconstructed video stream is smaller than the original resolution of the original video stream; and sending the reconstructed video stream to a playing end.
According to the cloud rebroadcasting method provided by the embodiment of the application, the picture frame of a small picture is used for moving and zooming on the full-field picture contained in the original video stream, so that the reconstructed video picture content can change along with the target position, the picture flexibility is improved, the data transmission is light by converting into the small picture size and the lower resolution, the picture effect of manually tracking a target object for shooting is simulated by a single camera of a fixed machine position, the whole shooting process is unattended, the shooting, reconstructing and rebroadcasting functions are automatically carried out, the cost investment of field shooting, broadcasting guide and post-production is reduced, and the rebroadcasting efficiency is improved.
According to an embodiment of the present application, reconstructing the original video stream in real time according to the recognition result to obtain a reconstructed video stream includes: and if the target object is identified, reconstructing the original video stream according to the position of the target object to obtain a reconstructed video stream, wherein the reconstructed frame is smaller than the original frame, and the frame of the reconstructed video stream contains the target object.
According to an embodiment of the present application, reconstructing the original video stream according to the position of the target object to obtain a reconstructed video stream includes: and if the identified target object is completely located in a preset area of the reconstruction frame, reconstructing the original video stream according to the position of the target object and the preset area to obtain a reconstruction video stream, wherein the central point of the preset area is coincident with the central point of the reconstruction frame, and the current reconstruction position of the reconstruction frame in the original frame is the same as the reconstruction position of the reconstruction frame during the last reconstruction.
According to one embodiment of the application, the target object comprises a primary target and a secondary target.
According to an embodiment of the present application, reconstructing the original video stream according to the position of the target object to obtain a reconstructed video stream includes: and if the identified target object comprises the main target, reconstructing the original video stream according to the position of the main target to obtain a reconstructed video stream, wherein the main target is positioned in the central area of the picture of the reconstructed video stream.
According to an embodiment of the present application, reconstructing the original video stream according to the position of the target object to obtain a reconstructed video stream includes: and if the identified target object comprises the secondary target and does not comprise the primary target, determining a reconstruction position of the reconstructed picture in the original picture according to the distribution position of the secondary target, and reconstructing the original video stream according to the reconstruction position to obtain a reconstructed video stream.
According to an embodiment of the present application, reconstructing the original video stream in real time according to the recognition result to obtain a reconstructed video stream includes: if the target object is not identified, when the original video stream is reconstructed, the current reconstruction position of the reconstructed picture in the original picture is the same as the reconstruction position of the reconstructed picture in the last reconstruction until the target object is identified in the picture of the original video stream.
According to one embodiment of the application, the target object recognition of the picture of the original video stream comprises: performing frame extraction on the original video stream to obtain a sampling image; and carrying out target object identification on the sampling image.
To achieve the above object, a second aspect of the present application provides a computer-readable storage medium, on which a cloud relaying program is stored, and the cloud relaying program, when executed by a processor, implements the cloud relaying method as in the above embodiments.
In order to achieve the above object, an embodiment of a third aspect of the present application provides a cloud relay device, which includes a memory, a processor, and a cloud relay program that is stored in the memory and is capable of running on the processor, and when the processor executes the cloud relay program, the cloud relay method in the foregoing embodiment is implemented.
To achieve the above object, a fourth aspect of the present application provides a cloud relay system, including: the video acquisition equipment is used for shooting a target area and transmitting an original video stream obtained through shooting to the streaming media server, the streaming media server is used for receiving the original video stream related to the target area and continuously carrying out target object identification on a picture of the original video stream, an identification result comprises the position of a target object in the picture, the original video stream is reconstructed in real time according to the identification result to obtain a reconstructed video stream, so that the original video stream is simulated to move and zoom in the frame of the original video stream through the reconstructed video stream, the reconstructed frame of the reconstructed video stream is smaller than the original frame of the original video stream, the reconstructed resolution of the reconstructed video stream is smaller than the original resolution of the original video stream, and the reconstructed video stream is sent to a playing end.
Additional aspects and advantages of the present application will be set forth in part in the description which follows and, in part, will be obvious from the description, or may be learned by practice of the present application.
Drawings
Fig. 1 is a schematic flow chart of a cloud relay method according to an embodiment of the present application.
Fig. 2 is a schematic diagram of a thread call of a central processing module of a streaming server according to an embodiment of the present application.
FIG. 3 is a schematic representation of reconstructed video frames of dancer A when boarding a stage according to an embodiment of the present application.
Fig. 4 is a schematic view of a reconstructed video frame of dancer a walking to the center of the stage according to an embodiment of the present application.
Fig. 5 is a schematic view of a reconstructed video frame of a dancer a in the center of a stage with an increased zoom factor according to an embodiment of the present application.
Fig. 6 is a schematic diagram of a reconstructed video picture of a dancer a dancing in the center of a stage according to an embodiment of the present application.
Fig. 7 is a schematic diagram of a reconstructed video frame when basketball is used as a sole focusing object in an embodiment of the present application.
Fig. 8 is a schematic diagram of a reconstructed video frame after reducing zoom factor when the positions of the players are scattered in one embodiment of the application.
Fig. 9 is a schematic diagram of a clustering result of clustering player positions in one embodiment of the present application.
Fig. 10 is a block diagram of a cloud relay device according to an embodiment of the present application.
Fig. 11 is a block diagram illustrating a configuration of a cloud relay system according to an embodiment of the present application.
Detailed Description
Reference will now be made in detail to the embodiments of the present application, examples of which are illustrated in the accompanying drawings, wherein like or similar reference numerals refer to the same or similar elements or elements having the same or similar function throughout. The embodiments described below with reference to the drawings are exemplary and intended to be used for explaining the present application and should not be construed as limiting the present application.
A cloud relay method, apparatus, medium, and system according to embodiments of the present application are described below with reference to the drawings.
Referring to fig. 1, the cloud broadcasting method in the embodiment of the present application includes the following steps 100 to 400.
S100, an original video stream is received with respect to a target area.
The target area is a meeting place for human activities, and if the target area is a dance match site of a theater, the stage is shot by arranging the video acquisition equipment on the site so as to shoot and upload pictures of dance matches of players on the match site.
The video acquisition equipment can adopt the camera of single fixed position of machine, because adopt single camera to shoot, but not many cameras carry out multi-angle respectively and shoot, consequently to all kinds of target area, the camera all need set up in the eminence to in the target area go on bowing, thereby cover the whole region of target area, and need adopt the higher camera of resolution ratio, for example high definition or super high definition camera, avoid the blurred problem of picture when shooing large tracts of land target area. Specifically, for the dance game of a theater, an ultra-high definition camera, for example, a 4K camera, may be arranged at a high position of the theater, and the camera is used to shoot the whole stage of the theater, rather than just shooting people on the stage, so as to obtain an ultra-high definition original video stream.
After the dance match starts, the camera shoots the meeting place in real time, continuously encodes the shot original video stream and uploads the encoded original video stream, specifically, the original video stream can be uploaded to a streaming media server, and the streaming media server receives the original video stream related to the target area.
Referring to fig. 2, after the streaming media server starts to receive the original video stream, the central processing module of the streaming media server may invoke the decoding thread to pull the original video from the streaming media server and decode the pulled video.
S200, continuously carrying out target object recognition on the picture of the original video stream, wherein the recognition result comprises the position of the target object in the picture.
The central processing module can identify the target object in the decoded video by calling the identification thread, and can perform downsampling on the decoded video before identification so as to reduce the resolution of the picture, and then identify the target object on the video picture with reduced resolution. In addition, the down-sampling may be performed by the decoding thread, that is, the decoding thread decodes the original video stream, and then down-samples the decoded video image and transmits the down-sampled video image to the identification thread.
The target object is a preset object, which can be a character object or an article object, for example, in speech games, dance games and track and field games, the target object usually only contains the character object, so that only human body recognition or face recognition can be performed; in javelin sports and ball sports, the target object can only comprise a character object (athlete) or can simultaneously comprise the character object and an object (javelin and ball), so that human body recognition or face recognition can be carried out, and javelin and ball recognition can be carried out simultaneously; in both racing and drone activities, the target object may contain only item objects, i.e. both the race and the drone themselves, and thus may only identify the race and the drone. In the identification process, if the target object is identified, the appearance position of the target object in the 4K picture can be determined at the same time.
It is understood that the recognition result may also include the size of the target object, and whether the recognition result is correct may be determined by the size of the target object.
And S300, reconstructing the original video stream in real time according to the identification result to obtain a reconstructed video stream so as to facilitate moving and zooming in the frame of the original video stream through simulation of the reconstructed video stream, wherein the reconstructed frame of the reconstructed video stream is smaller than the original frame of the original video stream, and the reconstruction resolution of the reconstructed video stream is smaller than the original resolution of the original video stream.
The central processing module calls the composition thread to cache the original video when calling the identification thread to identify the target object of the original video stream, sends the identification result to the composition thread after the identification thread obtains the identification result, and the composition thread reconstructs the cached original video according to the identification result to obtain a reconstructed video.
Because the original video stream is large in picture size and high in resolution, in order to reduce the requirements on the playing capacity and the data transmission rate of the playing end equipment, when the central processing module calls a composition thread to reconstruct the original video stream, the picture frame of the small picture is used for moving and zooming on the full-field picture contained in the original video stream, the content of the reconstructed video picture can change along with the target position, so that the picture flexibility is improved, the data transmission is light in weight by converting the down-sampling into the small picture size and the lower resolution, and the effect of simulating manual tracking of a target object to shoot through a single camera of a fixed machine position is realized.
It is understood that since the target objects appearing in the photographing region at different times may be different, the target objects included in the reconstructed video streams reconstructed at different times may be different.
Taking the shooting of the flight performance match of the small unmanned aerial vehicle as an example, because the flight performance space region of the unmanned aerial vehicle is large, the shooting can be carried out by using an 8K camera, the shooting region covers the flight action region, the picture size of the original video stream obtained by shooting is W0 × H0, and the picture resolution is 7860 × 4320. When the unmanned aerial vehicle enters the performance area, the streaming media server identifies the unmanned aerial vehicle from the picture of the original video stream, so that the default picture size of the reconstructed video stream obtained after reconstruction is w0 h0, w0=0.5 w0, h0=0.5 h0, the reconstructed picture is one fourth of the original picture, and the reconstructed picture can be focused on the unmanned aerial vehicle, so that the reconstructed picture contains the unmanned aerial vehicle. When the unmanned aerial vehicle leaves the performance area, the new unmanned aerial vehicle enters the performance area from the beginning, the streaming media server identifies the new unmanned aerial vehicle from the picture of the original video stream again, the position of the reconstructed picture in the original picture is adjusted along with the flight of the new unmanned aerial vehicle, and the follow-up unmanned aerial vehicle repeats the process.
S400, the reconstructed video stream is sent to a playing end.
When the reconstructed video stream is reconstructed by the composition thread called by the central processing module, the reconstructed video stream can be encoded in real time, and the ultra-high-definition playing end can pull and play the reconstructed video from the streaming media server, that is, what a viewer sees is the content in a rectangular dashed frame in the figure.
According to the cloud rebroadcasting method provided by the embodiment of the application, the picture frame of a small picture is used for moving and zooming on the full-field picture contained in the original video stream, the content of the reconstructed video picture can be changed along with the target position, so that the picture flexibility is improved, the data transmission is light by converting the reconstructed video picture into the small picture size and the lower resolution, the picture effect of manually tracking a target object for shooting is simulated through a single camera of a fixed machine position, the whole shooting process is unattended, the shooting, reconstructing and rebroadcasting functions are automatically carried out, the cost investment of field shooting, broadcasting guide and post-production is reduced, and the rebroadcasting efficiency is improved.
In some embodiments, the method for reconstructing the original video stream in real time according to the recognition result in step 300 to obtain the reconstructed video stream may specifically include the following step 310.
S310, if the target object is identified, reconstructing the original video stream according to the position of the target object to obtain a reconstructed video stream. The reconstructed frame is smaller than the original frame, and the frame of the reconstructed video stream contains the target object.
Referring to fig. 3 and 4, taking the example of shooting a dance game in a theater, the cuboid in the figure is a stage, and the stage area is small, so that the shooting can be performed by using a 4K camera, the shooting area covers the stage area, the size of the shot original video stream is W1 × H1, as shown by the solid rectangular frame in the figure, and the picture resolution is 4096 × 2160. As shown in fig. 3, when a dancer a lands on the stage, the streaming media server identifies a character object from the pictures of the original video stream, so that the default frame size of the reconstructed video stream obtained after reconstruction is w1 × h1, as shown by the dashed box in the figure, and w1=0.5 × w1, h1=0.5 × h1, and the default frame resolution is 2048 × 1080, which corresponds to the reconstructed frame being one fourth of the original frame. And the reconstructed picture can be focused on a person object, so that the reconstructed picture contains the dancer A and the dancer A can be kept to be always positioned at the center of the reconstructed picture. As shown in fig. 4, as the dancer a is continuously identified (going from the stage step to the stage center), the reconstructed frame moves (the shooting angle is not changed) in the original frame along with the movement of the dancer a, so that the dancer a always remains located in the frame of the reconstructed frame or even in the frame center position, thereby obtaining a segment of reconstructed video stream as the dancer a gets on the stage, performs dancing and leaves the stage. When the dancer A leaves the stage, a new dancer starts to board the stage, the streaming media server identifies the character object from the picture of the original video stream again, the position of the reconstructed picture in the original picture is adjusted along with the movement of the new person, and the rest persons do the same.
In some embodiments, when performing video reconstruction in step 310, the reconstruction may also be performed according to the size of the target object, that is, according to the position and size of the target object, wherein the position of the reconstructed frame in the original frame is determined by the position of the target object, and the zoom factor of the reconstructed frame is determined by the size of the target object. The zoom multiple is used for performing simulated zooming-in or simulated zooming-out on an original picture in the reconstructed picture or keeping a current visual field unchanged, and the zoom multiple can be specifically realized by adjusting the size of the reconstructed picture.
The size of the minimum envelope rectangle of the target object can be used as the size of the target object, if the size of the target object is smaller than a preset first size threshold, the size of the target object is too small, so that the zoom factor can be increased, for example, the size of the reconstructed frame is reduced, the central position of the reconstructed frame is kept unchanged, the reconstructed frame obtained after the size of the reconstructed frame is reduced, when the reconstructed frame is enlarged to the default size of the reconstructed frame during reconstruction, the target object in the reconstructed frame is enlarged along with the enlargement, so that the target object is simulated to be drawn close, and the reconstructed frame can clearly display the target object; if the size of the target object is larger than a preset second size threshold (the second size threshold is larger than the first size threshold), the size of the target object is too large, so that the zoom factor can be reduced, for example, the size of the reconstructed frame is increased while the central position of the reconstructed frame is kept unchanged, the reconstructed frame obtained after the size of the reconstructed frame is increased, and when the reconstructed frame is reduced to the default size of the reconstructed frame during reconstruction, the target object in the reconstructed frame is reduced along with the reduction, so that the target object is pulled away, and the reconstructed frame can completely display the target object and the surrounding contents; if the size of the target object is between the first size threshold and the second size threshold, the zoom multiple does not need to be changed, the zoom multiple at the moment is a default value, and the zoom multiple of the reconstructed frame is not changed, wherein the default value of the zoom multiple is a numerical value when zooming is not performed, and corresponds to the default size of the reconstructed frame.
It can be understood that, if the size of the reconstructed frame is changed, the size of the reconstructed frame obtained under the changed reconstructed frame needs to be adjusted to the default size of the reconstructed frame during reconstruction, so as to make the sizes of the video frames of the played video consistent.
Referring to fig. 5, assuming that the stage is large and the size of the dancer a in the original frame is small, when the original video stream is reconstructed according to the position of the dancer a, the zoom factor of the reconstructed frame is further increased according to the size of the dancer a, so that the size of the dancer a in the reconstructed frame is increased, so as to be able to clearly see the dancing action of the dancer. It can be understood that no matter what value the current zoom multiple of the reconstructed picture is, the reconstructed picture is not influenced to move along with the movement of the target object.
In some embodiments, the reconstructing the original video stream according to the position of the target object in step 310 to obtain a reconstructed video stream may specifically include: and if the identified target object is completely positioned in the preset area of the reconstruction frame, reconstructing the original video stream according to the position of the target object and the preset area to obtain the reconstructed video stream. The center point of the preset area coincides with the center point of the reconstructed picture, and the current reconstruction position of the reconstructed picture in the original picture is the same as the reconstruction position of the reconstructed picture in the last reconstruction.
Referring to fig. 6, the dotted line rectangular frame in fig. 6 is a preset region, the preset region is located at the center of the reconstructed frame, and the width and height of the preset region are respectively smaller than the width and height of the reconstructed frame. Specifically, the size of the preset area may be adjusted according to the shooting scene, for example, the width and the height of the preset area may be respectively half of the width and the height of the reconstructed frame. Taking the example of taking a picture of a dance match in a theater as an example, after a dancer a gets on the stage and starts a dance performance, if the dancer a keeps the reconstructed frame (shown by a dotted line in the figure) as a frame center, the reconstructed position (the position of the reconstructed frame in the original frame) will always move with the dancer a, and the frame will also sway with the dancer a, so that the reconstructed position of the reconstructed frame will not move as the dancer a moves in the original frame when the dancer a does not move outside the preset area through step 311, but will not move as the dancer a moves in the original frame when the dancer a does not move outside the preset area, and is fixed at the position currently located in the original frame, so that when the dancer a dances in the preset area, the position of the reconstructed frame will be stable and not move, for example, in fig. 6, when the dancer a is located in the frame of the dotted frame, the dotted frame will keep itself in the position in the solid frame, and avoid the dancer a feeling of the reconstructed frame and the audience will shake and the dancer, thereby improving the effect of the video. When the dancer A moves to the edge of the preset area and gradually moves out of the preset area, the reconstruction position of the reconstruction frame in the original frame is adjusted according to the position of the dancer A identified from the current original frame, for example, when the dancer A moves out of a dotted line frame from the left side of the frame, the dotted line frame moves to the left according to the identification result, and the dancer A is kept appearing in the reconstruction frame.
In some embodiments, the target object includes a primary target and a secondary target.
Since there may be a plurality of objects attracting the interests of the audience during the live events and games, and thus there may be a plurality of target objects, for example, in the basketball game, the audience watching the broadcast pays attention to both the basketball and the players, that is, both the basketball and the players are target objects and appear in the video picture at the same time, so that both the basketball and the players need to be considered during the broadcast. Since the movement of the basketball is the core point viewed by the audience, the basketball can be used as the main target, and each player can be used as the secondary target, that is, the focusing priority of the main target is higher than that of the secondary target, and the basketball is preferentially used as the center of the reconstructed picture frame, so that the picture of the basketball movement can be conveniently tracked by the analog camera.
In some embodiments, the reconstructing the original video stream according to the position of the target object in step 310 may specifically include: and if the identified target object comprises a main target, reconstructing the original video stream according to the position of the main target to obtain a reconstructed video stream. Wherein the primary target is located in a picture center region of the reconstructed video stream.
Referring to fig. 7, taking the rebroadcasting of the basketball game as an example, the basketball is a main target, each player is a secondary target, the trapezoid in the figure is a basketball court, the black circle is the basketball, the humanoid pattern is the player, the rebroadcasting of the game is shot by using an 8K camera, the shooting area covers the basketball court area, the size of the shot original video stream is W2 × H2, as shown by the solid rectangle frame in the figure, and the picture resolution is 8K resolution. When the recognition thread of the central processing module recognizes basketball from the pictures of the original video stream, no matter whether the player is recognized at the same time or not, the reconstructed picture can focus on basketball, the default picture size of the reconstructed video stream after reconstruction is w2 h2, and the picture resolution is 4K resolution, as shown by a dashed box in the figure, and w2=0.5 w2, h2=0.5 h2, which is equivalent to that the reconstructed picture is one fourth of the original picture. As the basketball is continuously identified, the reconstructed frame moves along with the movement of the basketball in the original frame (the shooting visual angle is not changed), so that the basketball always keeps in the central area of the picture of the reconstructed frame or even the central position of the picture, and a section of reconstructed video stream moving along with the basketball is obtained. The center area of the picture refers to an area including a center point of the picture.
Because the size of the basketball is small relative to the size of the field, the zoom factor can be increased, which is equivalent to enlarging the basketball and other contents, so that the size of the basketball and other contents in the reconstructed picture becomes larger (not shown in fig. 7), and the basketball and other contents are convenient for the audience to watch. Meanwhile, the position of the basketball can be judged through the preset area of the reconstructed frame, so that the position of the reconstructed frame in the original frame does not need to be changed when the basketball does not leave the preset area (not shown in figure 7), the image appreciation of the reconstructed video is enhanced, and the image is prevented from frequently shaking along with the basketball.
In some embodiments, the manner of reconstructing the original video stream according to the position of the target object in step 310 to obtain a reconstructed video stream may further include: and if the identified target object comprises the secondary target and does not comprise the primary target, determining the reconstruction position of the reconstructed picture in the original picture according to the distribution position of the secondary target, and reconstructing the original video stream according to the reconstruction position to obtain the reconstructed video stream.
If the basketball is not identified but one or more players are identified, the reconstructed position is determined according to the standing positions of the players, namely when the basketball is shot out of the field or cannot be identified from the original video due to other reasons, the basketball is focused on the basketball, and the picture of the players is mainly relayed so as to facilitate the simulation camera to track the picture of the movement of the players. When the basketball is identified later, the focus on the basketball is started again and the picture of the basketball is mainly rebroadcast.
In some embodiments, the manner of determining the reconstruction position of the reconstructed frame in the original frame according to the distribution position of the secondary object in step 310 may specifically be: determining a minimum envelope rectangle for enveloping all the secondary targets according to the distribution positions of the secondary targets; and determining the reconstruction position of the reconstructed picture in the original picture and the zoom multiple of the reconstructed picture according to the minimum envelope rectangle.
Referring to fig. 8, the dispersion degree of the player positions can be known from the distribution positions of the players, for example, all the players are used as each vertex of a closed polygon, a minimum envelope rectangle of the polygon is generated, then the minimum envelope rectangle and the reconstructed frame are compared in size, if the reconstructed frame cannot cover the minimum envelope rectangle, the zoom factor of the reconstructed frame is reduced, the zoom factor is used for performing simulated zooming on the original frame in the reconstructed frame, so that the reconstructed frame can contain more contents (players), and specifically, the zoom factor can be correspondingly reduced according to the size ratio of the minimum envelope rectangle to the reconstructed frame until the reconstructed frame can reach the degree of containing all the players. At this time, the center point of the reconstructed picture is determined according to the specific positions of the players, which is equivalent to determining the reconstruction position, so that all the players appear in the reconstructed picture. It is to be understood that, when the zoom factor is adjusted to be smaller than the default value, the change of the reconstruction position may be restricted so that the reconstruction position is maintained at the position at the time when the last zoom factor was equal to the default value.
If the player station positions are changed from dispersion to aggregation and the reconstructed picture can cover the minimum envelope rectangle, the zoom multiple of the reconstructed picture is improved, and the zoom multiple is used for carrying out simulated zooming on the original picture in the reconstructed picture, so that the displayed content of the reconstructed picture is clearer.
In some embodiments, the manner of determining the reconstruction position of the reconstructed frame in the original frame according to the distribution position of the secondary object in step 310 may further be: clustering the positions of all secondary targets through a clustering algorithm to obtain a plurality of clusters and a central point of each cluster; and determining the reconstruction position of the reconstructed picture in the original picture according to the central point position of each cluster and the number of the secondary targets.
Referring to fig. 9, taking the hierarchical clustering algorithm as an example, assuming that all 10 players are identified, each player is first taken as a cluster, and the position of the player in the original picture is the position of the corresponding cluster, and then two clusters that are closest to each other are repeatedly merged until the distance between each two clusters is greater than the preset distance threshold, so as to finally obtain three clusters of C1, C2 and C3. Wherein, the distance between two clusters is the average value of the sum of the distances between each member in one cluster and each member in the other cluster, namely the average-linking algorithm. The distance threshold may be set according to the size of the reconstructed frame, and since the more times of merging, the larger the distance between clusters, the distance between clusters obtained finally may be a certain distance by the distance threshold, and a certain number of clusters may also be reserved. And if the distance between the clusters is not greater than the distance threshold value until the two clusters are merged into the remaining two clusters, taking the two clusters as the finally obtained clusters.
In the three clusters in fig. 9, C1 contains the largest number of players, so that taking the center point of C1 as the center point of the current reconstructed frame can maximize the number of players appearing in the frame, but the distance between the center point of C1 and the center point of the last reconstructed frame is relatively long, thus causing the frame jump distance to be too large; c3 contains the minimum number of players, and the distance between the center point of C3 and the center point of the last reconstructed picture is longer, so that the number of players in the picture is smaller, and the picture jump distance is too large; the number of the players contained in the C2 is large, and the distance between the center point of the C2 and the center point of the previous reconstructed picture is short, so that the number of the players appearing in the picture is large, the picture jump distance can be avoided, the continuity of the picture is guaranteed, and the center point of the C2 is used as the center of the current reconstructed position.
Specifically, a first weight may be set for a distance between a center point position of the cluster and a center point of the last reconstructed frame, a second weight may be set for a number of secondary objects included in the cluster, a score may be calculated according to the distance between the center point position of the cluster and the center point of the last reconstructed frame, the number of secondary objects included in the cluster, and the corresponding first weight and second weight, and the reconstructed position may be selected according to a size of the score. For example, the first weight is set to be a negative value, the closer the distance, the higher the score, the second weight is set to be a positive value, the larger the number, the higher the score, the two scores are added to obtain the total score of the cluster, and the center point of the cluster with the highest total score is taken as the center of the reconstruction position.
In some embodiments, when any kind of target object changes from being identifiable to being unidentifiable, the reconstruction position is kept unchanged when the original video stream is reconstructed, timing is started, and when the timing value reaches the time threshold value, the original video stream is reconstructed according to the identification result at the moment when the timing value reaches the time threshold value to obtain the reconstructed video stream.
The target objects may be of various kinds, and the primary and secondary targets may each comprise one or more target objects, for example a basketball may be the only primary target and two secondary targets for players. If the target objects at this time are only two types, namely a basketball (main target) and all players (secondary targets), the basketball is identified from the original picture at the previous time, and the basketball cannot be identified at the current time, at this time, the basketball may move to the back of the body of a certain player in the jumping process, so that the shooting side of the camera cannot shoot the basketball, and if the focused object is directly changed from the basketball to the players, when the basketball quickly bounces from the players and is identified at the next time, the focused object is quickly changed back to the basketball from the players.
Therefore, a short time threshold value, for example, a time threshold value of 1 second, may be set, the timing is started from the moment when the basketball is not recognized, and the reconstruction position is kept unchanged at this moment, in the obtained reconstructed video, the picture is fixed at the current position, but the picture content continues to change with time, if the basketball cannot be recognized all the time when the timing value reaches 1 second, the focusing object is changed to the recognized player, and the reconstruction position is determined by the distribution position of the recognized player; if the basketball is identified before the timing value reaches 1 second, the basketball is continuously used as a focusing object, and the reconstruction position is determined according to the position of the basketball or the position of the basketball and the preset area, so that unnecessary rapid skip of pictures is avoided.
In some embodiments, the method for reconstructing the original video stream according to the identification result in step 300 to obtain a reconstructed video stream may further include: if the target object is not identified, when the original video stream is reconstructed, the current reconstruction position of the reconstruction frame in the original frame is the same as the reconstruction position of the reconstruction frame in the last reconstruction time until the target object is identified in the frame of the original video stream.
If neither basketball nor any player is identified in the original picture received from the current time due to an algorithm error or other reasons, all the target objects are not identified, the reconstruction position is kept unchanged when the original picture at the current time is reconstructed, the zoom multiple can also be kept unchanged, at this time, the basketball or the player may be included in the reconstructed picture but is not identified, and therefore the reconstruction position is continuously used, if no target object is identified in the original picture received at the next time, the reconstruction position is continuously used, if the target object is identified in the original picture received at one time, the reconstruction position is adjusted according to the identified target object, for example, if basketball is identified, the reconstruction position is adjusted according to the basketball position, or if only a plurality of players are identified, the reconstruction position is adjusted according to the player station position.
In some embodiments, the manner of performing target object recognition on the picture of the original video stream in step 200 may specifically include: and performing frame extraction on the original video stream to obtain a sampling image, and performing target object identification on the sampling image.
Specifically, the original video stream may be subjected to frame extraction according to a preset number of interval frames or a preset interval period to obtain a sampling picture in the original video, and only the sampling picture is subjected to target object identification, so as to increase the identification speed.
The identification thread and the composition thread are respectively configured with a data queue, the data queue of the identification thread is assumed to be a first queue, the data queue of the composition thread is assumed to be a second queue, the data access of the first queue and the data access of the second queue are synchronous, the two queues have real-time interaction, the frame extraction interval is 4 frames, namely, one frame is extracted every 4 frames to be used as a sampling image. When the first queue identifies a target object for an nth frame image (sampling picture), if a basketball is identified, the identification result is sent to the composition thread as the identification results of the (n-4) th to nth frames, and the composition thread performs corresponding reconstruction on the pictures of the (n-4) th to nth frames, wherein the total of the 5 frames are from the (n-4) th to nth frames according to the identification result.
In addition, embodiments of the present application also provide a computer readable storage medium, on which a cloud relay program is stored, and when executed by a processor, the cloud relay program implements the cloud relay method in the above embodiments.
According to the computer-readable storage medium provided by the embodiment of the application, the picture frame of the small picture is used for moving and zooming on the full-field picture contained in the original video stream, so that the reconstructed video picture content can change along with the target position, the picture flexibility is improved, the data transmission is light by converting into the small picture size and the lower resolution, the picture effect of manually tracking a target object for shooting is simulated through a single camera of a fixed machine position, the whole shooting process is unattended, the shooting, reconstructing and rebroadcasting functions are automatically carried out, the cost investment of field shooting, broadcasting guide and post-production is reduced, and the rebroadcasting efficiency is improved.
In addition, referring to fig. 10, an embodiment of the present application further provides a cloud relay device 10, which includes a memory 11, a processor 12, and a cloud relay program that is stored in the memory 11 and can be run on the processor 12, and when the processor 12 executes the cloud relay program, the cloud relay method in the foregoing embodiment is implemented.
According to the cloud rebroadcasting device that this application embodiment provided, frame through using little picture moves and zooms on the full field picture that original video stream contains, thereby make the video picture content after the reconsitution can change along with the target location and promoted the picture flexibility, thereby it is lightweight to make data transmission through turning into little picture size and lower resolution ratio, thereby realize simulating the picture effect that artifical tracking target object carried out the camera shooting through the single camera of fixed machine position, and whole shooting process realizes unmanned on duty, shoot automatically, reconsitution and rebroadcasting function, on-the-spot shooting has been reduced, the cost input of director and post production, rebroadcasting efficiency has been improved.
In addition, referring to fig. 11, an embodiment of the present application further provides a cloud relay broadcasting system 20, which includes a video acquisition device 21 and a streaming media server 22, where the video acquisition device 21 may adopt an ultra high definition camera, and is configured to shoot a target area and transmit an original video stream obtained by shooting to the streaming media server 22. The streaming media server 22 is configured to receive an original video stream related to a target area, continuously perform target object identification on a picture of the original video stream, where the identification result includes a position of a target object in the picture, reconstruct the original video stream in real time according to the identification result to obtain a reconstructed video stream, so as to simulate movement and zooming in a frame of the original video stream through the reconstructed video stream, where the reconstructed frame of the reconstructed video stream is smaller than the original frame of the original video stream, and a reconstruction resolution of the reconstructed video stream is smaller than an original resolution of the original video stream, and send the reconstructed video stream to the playing terminal 23.
In some embodiments, the manner in which the streaming media server 22 reconstructs the original video stream in real time according to the identification result to obtain the reconstructed video stream may specifically include: and if the target object is identified, reconstructing the original video stream according to the position of the target object to obtain a reconstructed video stream. The reconstructed frame is smaller than the original frame, and the frame of the reconstructed video stream contains the target object.
In some embodiments, when the streaming media server 22 performs video reconstruction, the reconstruction may also be performed according to the size of the target object, that is, according to the position and size of the target object, where the position of the reconstructed frame in the original frame is determined by the position of the target object, and the zoom factor of the reconstructed frame is determined by the size of the target object.
In some embodiments, the manner in which the streaming media server 22 reconstructs the original video stream according to the position of the target object to obtain the reconstructed video stream may specifically include: and if the identified target object is completely positioned in the preset area of the reconstructed picture, reconstructing the original video stream according to the position of the target object and the preset area to obtain a reconstructed video stream. The center point of the preset area coincides with the center point of the reconstructed picture, and the current reconstruction position of the reconstructed picture in the original picture is the same as the reconstruction position of the reconstructed picture in the last reconstruction.
In some embodiments, the target object includes a primary target and a secondary target.
In some embodiments, the method for reconstructing the original video stream by the streaming media server 22 according to the position of the target object to obtain the reconstructed video stream may specifically include: and if the identified target object comprises a main target, reconstructing the original video stream according to the position of the main target to obtain a reconstructed video stream. Wherein the primary target is located in a picture center region of the reconstructed video stream.
In some embodiments, the manner in which the streaming media server 22 reconstructs the original video stream according to the position of the target object to obtain the reconstructed video stream may further include: and if the identified target object comprises the secondary target and does not comprise the primary target, determining a reconstruction position of the reconstruction frame in the original frame according to the distribution position of the secondary target, and reconstructing the original video stream according to the reconstruction position to obtain a reconstruction video stream.
In some embodiments, the manner for the streaming media server 22 to determine the reconstruction position of the reconstructed frame in the original frame according to the distribution position of the secondary target may specifically be: determining a minimum envelope rectangle for enveloping all the secondary targets according to the distribution positions of the secondary targets; and determining the reconstruction position of the reconstructed picture in the original picture and the zoom multiple of the reconstructed picture according to the minimum envelope rectangle.
In some embodiments, the manner in which the streaming media server 22 determines the reconstruction position of the reconstructed frame in the original frame according to the distribution position of the secondary target may also be: clustering the positions of all secondary targets through a clustering algorithm to obtain a plurality of clusters and a central point of each cluster; and determining the reconstruction position of the reconstructed picture in the original picture according to the central point position of each cluster and the number of the secondary targets.
In some embodiments, when any kind of target object changes from being identifiable to being unidentifiable, the streaming media server 22 keeps the reconstruction position unchanged while reconstructing the original video stream, starts timing, and reconstructs the original video stream to obtain a reconstructed video stream according to the identification result at the time when the timing value reaches the time threshold.
In some embodiments, the manner in which the streaming media server 22 reconstructs the original video stream according to the identification result to obtain the reconstructed video stream may further include: if the target object is not identified, when the original video stream is reconstructed, the current reconstruction position of the reconstruction frame in the original frame is the same as the reconstruction position of the reconstruction frame in the last reconstruction time until the target object is identified in the frame of the original video stream.
In some embodiments, the manner of the streaming media server 22 performing target object identification on the picture of the original video stream may specifically include: and performing frame extraction on the original video stream to obtain a sampling image, and performing target object identification on the sampling image.
According to the cloud rebroadcasting system that this application embodiment provided, frame through using little picture moves and zooms on the full field picture that original video stream contains, thereby make the video picture content after the reconsitution can change along with the target location and promoted the picture flexibility, thereby it is lightweight to make data transmission through turning into little picture size and lower resolution ratio, thereby realize simulating the picture effect that artifical tracking target object carried out the camera shooting through the single camera of fixed machine position, and whole shooting process realizes unmanned on duty, shoot automatically, reconsitution and rebroadcasting function, the cost input of scene shooting, the director and post-production has been reduced, rebroadcasting efficiency has been improved.
It should be noted that the logic and/or steps represented in the flowcharts or otherwise described herein, such as an ordered listing of executable instructions that can be considered to implement logical functions, can be embodied in any computer-readable medium for use by or in connection with an instruction execution system, apparatus, or device, such as a computer-based system, processor-containing system, or other system that can fetch the instructions from the instruction execution system, apparatus, or device and execute the instructions. For the purposes of this description, a "computer-readable medium" can be any means that can contain, store, communicate, propagate, or transport the program for use by or in connection with the instruction execution system, apparatus, or device. More specific examples (a non-exhaustive list) of the computer-readable medium would include the following: an electrical connection (electronic device) having one or more wires, a portable computer diskette (magnetic device), a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable compact disc read-only memory (CDROM). Additionally, the computer-readable medium could even be paper or another suitable medium upon which the program is printed, as the program can be electronically captured, via for instance optical scanning of the paper or other medium, then compiled, interpreted or otherwise processed in a suitable manner if necessary, and then stored in a computer memory.
It should be understood that portions of the present application may be implemented in hardware, software, firmware, or a combination thereof. In the above embodiments, the various steps or methods may be implemented in software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, any one or combination of the following technologies, which are well known in the art, may be used: a discrete logic circuit having a logic gate circuit for implementing a logic function on a data signal, an application specific integrated circuit having an appropriate combinational logic gate circuit, a Programmable Gate Array (PGA), a Field Programmable Gate Array (FPGA), or the like.
In the description herein, reference to the description of the term "one embodiment," "some embodiments," "an example," "a specific example," or "some examples," etc., means that a particular feature, structure, material, or characteristic described in connection with the embodiment or example is included in at least one embodiment or example of the application. In this specification, the schematic representations of the terms used above do not necessarily refer to the same embodiment or example. Furthermore, the particular features, structures, materials, or characteristics described may be combined in any suitable manner in any one or more embodiments or examples.
Furthermore, the terms "first", "second" and "first" are used for descriptive purposes only and are not to be construed as indicating or implying relative importance or to implicitly indicate the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of the feature. In the description of the present application, "plurality" means at least two, e.g., two, three, etc., unless specifically limited otherwise. Relational terms such as first and second, and the like may be used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Also, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrases "comprising a," "8230," "8230," or "comprising" does not exclude the presence of additional like elements in a process, method, article, or apparatus that comprises the element.
All the embodiments in the present specification are described in a related manner, and the same and similar parts among the embodiments may be referred to each other, and each embodiment focuses on differences from other embodiments. In particular, as for the apparatus, the electronic device, and the computer-readable storage medium embodiments, since they are substantially similar to the method embodiments, the description is relatively simple, and in relation to the description, reference may be made to some portions of the description of the method embodiments.
The above description is only for the preferred embodiment of the present application, and is not intended to limit the scope of the present application. Any modification, equivalent replacement, improvement and the like made within the spirit and principle of the present application are included in the scope of protection of the present application.
Claims (11)
1. A cloud relay method, comprising:
receiving an original video stream regarding a target area;
continuously carrying out target object recognition on a picture of the original video stream, wherein the recognition result comprises the position of a target object in the picture;
reconstructing the original video stream in real time according to the identification result to obtain a reconstructed video stream so as to facilitate moving and zooming in the frame of the original video stream through simulation of the reconstructed video stream, wherein the reconstructed frame of the reconstructed video stream is smaller than the original frame of the original video stream, and the reconstruction resolution of the reconstructed video stream is smaller than the original resolution of the original video stream;
and sending the reconstructed video stream to a playing end.
2. The method of claim 1, wherein reconstructing the original video stream in real time according to the recognition result to obtain a reconstructed video stream comprises:
and if the target object is identified, reconstructing the original video stream according to the position of the target object to obtain a reconstructed video stream, wherein the reconstructed frame is smaller than the original frame, and the frame of the reconstructed video stream contains the target object.
3. The method of claim 2, wherein reconstructing the original video stream according to the position of the target object to obtain a reconstructed video stream comprises:
and if the identified target object is completely located in a preset area of the reconstruction frame, reconstructing the original video stream according to the position of the target object and the preset area to obtain a reconstruction video stream, wherein the central point of the preset area is coincident with the central point of the reconstruction frame, and the current reconstruction position of the reconstruction frame in the original frame is the same as the reconstruction position of the reconstruction frame during the last reconstruction.
4. A method according to claim 2 or 3, wherein the target objects comprise primary and secondary targets.
5. The method of claim 4, wherein reconstructing the original video stream according to the position of the target object to obtain a reconstructed video stream comprises:
and if the identified target object comprises the main target, reconstructing the original video stream according to the position of the main target to obtain a reconstructed video stream, wherein the main target is positioned in the central area of the picture of the reconstructed video stream.
6. The method of claim 4, wherein reconstructing the original video stream according to the position of the target object to obtain a reconstructed video stream comprises:
and if the identified target object comprises the secondary target and does not comprise the primary target, determining a reconstruction position of the reconstructed picture in the original picture according to the distribution position of the secondary target, and reconstructing the original video stream according to the reconstruction position to obtain a reconstructed video stream.
7. The method according to claim 2 or 3, wherein reconstructing the original video stream in real time according to the recognition result to obtain a reconstructed video stream comprises:
if the target object is not identified, when the original video stream is reconstructed, the current reconstruction position of the reconstructed picture in the original picture is the same as the reconstruction position of the reconstructed picture in the last reconstruction until the target object is identified in the picture of the original video stream.
8. The method according to claim 1, wherein performing object recognition on the picture of the original video stream comprises:
performing frame extraction on the original video stream to obtain a sampling image;
and carrying out target object identification on the sampling image.
9. A computer-readable storage medium, characterized in that the computer-readable storage medium has stored thereon a cloud relay program that, when executed by a processor, implements the cloud relay method according to any one of claims 1 to 8.
10. A cloud relay device, comprising a memory, a processor, and a cloud relay program stored in the memory and executable on the processor, wherein the processor executes the cloud relay program to implement the cloud relay method according to any one of claims 1 to 8.
11. A cloud relay system, comprising: the video acquisition equipment is used for shooting a target area and transmitting an original video stream obtained through shooting to the streaming media server, the streaming media server is used for receiving the original video stream related to the target area and continuously carrying out target object identification on a picture of the original video stream, an identification result comprises the position of a target object in the picture, the original video stream is reconstructed in real time according to the identification result to obtain a reconstructed video stream, so that the original video stream is simulated to move and zoom in the frame of the original video stream through the reconstructed video stream, the reconstructed frame of the reconstructed video stream is smaller than the original frame of the original video stream, the reconstructed resolution of the reconstructed video stream is smaller than the original resolution of the original video stream, and the reconstructed video stream is sent to a playing end.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211169502.8A CN115767113B (en) | 2022-09-22 | 2022-09-22 | Cloud rebroadcasting method, device, medium and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211169502.8A CN115767113B (en) | 2022-09-22 | 2022-09-22 | Cloud rebroadcasting method, device, medium and system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN115767113A true CN115767113A (en) | 2023-03-07 |
| CN115767113B CN115767113B (en) | 2023-09-01 |
Family
ID=85351920
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211169502.8A Active CN115767113B (en) | 2022-09-22 | 2022-09-22 | Cloud rebroadcasting method, device, medium and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115767113B (en) |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2018042105A (en) * | 2016-09-07 | 2018-03-15 | 東芝テリー株式会社 | Monitoring image processing device and monitoring image processing method |
| CN109040780A (en) * | 2018-08-07 | 2018-12-18 | 北京优酷科技有限公司 | A kind of method for processing video frequency and server |
| CN109640150A (en) * | 2018-10-26 | 2019-04-16 | 广州华多网络科技有限公司 | Method of adjustment, equipment and the device of reference video picture |
| CN110062252A (en) * | 2019-04-30 | 2019-07-26 | 广州酷狗计算机科技有限公司 | Live broadcasting method, device, terminal and storage medium |
| CN114222065A (en) * | 2021-12-20 | 2022-03-22 | 北京奕斯伟计算技术有限公司 | Image processing method, image processing apparatus, electronic device, storage medium, and program product |
| CN114401417A (en) * | 2022-01-28 | 2022-04-26 | 广州方硅信息技术有限公司 | Live stream object tracking method and device, equipment and medium thereof |
-
2022
- 2022-09-22 CN CN202211169502.8A patent/CN115767113B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2018042105A (en) * | 2016-09-07 | 2018-03-15 | 東芝テリー株式会社 | Monitoring image processing device and monitoring image processing method |
| CN109040780A (en) * | 2018-08-07 | 2018-12-18 | 北京优酷科技有限公司 | A kind of method for processing video frequency and server |
| CN109640150A (en) * | 2018-10-26 | 2019-04-16 | 广州华多网络科技有限公司 | Method of adjustment, equipment and the device of reference video picture |
| CN110062252A (en) * | 2019-04-30 | 2019-07-26 | 广州酷狗计算机科技有限公司 | Live broadcasting method, device, terminal and storage medium |
| CN114222065A (en) * | 2021-12-20 | 2022-03-22 | 北京奕斯伟计算技术有限公司 | Image processing method, image processing apparatus, electronic device, storage medium, and program product |
| CN114401417A (en) * | 2022-01-28 | 2022-04-26 | 广州方硅信息技术有限公司 | Live stream object tracking method and device, equipment and medium thereof |
Also Published As
| Publication number | Publication date |
|---|---|
| CN115767113B (en) | 2023-09-01 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CA2798298C (en) | Systems and methods for video processing | |
| CN106789991B (en) | Multi-person interactive network live broadcast method and system based on virtual scene | |
| US8665374B2 (en) | Interactive video insertions, and applications thereof | |
| CN103428461B (en) | System and method for recording teaching video | |
| JP6267961B2 (en) | Image providing method and transmitting apparatus | |
| RU2666137C2 (en) | Video product production method and system | |
| CN107105315A (en) | Live broadcasting method, the live broadcasting method of main broadcaster's client, main broadcaster's client and equipment | |
| CN106165393A (en) | Method and system for automatic television production | |
| US20200304707A1 (en) | System and Camera Device for Capturing Images | |
| CN106658032B (en) | Multi-camera live broadcasting method and system | |
| JP2015514339A (en) | System and method for video recording and webcasting of sports events | |
| CN107124662A (en) | Net cast method, device, electronic equipment and computer-readable recording medium | |
| CN103634503A (en) | Video manufacturing method based on face recognition and behavior recognition and video manufacturing method based on face recognition and behavior recognition | |
| CN110087081A (en) | Method for video coding, device, server and storage medium | |
| JP2010232814A (en) | Video editing program, and video editing device | |
| CN111953910A (en) | Video processing method and device based on artificial intelligence and electronic equipment | |
| CN114449252A (en) | Method, device, equipment, system and medium for dynamically adjusting live video based on explication audio | |
| CN110798692A (en) | Video live broadcast method, server and storage medium | |
| CN114025183A (en) | Live broadcast method, device, equipment, system and storage medium | |
| CN115734007B (en) | Video editing method, device, medium and video processing system | |
| CN115767113A (en) | Cloud rebroadcasting method, device, medium and system | |
| US11736775B1 (en) | Artificial intelligence audio descriptions for live events | |
| Drucker et al. | Spectator games: A new entertainment modality of networked multiplayer games | |
| CN110213611A (en) | A kind of ball competition field camera shooting implementation method based on artificial intelligence Visual identification technology | |
| KR20210158362A (en) | Method for providing virtual sports stadium service to watch a game in realtime in virtual space of sports stadium and system therefore |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |