Method, device, equipment and medium for obtaining motion trajectory

Abstract

The present application relates to artificial intelligence technology and provides a method, apparatus, device, and medium for obtaining motion trajectories. The method includes: obtaining 3D video data and first label information of the 3D video data, the first label information including: a first object identifier of a main moving object in the 3D video data, a second object identifier of an auxiliary moving object, and first scene information; identifying the motion trajectory of the main moving object corresponding to the first object identifier to obtain the motion trajectory of the main moving object; processing the first scene information, the second object identifier, and the motion trajectory of the main moving object using a trained composition model to obtain the motion trajectory of the auxiliary moving object corresponding to the second object identifier; and assisting in filming based on the predicted motion trajectory of the auxiliary moving object, thereby achieving efficient and convenient production of 3D video data, enhancing the viewer's perception of changes in scene features, and enabling the viewer to experience more varied and intense dynamic 3D scenes.

Description

Technical Field

[0001] The present application relates to the field of artificial intelligence technology, and in particular to a method, device, equipment and medium for obtaining a motion trajectory. Background Art

[0002] The advantage of 3D video data over 2D video data lies in the inclusion of 3D depth information, allowing for a more realistic scene presentation. However, if objects remain at the same depth in 3D video data, viewers will have difficulty perceiving changes in their features. To address this issue, a professional 3D animation team is required to model characters and scenes, which is difficult for non-professionals to perform, making the production of 3D video data challenging. Therefore, efficiently and conveniently producing 3D video data while ensuring its quality is a pressing technical challenge. Summary of the Invention

[0003] The embodiments of the present application provide a method, apparatus, device, and medium for obtaining motion trajectories, which can assist in filming based on the predicted motion trajectories of auxiliary moving objects, achieve efficient and convenient production of 3D video data, and enhance the viewer's perception of changes in scene features, enabling the viewer to experience more varied and intense dynamic 3D scenes.

[0004] In one aspect, an embodiment of the present application provides a method for obtaining a motion trajectory, the method comprising:

[0005] Acquire 3D video data and first tag information of the 3D video data, where the first tag information includes: a first object identifier of a main moving object in the 3D video data, a second object identifier of an auxiliary moving object, and first scene information;

[0006] Performing motion trajectory recognition on the main moving object corresponding to the first object identifier to obtain the motion trajectory of the main moving object;

[0007] The first scene information, the second object identifier, and the motion trajectory of the main moving object are processed by the trained composition model to obtain the motion trajectory of the auxiliary moving object corresponding to the second object identifier.

[0008] In one embodiment, the first object identifier and the first scene information are obtained by user annotation;

[0009] The specific implementation process of obtaining the first tag information of 3D video data is as follows:

[0010] Identify the 3D video data according to the first object identifier to obtain multiple frames of foreground images and background images, where the multiple frames of foreground images include a main moving object corresponding to the first object identifier;

[0011] Performing auxiliary moving object recognition on the background image to obtain a second object identifier of the auxiliary moving object;

[0012] First label information including a first object identifier, a second object identifier, and first scene information is generated, where the first label information is first label information of 3D video data.

[0013] In one embodiment, the specific implementation process of obtaining the first tag information of the 3D video data is as follows:

[0014] Recognize 3D video data to obtain multiple frames of foreground images and background images;

[0015] Performing main moving object recognition on multiple frames of foreground images to obtain a first object identifier of the main moving object;

[0016] Performing scene recognition on the background image to obtain first scene information;

[0017] Performing auxiliary moving object recognition on the background image to obtain a second object identifier of the auxiliary moving object;

[0018] First label information including a first object identifier, a second object identifier, and first scene information is generated, where the first label information is first label information of 3D video data.

[0019] In one embodiment, after performing auxiliary moving object recognition on the background image and obtaining the second object identification of the auxiliary moving object, the following process may be performed:

[0020] Outputting first prompt information, where the first prompt information is used to prompt the user whether to use the object corresponding to the second object identifier as an auxiliary moving object;

[0021] If the confirmation information is received, the first tag information including the first object identifier, the second object identifier and the first scene information is generated;

[0022] If a non-confirmation message is received, a second object identification submission interface is output;

[0023] Responding to a user operation on the second object identifier submission interface, obtaining a target second object identifier;

[0024] First label information including a first object identifier, a target second object identifier, and first scene information is generated, where the first label information is first label information of 3D video data.

[0025] In one embodiment, the following process may also be implemented:

[0026] Output recommendation information, where the recommendation information includes at least the motion trajectory of the auxiliary moving object, and the recommendation information further includes one or more of the following: a second object identifier, a target video segment in the 3D video data, each frame image in the target video segment includes the auxiliary moving object, and the target video segment matches the motion trajectory of the auxiliary moving object.

[0027] In one embodiment, the following process may also be implemented:

[0028] Based on the motion trajectory of the auxiliary moving object, the 3D video data is segmented to obtain a target video segment, each frame of the target video segment includes the auxiliary moving object, and the target video segment matches the motion trajectory of the auxiliary moving object;

[0029] performing image processing on the target video segment based on the motion trajectory of the auxiliary moving object to obtain a processed video segment, wherein the motion trajectory of the auxiliary moving object in the processed video segment matches the motion trajectory of the auxiliary moving object;

[0030] splicing the processed video clips with video clips other than the target video clip in the 3D video data to obtain target 3D video data;

[0031] Output target 3D video data.

[0032] In one embodiment, the following process may also be implemented:

[0033] Acquire a training sample, where the training sample includes training video data and second label information of the training video data, where the second label information includes: a third object identifier of a main moving object in the training video data, a fourth object identifier of an auxiliary moving object, and second scene information;

[0034] performing motion trajectory recognition on the auxiliary moving object corresponding to the fourth object identifier to obtain the motion trajectory of the auxiliary moving object corresponding to the fourth object identifier;

[0035] performing motion trajectory recognition on the main moving object corresponding to the third object identifier to obtain the motion trajectory of the main moving object corresponding to the third object identifier;

[0036] Processing the motion trajectory of the main moving object corresponding to the second scene information, the fourth object identifier, and the third object identifier through the composition model to obtain the motion trajectory of the auxiliary moving object corresponding to the fourth object identifier;

[0037] The composition model is trained according to the motion trajectory of the auxiliary motion object corresponding to the fourth object identification obtained by motion trajectory recognition and the motion trajectory of the auxiliary motion object corresponding to the fourth object identification obtained by composition model processing to obtain a trained composition model.

[0038] On the other hand, an embodiment of the present application provides a motion trajectory acquisition device, the motion trajectory acquisition device comprising:

[0039] an acquiring unit, configured to acquire 3D video data and first tag information of the 3D video data, the first tag information including: a first object identifier of a main moving object in the 3D video data, a second object identifier of an auxiliary moving object, and first scene information;

[0040] A processing unit, configured to identify a motion trajectory of a main moving object corresponding to the first object identifier to obtain a motion trajectory of the main moving object;

[0041] The processing unit is further configured to process the first scene information, the second object identifier, and the motion trajectory of the main moving object using the trained composition model to obtain the motion trajectory of the auxiliary moving object corresponding to the second object identifier.

[0042] On the other hand, an embodiment of the present application provides an electronic device, including a processor, a memory and a communication interface, which are interconnected, wherein the memory is used to store a computer program that supports a terminal to execute the above method, the computer program includes program instructions, and the processor is configured to call the program instructions to perform the following steps: obtaining 3D video data and first label information of the 3D video data, the first label information including: a first object identifier of a main moving object in the 3D video data, a second object identifier of an auxiliary moving object and first scene information; identifying the motion trajectory of the main moving object corresponding to the first object identifier to obtain the motion trajectory of the main moving object; processing the first scene information, the second object identifier and the motion trajectory of the main moving object through a trained composition model to obtain the motion trajectory of the auxiliary moving object corresponding to the second object identifier.

[0043] On the other hand, an embodiment of the present application provides a computer-readable storage medium, which stores a computer program. The computer program includes program instructions. When the program instructions are executed by a processor, the processor executes the above-mentioned method for obtaining the motion trajectory.

[0044] In an embodiment of the present application, 3D video data and first label information of the 3D video data are obtained, where the first label information includes: a first object identifier of a main moving object in the 3D video data, a second object identifier of an auxiliary moving object, and first scene information; motion trajectory identification is performed on the main moving object corresponding to the first object identifier to obtain the motion trajectory of the main moving object; the first scene information, the second object identifier, and the motion trajectory of the main moving object are processed by a trained composition model to obtain the motion trajectory of the auxiliary moving object corresponding to the second object identifier; and assisted shooting can be achieved based on the predicted motion trajectory of the auxiliary moving object without the need for a professional 3D animation production team to re-model the characters and the scene. 3D video data can be produced efficiently and conveniently, and the viewer's perception of changes in scene features can be enhanced, prompting the viewer to experience more varied and intense dynamic 3D scenes. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0046] Figure 1 This is a flow chart of a method for obtaining a motion trajectory provided in an embodiment of the present application;

[0047] Figure 2 is a schematic diagram of an image provided in an embodiment of the present application;

[0048] Figure 3 1 is a flow chart of a method for training a composition model provided in an embodiment of the present application;

[0049] Figure 4 This is a schematic diagram of the architecture of a training system for a composition model provided in an embodiment of the present application;

[0050] Figure 5 1 is a schematic structural diagram of a motion trajectory acquisition device provided in an embodiment of the present application;

[0051] Figure 6 This is a structural diagram of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0052] In a 3D scene, the presence of relative depth information can enhance the perception of changes in the absolute depth of an object, but objects at the same relative depth position are easily identified as the same object. Therefore, in order to enhance the viewer's perception of changes in scene features, auxiliary objects can be reasonably added around the motion trajectory of the main moving object, prompting the viewer to experience a more varied and intense dynamic 3D scene. Based on this, an embodiment of the present application trains a composition model, which can input the scene information of the 3D video data, the motion trajectory of the main moving object, and the second object identifier of the auxiliary moving object into the trained composition model to obtain the motion trajectory of the auxiliary moving object. In one example, the motion trajectory of the auxiliary moving object can conveniently and efficiently assist the 3D animation production team to provide shooting suggestions for the placement and motion trajectory of the auxiliary moving object when producing intense or varied scenes (such as the flight of birds, explosions, and large-scale fighting). The shooting suggestion can improve the utilization of depth information, ensure visual effects, and help 3D video data to exert its greater sensory enhancement capabilities, thereby saving unnecessary manpower and material costs when shooting large-scale movies. In another example, for small-scale film shoots, even without good art direction, a more reasonable 3D dynamic scene setting can be quickly performed based on the motion trajectory of the auxiliary moving object, thereby improving the utilization of depth information and ensuring the visual effect. In another example, the 3D video data can be updated based on the predicted motion trajectory of the auxiliary moving object, so that the motion trajectory of the auxiliary moving object in the updated 3D video data matches the motion trajectory of the auxiliary moving object. The above examples can achieve efficient and convenient production of 3D video data, and enhance the viewer's perception of changes in scene features, prompting the viewer to experience more varied and intense dynamic 3D scenes.

[0053] The 3D video data may include 3D movies, 3D animations, or 3D video clips.

[0054] The embodiments of the present application can acquire and process relevant data based on artificial intelligence technology. Artificial Intelligence (AI) is the theory, method, technology, and application system that uses digital computers or machines controlled by digital computers to simulate, extend, and expand human intelligence, perceive the environment, acquire knowledge, and use knowledge to achieve optimal results.

[0055] Fundamental AI technologies generally include sensors, dedicated AI chips, cloud computing, distributed storage, big data processing, operating / interaction systems, and mechatronics. AI software technologies primarily encompass computer vision, robotics, biometrics, speech processing, natural language processing, and machine learning / deep learning.

[0056] See Figure 1 , Figure 1 is a flow chart of a method for obtaining a motion trajectory provided in an embodiment of the present application; Figure 1 The method for obtaining the motion trajectory shown can be executed by the first electronic device. The method includes but is not limited to steps S101 to S103, wherein:

[0057] S101 , acquiring 3D video data and first tag information of the 3D video data, where the first tag information includes: a first object identifier of a main moving object in the 3D video data, a second object identifier of an auxiliary moving object, and first scene information.

[0058] Among them, the first electronic device can be any one or more of a smart phone, a tablet computer, a laptop computer, a desktop computer, an intelligent vehicle-mounted device, and an intelligent wearable device. Optionally, the first electronic device can also be a server, which can be an independent physical server, or a server cluster or distributed system composed of multiple physical servers. In other words, the server can be an independent server, or it can be a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, content delivery networks (CDNs), and big data and artificial intelligence platforms.

[0059] Taking the first electronic device as any one or more of a smart phone, a tablet computer, a laptop computer, a desktop computer, a smart car device and a smart wearable device as an example, the trained composition model can be run in the first electronic device. Assuming that a user (such as a 3D animator) wants to give a motion trajectory suggestion of a specified auxiliary moving object based on a certain 3D video data through the trained composition model, the user can submit the 3D video data to the first electronic device and manually mark the main moving object, auxiliary moving object and scene in the 3D video data in the first electronic device, so that the electronic device can obtain the first label information of the 3D video data, that is, the first object identification of the main moving object, the second object identification of the auxiliary moving object, and the first scene information. Optionally, the 3D video data can be uploaded to the first electronic device by the user through the data transmission interface, or downloaded by the first electronic device through the Internet in response to the user's download instruction, or pre-stored in the first electronic device, which is not limited by the embodiment of the present application. After the first electronic device obtains the 3D video data and the first label information of the 3D video data, it can further perform steps S102 and S103.

[0060] Taking the first electronic device as an example, assuming that a user (e.g., a 3D animator) wants to use a trained composition model to give motion trajectory suggestions for a specified auxiliary moving object based on certain 3D video data, the user can log in to the client corresponding to the server through an account, upload the 3D video data through the client, and manually annotate the main moving object, auxiliary moving object, and scene in the 3D video data in the client. The client can then obtain the first label information of the 3D video data, and the client can send the 3D video data and the first label information of the 3D video data to the server. After the server obtains the 3D video data and the first label information of the 3D video data, it can further execute steps S102 and S103.

[0061] In an embodiment of the present application, the first object identification of the main moving object, the second object identification of the auxiliary moving object, and the first scene information are obtained through manual labeling by the user, which can ensure that the motion trajectory of the auxiliary moving object obtained through the trained composition model meets the user's wishes and improve the accuracy of the motion trajectory of the auxiliary moving object.

[0062] As a feasible implementation manner, the first tag information of the 3D video data in the embodiment of the present application may be obtained through manual labeling by a user, or may be automatically recognized by the first electronic device.

[0063] In a specific implementation, after acquiring 3D video data, the first electronic device may identify the 3D video data to obtain multiple frames of foreground images and background images. The first electronic device may then identify a main moving object in the multiple frames of foreground images to obtain a first object identifier for the main moving object. For example, the first electronic device may identify an object that appears in most of the multiple frames of foreground images as the main moving object, and then obtain the first object identifier for the main moving object. In another example, the first electronic device may first determine at least one object that appears in most of the multiple frames of foreground images, and then determine a target object within the at least one object, where the position of the target object changes between different images in the 3D video data. The first electronic device may then identify the target object as the main moving object and obtain the first object identifier for the main moving object. In another example, the first electronic device may first determine at least one object that appears in most of the multiple frames of foreground images, and then determine the target object within the at least one object, where the depth information of the target object changes between different images in the 3D video data. The first electronic device may then identify the target object as the main moving object and obtain the first object identifier for the main moving object.

[0064] At the same time, the first electronic device can perform auxiliary moving object recognition on the background image to obtain the second object identification of the auxiliary moving object. For example, the first electronic device can identify the auxiliary moving object in the background image based on the image and background image where the main moving object is located in the 3D video data, and then obtain the second object identification of the auxiliary moving object. Since the inventive concept of this application is: in order to enhance the viewer's perception of changes in scene features, auxiliary objects can be reasonably added around the motion trajectory of the main moving object, prompting the viewer to experience a more varied and intense dynamic 3D scene. Then the auxiliary moving object is usually located around the main moving object. Based on this, the first electronic device can determine the candidate object whose distance from the main moving object is less than a preset distance threshold in the image where the main moving object is located, and then determine whether the candidate object is in the background image. If the candidate object is in the background image, the candidate object is identified as an auxiliary moving object, and then the second object identification of the auxiliary moving object is obtained.

[0065] At the same time, the first electronic device can perform scene recognition on the background image to obtain first scene information. For example, the first electronic device can identify the scene in the background image based on the image of the main moving object and the background image in the 3D video data, and then obtain the first scene information. For example, the first electronic device can determine the position of the main moving object and the position of the auxiliary moving object in the image where the main moving object is located, and then determine the candidate scene where the main moving object and the auxiliary moving object are located in the image where the main moving object is located based on the position of the main moving object and the position of the auxiliary moving object. If the candidate scene is located in the background image, the candidate scene is identified as a scene, and then the first scene information of the scene is obtained.

[0066] Furthermore, after acquiring the first object identification of the main moving object, the second object identification of the auxiliary moving object, and the first scene information, the first electronic device can generate the first tag information of the 3D video data, and the first tag information includes the first object identification of the main moving object, the second object identification of the auxiliary moving object, and the first scene information. The first object identification information is used to uniquely identify the main moving object, such as the object name of the main moving object, the position of the main moving object in a certain frame image of the 3D video data, or the description information for describing the main moving object (such as the color, shape or size of the main moving object). The second object identification information is used to uniquely identify the auxiliary moving object, such as the object name of the auxiliary moving object, the position of the auxiliary moving object in a certain frame image of the 3D video data, or the description information for describing the auxiliary moving object. The first scene information is used to indicate the scene in which the main moving object and the auxiliary moving object are located. Exemplarily, with Figure 2Taking the schematic diagram of the image shown as an example, the image is a frame image in 3D video data, the main moving object may be a flying bird, the auxiliary moving objects may be the trees that the bird flies through, and the scene may be an "open street".

[0067] In an embodiment of the present application, compared with the first label information obtained by manual labeling by the user, the present application does not require user operation and can realize automatic recognition of the main moving object, auxiliary moving object and scene, and then generate the first label information based on the recognition result, which can improve the efficiency of obtaining the motion trajectory of the auxiliary moving object.

[0068] As a feasible implementation method, after the first electronic device automatically identifies the first object identifier, it can confirm with the user whether to use the first object identifier identified by the first electronic device as the first object identifier of the main moving object. The specific implementation process is: after the first electronic device identifies the first object identifier, it can output a first prompt message, and the first prompt message is used to prompt the user whether to use the object corresponding to the first object identifier as the main moving object. If the user clicks the "Confirm" button or replies to the field "Yes", the first object identifier is determined as the first object identifier of the main moving object; if the user clicks the "Reject" button or replies to the field "No", the first object identifier submission interface is output, so that the user can submit the first object identifier by input or selection in the first object identifier submission interface, and then the first object identifier submitted by the user is determined as the first object identifier of the main moving object.

[0069] As a feasible implementation method, after the first electronic device automatically identifies the second object identifier, it can confirm with the user whether to use the second object identifier identified by the first electronic device as the second object identifier of the auxiliary moving object. The specific implementation process is: after the first electronic device identifies the second object identifier, it can output a second prompt information, and the second prompt information is used to prompt the user whether to use the object corresponding to the second object identifier as the auxiliary moving object. If the user clicks the "Confirm" button or replies to the field "Yes", the second object identifier is determined as the second object identifier of the auxiliary moving object; if the user clicks the "Reject" button or replies to the field "No", the second object identifier submission interface is output, so that the user can submit the second object identifier by input or selection in the second object identifier submission interface, and then the second object identifier submitted by the user is determined as the second object identifier of the auxiliary moving object.

[0070] As a feasible implementation method, after the first electronic device automatically identifies the first scene information, it can confirm with the user whether the first scene information identified by the first electronic device is correct. The specific implementation process is: after the first electronic device identifies the first scene information, it can output a third prompt information, and the third prompt information is used to prompt the user whether the scene corresponding to the first scene information is correct. If the user clicks the "Confirm" button or replies to the field "Yes", it is determined that the scene information is correct, and then the first tag information containing the first scene information can be generated; if the user clicks the "Reject" button or replies to the field "No", the scene information submission interface is output, so that the user can submit the scene information by input or selection in the scene information submission interface, and then the scene information submitted by the user is used as the first scene information.

[0071] S102 : Identify the motion trajectory of the main moving object corresponding to the first object identifier to obtain the motion trajectory of the main moving object.

[0072] The first electronic device may use image tracking technology to identify the motion trajectory of the main moving object corresponding to the first object identifier to obtain the motion trajectory of the main moving object. In a specific implementation, the first electronic device may determine the main moving object corresponding to the first object identifier, then obtain the position of the main moving object in each frame of the 3D video data, and obtain the motion trajectory of the main moving object based on the position of the main moving object in each frame and the playback order of the frames.

[0073] S103: Process the first scene information, the second object identifier, and the motion trajectory of the main moving object through the trained composition model to obtain the motion trajectory of the auxiliary moving object corresponding to the second object identifier.

[0074] In a specific implementation, the first electronic device can input the first scene information, the second object identification and the motion trajectory of the main moving object into the trained composition model, and process the first scene information, the second object identification and the motion trajectory of the main moving object through the trained composition model to obtain the motion trajectory of the auxiliary moving object corresponding to the second object identification. The 3D animation production team can then place the auxiliary moving object based on the suggestion of the motion trajectory of the auxiliary moving object given by the trained composition model, and set the motion mode of the auxiliary object based on the suggestion of the motion trajectory of the auxiliary moving object given by the trained composition model, and shoot the auxiliary moving object based on the motion mode of the auxiliary object. Alternatively, the 3D animation production team can adjust the 3D video data based on the suggestion of the motion trajectory of the auxiliary moving object given by the trained composition model, so that the motion trajectory of the auxiliary moving object in the adjusted 3D video data matches the motion trajectory of the auxiliary moving object given based on the trained composition model.

[0075] As a feasible implementation, the first electronic device can use a trained composition model to provide a video clip that matches the motion trajectory of the auxiliary moving object, so that the 3D animator can process the video clip based on the motion trajectory of the auxiliary object provided by the trained composition model, thereby obtaining a processed video clip in which the motion trajectory of the auxiliary moving object in the processed video clip matches the motion trajectory of the auxiliary moving object provided based on the trained composition model. The first electronic device then splices the processed video clip with other video clips in the 3D video data to obtain new 3D video data, so that the new 3D video data presents a more exciting and exciting scene to the audience, enhancing the immersive viewing experience.

[0076] Exemplarily, the way in which the first electronic device provides a video clip that matches the motion trajectory of the auxiliary moving object through the trained composition model can be: the first electronic device searches for an image containing the auxiliary moving object in the 3D video data through the trained composition model, and then determines a target image that matches the motion trajectory of the auxiliary moving object in the image containing the auxiliary moving object, wherein the target image is continuous in time, thereby obtaining a video clip containing the target image.

[0077] As a feasible implementation method, the electronic device can also output the second object identification of the auxiliary moving object through the trained composition model, so that the 3D animation production team can know that the moving object corresponding to the second object identification is the auxiliary moving object suggestion given by the trained composition model.

[0078] In an embodiment of the present application, 3D video data and first label information of the 3D video data are obtained, where the first label information includes: a first object identifier of a main moving object in the 3D video data, a second object identifier of an auxiliary moving object, and first scene information; motion trajectory identification is performed on the main moving object corresponding to the first object identifier to obtain the motion trajectory of the main moving object; the first scene information, the second object identifier, and the motion trajectory of the main moving object are processed by a trained composition model to obtain the motion trajectory of the auxiliary moving object corresponding to the second object identifier; auxiliary shooting can be performed based on the predicted motion trajectory of the auxiliary moving object, thereby achieving efficient and convenient production of 3D video data, and enhancing the viewer's perception of changes in scene features, prompting the viewer to experience more varied and intense dynamic 3D scenes.

[0079] See Figure 3 , Figure 3 is a flow chart of a method for training a composition model provided in an embodiment of the present application; the method for training a composition model may be performed by a second electronic device, and the scheme may include but is not limited to steps S301 to S304, wherein:

[0080] S301, obtaining a training sample, where the training sample includes training video data and second label information of the training video data, where the second label information includes: a third object identifier of a main moving object in the training video data, a fourth object identifier of an auxiliary moving object, and second scene information.

[0081] by Figure 4 Taking the architectural diagram of the training system for the composition model shown as an example, assuming that the training video data is the original video, the second electronic device can obtain the original video and obtain the third object identifier of the main moving object, the fourth object identifier of the auxiliary moving object, and the second scene information obtained by manually annotating the original video. The second electronic device can use image tracking technology to identify the motion trajectory of the main moving object corresponding to the third object identifier to obtain the motion trajectory of the main moving object, and use image tracking technology to identify the motion trajectory of the auxiliary moving object corresponding to the fourth object identifier to obtain the motion trajectory of the auxiliary moving object. The second electronic device obtains a training sample, and the training sample includes the manually annotated second scene information, the fourth object identifier of the auxiliary moving object, the motion trajectory of the main moving object, and the motion trajectory of the auxiliary moving object. The second electronic device then inputs the second scene information, the fourth object identifier of the auxiliary moving object, and the motion trajectory of the main moving object in the training sample into the composition model, processes the second scene information, the fourth object identifier, and the motion trajectory of the main moving object through the composition model, and obtains the auxiliary object motion trajectory. The auxiliary object motion trajectory obtained by the composition model is then compared with the auxiliary object motion trajectory identified by the image tracking technology to obtain a loss value of the composition model. The composition model is trained based on the loss value to obtain a trained composition model. Exemplarily, the composition model can be a DensentNet convolutional neural network structure.

[0082] Optionally, one or more of the third object identifier of the main moving object, the fourth object identifier of the auxiliary moving object, or the second scene information may not be manually labeled, but may be automatically recognized by the second electronic device. The manner in which the second electronic device automatically recognizes the third object identifier of the main moving object, the fourth object identifier of the auxiliary moving object, or the second scene information can be found in Figure 1 The related description of the first electronic device automatically identifying the first object identifier of the main moving object, the second object identifier of the auxiliary moving object or the first scene information is not repeated here.

[0083] The second electronic device and the first electronic device may be the same device or different devices, and are not limited by the embodiments of the present application. The second electronic device may be any one or more of a smartphone, a tablet computer, a laptop computer, a desktop computer, a smart car device, and a smart wearable device. Optionally, the second electronic device may also be a server, which may be an independent physical server, a server cluster composed of multiple physical servers, or a distributed system.

[0084] S302: Perform motion trajectory recognition on the auxiliary moving object corresponding to the fourth object identifier to obtain the motion trajectory of the auxiliary moving object corresponding to the fourth object identifier.

[0085] The second electronic device identifies the motion trajectory of the auxiliary moving object corresponding to the fourth object identifier. The method of obtaining the motion trajectory of the auxiliary moving object corresponding to the fourth object identifier can be referred to in Figure 1 The first electronic device identifies the motion trajectory of the main moving object corresponding to the first object identifier and obtains a description of the motion trajectory of the main moving object corresponding to the first object identifier, which will not be repeated here.

[0086] S303 : Perform motion trajectory recognition on the main moving object corresponding to the third object identifier to obtain the motion trajectory of the main moving object corresponding to the third object identifier.

[0087] The second electronic device identifies the motion trajectory of the main moving object corresponding to the third object identifier. The method of obtaining the motion trajectory of the main moving object corresponding to the third object identifier can be found in Figure 1 The first electronic device identifies the motion trajectory of the main moving object corresponding to the first object identifier and obtains a description of the motion trajectory of the main moving object corresponding to the first object identifier, which will not be repeated here.

[0088] S304 : Process the motion trajectory of the main moving object corresponding to the second scene information, the fourth object identifier, and the third object identifier through the composition model to obtain the motion trajectory of the auxiliary moving object corresponding to the fourth object identifier.

[0089] In a specific implementation, the second electronic device can input the motion trajectory of the main moving object corresponding to the second scene information, the fourth object identification and the third object identification into the composition model, and process the motion trajectory of the main moving object corresponding to the second scene information, the fourth object identification and the third object identification through the composition model to obtain the motion trajectory of the auxiliary moving object corresponding to the fourth object identification.

[0090] S305 , training the composition model according to the motion trajectory of the auxiliary motion object corresponding to the fourth object identifier obtained by motion trajectory recognition and the motion trajectory of the auxiliary motion object corresponding to the fourth object identifier obtained by composition model processing to obtain a trained composition model.

[0091] In a specific implementation, the second electronic device can compare the motion trajectory of the auxiliary moving object corresponding to the fourth object identifier obtained by processing the composition model with the motion trajectory of the auxiliary moving object corresponding to the fourth object identifier obtained by motion trajectory recognition to obtain a loss value of the composition model, and train the composition model based on the loss value to obtain a trained composition model. In other words, the second electronic device can use the motion trajectory of the main moving object corresponding to the second scene information, the fourth object identifier, and the third object identifier as model input, and the motion trajectory of the auxiliary moving object corresponding to the fourth object identifier obtained by motion trajectory recognition as the expected output, and train the composition model based on the training samples to obtain a trained composition model.

[0092] In an embodiment of the present application, a training sample is obtained, the training sample includes training video data and second label information of the training video data, the second label information includes: a third object identifier of the main moving object in the training video data, a fourth object identifier of the auxiliary moving object and second scene information, a motion trajectory of the auxiliary moving object corresponding to the fourth object identifier is identified, and a motion trajectory of the auxiliary moving object corresponding to the fourth object identifier is obtained, a motion trajectory of the main moving object corresponding to the third object identifier is identified, and a motion trajectory of the main moving object corresponding to the third object identifier is obtained, and the motion trajectory of the main moving object corresponding to the second scene information, the fourth object identifier and the third object identifier is identified through a composition model. Processing is performed to obtain the motion trajectory of the auxiliary moving object corresponding to the fourth object identifier. The motion trajectory of the auxiliary moving object corresponding to the fourth object identifier obtained by motion trajectory recognition and the motion trajectory of the auxiliary moving object corresponding to the fourth object identifier obtained by composition model processing are used to train the composition model to obtain a trained composition model. The composition model trained in the above manner can accurately predict the motion trajectory of the auxiliary moving object, and then assist in shooting based on the predicted motion trajectory of the auxiliary moving object, which can realize efficient and convenient production of 3D video data, and effectively enhance the viewer's perception of changes in scene features, prompting the viewer to feel more varied and intense dynamic 3D scenes.

[0093] An embodiment of the present application further provides a computer storage medium, in which program instructions are stored. When the program instructions are executed, they are used to implement the corresponding methods described in the above embodiments.

[0094] See also Figure 5 , Figure 5It is a structural diagram of a motion trajectory acquisition device provided in an embodiment of the present application.

[0095] In one implementation of the device of the embodiment of the present application, the device includes the following structure.

[0096] An acquiring unit 501 is configured to acquire 3D video data and first tag information of the 3D video data, where the first tag information includes: a first object identifier of a main moving object, a second object identifier of an auxiliary moving object, and first scene information in the 3D video data;

[0097] The processing unit 502 is configured to identify a motion trajectory of the main moving object corresponding to the first object identifier to obtain the motion trajectory of the main moving object;

[0098] The processing unit 502 is further configured to process the first scene information, the second object identifier, and the motion trajectory of the main moving object using the trained composition model to obtain the motion trajectory of the auxiliary moving object corresponding to the second object identifier.

[0099] In one embodiment, the first object identifier and the first scene information are obtained by user annotation;

[0100] The acquiring unit 501 acquires first tag information of the 3D video data, including:

[0101] Identify the 3D video data according to the first object identifier to obtain multiple frames of foreground images and background images, where the multiple frames of foreground images include a main moving object corresponding to the first object identifier;

[0102] Performing auxiliary moving object recognition on the background image to obtain a second object identifier of the auxiliary moving object;

[0103] First label information including a first object identifier, a second object identifier, and first scene information is generated, where the first label information is first label information of 3D video data.

[0104] In one embodiment, the acquiring unit 501 acquires the first tag information of the 3D video data, including:

[0105] Recognize 3D video data to obtain multiple frames of foreground images and background images;

[0106] Performing main moving object recognition on multiple frames of foreground images to obtain a first object identifier of the main moving object;

[0107] Performing scene recognition on the background image to obtain first scene information;

[0108] Performing auxiliary moving object recognition on the background image to obtain a second object identifier of the auxiliary moving object;

[0109] First label information including a first object identifier, a second object identifier, and first scene information is generated, where the first label information is first label information of 3D video data.

[0110] In one embodiment, the apparatus may further include an output unit 503;

[0111] Output unit 503, configured to output first prompt information after processing unit 502 performs auxiliary moving object recognition on the background image and obtains a second object identifier of the auxiliary moving object, the first prompt information being used to prompt a user whether to select the object corresponding to the second object identifier as the auxiliary moving object;

[0112] The acquiring unit 501 is further configured to trigger the generation of first tag information including the first object identifier, the second object identifier, and the first scene information upon receiving the confirmation information;

[0113] The output unit 503 is further configured to output a second object identification submission interface if the acquisition unit 501 receives non-confirmation information;

[0114] The acquiring unit 501 is further configured to acquire a target second object identifier in response to a user operation on the second object identifier submission interface;

[0115] The acquiring unit 501 is further configured to generate first tag information including a first object identifier, a target second object identifier, and first scene information, where the first tag information is first tag information of 3D video data.

[0116] In one embodiment, the apparatus may further include an output unit 503;

[0117] The output unit 503 is used to output recommendation information, where the recommendation information includes at least the motion trajectory of the auxiliary moving object, and the recommendation information also includes one or more of the following: a second object identifier, a target video segment in the 3D video data, each frame image in the target video segment includes the auxiliary moving object, and the target video segment matches the motion trajectory of the auxiliary moving object.

[0118] In one embodiment, the processing unit 502 is further configured to perform segment capture on the 3D video data based on the motion trajectory of the auxiliary moving object to obtain a target video segment, wherein each frame of the target video segment includes the auxiliary moving object, and the target video segment matches the motion trajectory of the auxiliary moving object;

[0119] The processing unit 502 is further configured to perform image processing on the target video segment based on the motion trajectory of the auxiliary moving object to obtain a processed video segment, wherein the motion trajectory of the auxiliary moving object in the processed video segment matches the motion trajectory of the auxiliary moving object;

[0120] The processing unit 502 is further configured to perform splicing processing on the processed video segment with video segments other than the target video segment in the 3D video data to obtain target 3D video data;

[0121] The apparatus may further include an output unit 503 ; the output unit 503 is configured to output target 3D video data.

[0122] In one embodiment, the acquisition unit 501 is further configured to acquire a training sample, where the training sample includes training video data and second label information of the training video data, where the second label information includes: a third object identifier of a main moving object in the training video data, a fourth object identifier of an auxiliary moving object, and second scene information;

[0123] The processing unit 502 is further configured to identify a motion trajectory of the auxiliary moving object corresponding to the fourth object identifier to obtain a motion trajectory of the auxiliary moving object corresponding to the fourth object identifier;

[0124] The processing unit 502 is further configured to identify a motion trajectory of the main moving object corresponding to the third object identifier to obtain a motion trajectory of the main moving object corresponding to the third object identifier;

[0125] The processing unit 502 is further configured to process the motion trajectory of the main moving object corresponding to the second scene information, the fourth object identifier, and the third object identifier using the composition model to obtain the motion trajectory of the auxiliary moving object corresponding to the fourth object identifier;

[0126] The processing unit 502 is further used to train the composition model according to the motion trajectory of the auxiliary moving object corresponding to the fourth object identification obtained by motion trajectory recognition and the motion trajectory of the auxiliary moving object corresponding to the fourth object identification obtained by composition model processing to obtain a trained composition model.

[0127] In an embodiment of the present application, 3D video data and first label information of the 3D video data are obtained, where the first label information includes: a first object identifier of a main moving object in the 3D video data, a second object identifier of an auxiliary moving object, and first scene information; motion trajectory identification is performed on the main moving object corresponding to the first object identifier to obtain the motion trajectory of the main moving object; the first scene information, the second object identifier, and the motion trajectory of the main moving object are processed by a trained composition model to obtain the motion trajectory of the auxiliary moving object corresponding to the second object identifier; auxiliary shooting can be performed based on the predicted motion trajectory of the auxiliary moving object, thereby achieving efficient and convenient production of 3D video data, and enhancing the viewer's perception of changes in scene features, prompting the viewer to experience more varied and intense dynamic 3D scenes.

[0128] See also Figure 6 , Figure 66 is a schematic diagram of the structure of an electronic device provided in an embodiment of the present application. The electronic device in this embodiment of the present application includes a power supply module and other structures, and includes a processor 601, a memory 602, and a communication interface 603. The processor 601, the memory 602, and the communication interface 603 can exchange data, and the processor 601 implements the corresponding data processing solution.

[0129] The memory 602 may include volatile memory, such as random-access memory (RAM); the memory 602 may also include non-volatile memory, such as flash memory, solid-state drive (SSD), etc.; the memory 602 may also include a combination of the above types of memory.

[0130] The processor 601 may be a central processing unit (CPU). The processor 601 may also be a combination of a CPU and a GPU. In an electronic device, multiple CPUs and GPUs may be included as needed to perform corresponding data processing. In one embodiment, the memory 602 is used to store program instructions. The processor 601 may call the program instructions to implement the various methods described above in the embodiments of the present application.

[0131] In a first possible implementation, the processor 601 of the electronic device calls the program instructions stored in the memory 602 to perform the following operations:

[0132] Acquire 3D video data and first tag information of the 3D video data, where the first tag information includes: a first object identifier of a main moving object in the 3D video data, a second object identifier of an auxiliary moving object, and first scene information;

[0133] Performing motion trajectory recognition on the main moving object corresponding to the first object identifier to obtain the motion trajectory of the main moving object;

[0134] The first scene information, the second object identifier, and the motion trajectory of the main moving object are processed by the trained composition model to obtain the motion trajectory of the auxiliary moving object corresponding to the second object identifier.

[0135] In one embodiment, the first object identifier and the first scene information are obtained by user annotation;

[0136] When the processor 601 obtains the first tag information of the 3D video data, it specifically performs the following operations:

[0137] Identify the 3D video data according to the first object identifier to obtain multiple frames of foreground images and background images, where the multiple frames of foreground images include a main moving object corresponding to the first object identifier;

[0138] Performing auxiliary moving object recognition on the background image to obtain a second object identifier of the auxiliary moving object;

[0139] First label information including a first object identifier, a second object identifier, and first scene information is generated, where the first label information is first label information of 3D video data.

[0140] In one embodiment, when the processor 601 obtains the first tag information of the 3D video data, it specifically performs the following operations:

[0141] Recognize 3D video data to obtain multiple frames of foreground images and background images;

[0142] Performing main moving object recognition on multiple frames of foreground images to obtain a first object identifier of the main moving object;

[0143] Performing scene recognition on the background image to obtain the first scene information;

[0144] Performing auxiliary moving object recognition on the background image to obtain a second object identifier of the auxiliary moving object;

[0145] First label information including a first object identifier, a second object identifier, and first scene information is generated, where the first label information is first label information of 3D video data.

[0146] In one embodiment, after performing auxiliary moving object recognition on the background image and obtaining the second object identifier of the auxiliary moving object, the processor 601 may further perform the following operations:

[0147] Outputting first prompt information through the communication interface 603, the first prompt information is used to prompt the user whether to use the object corresponding to the second object identifier as an auxiliary moving object;

[0148] If the confirmation information is received through the communication interface 603, the first tag information including the first object identifier, the second object identifier and the first scene information is generated.

[0149] If non-confirmation information is received through the communication interface 603, the second object identification submission interface is output through the communication interface 603;

[0150] Responding to a user operation on the second object identifier submission interface, obtaining a target second object identifier;

[0151] First label information including a first object identifier, a target second object identifier, and first scene information is generated, where the first label information is first label information of 3D video data.

[0152] In one embodiment, the processor 601 calls the program instructions stored in the memory 602 and is further configured to perform the following operations:

[0153] The recommendation information is output through the communication interface 603. The recommendation information includes at least the motion trajectory of the auxiliary moving object. The recommendation information also includes one or more of the following: a second object identifier, a target video segment in the 3D video data, each frame image in the target video segment includes the auxiliary moving object, and the target video segment matches the motion trajectory of the auxiliary moving object.

[0154] In one embodiment, the processor 601 calls the program instructions stored in the memory 602 and is further configured to perform the following operations:

[0155] Based on the motion trajectory of the auxiliary moving object, the 3D video data is segmented to obtain a target video segment, each frame of the target video segment includes the auxiliary moving object, and the target video segment matches the motion trajectory of the auxiliary moving object;

[0156] performing image processing on the target video segment based on the motion trajectory of the auxiliary moving object to obtain a processed video segment, wherein the motion trajectory of the auxiliary moving object in the processed video segment matches the motion trajectory of the auxiliary moving object;

[0157] splicing the processed video clips with video clips other than the target video clip in the 3D video data to obtain target 3D video data;

[0158] The target 3D video data is output through the communication interface 603 .

[0159] In one embodiment, the processor 601 calls the program instructions stored in the memory 602 and is further configured to perform the following operations:

[0160] Acquire a training sample, where the training sample includes training video data and second label information of the training video data, where the second label information includes: a third object identifier of a main moving object in the training video data, a fourth object identifier of an auxiliary moving object, and second scene information;

[0161] performing motion trajectory recognition on the auxiliary moving object corresponding to the fourth object identifier to obtain the motion trajectory of the auxiliary moving object corresponding to the fourth object identifier;

[0162] performing motion trajectory recognition on the main moving object corresponding to the third object identifier to obtain the motion trajectory of the main moving object corresponding to the third object identifier;

[0163] Processing the motion trajectory of the main moving object corresponding to the second scene information, the fourth object identifier, and the third object identifier through the composition model to obtain the motion trajectory of the auxiliary moving object corresponding to the fourth object identifier;

[0164] The composition model is trained according to the motion trajectory of the auxiliary motion object corresponding to the fourth object identification obtained by motion trajectory recognition and the motion trajectory of the auxiliary motion object corresponding to the fourth object identification obtained by composition model processing to obtain a trained composition model.

[0165] In an embodiment of the present application, 3D video data and first label information of the 3D video data are obtained, where the first label information includes: a first object identifier of a main moving object in the 3D video data, a second object identifier of an auxiliary moving object, and first scene information; motion trajectory identification is performed on the main moving object corresponding to the first object identifier to obtain the motion trajectory of the main moving object; the first scene information, the second object identifier, and the motion trajectory of the main moving object are processed by a trained composition model to obtain the motion trajectory of the auxiliary moving object corresponding to the second object identifier; auxiliary shooting can be performed based on the predicted motion trajectory of the auxiliary moving object, thereby achieving efficient and convenient production of 3D video data, and enhancing the viewer's perception of changes in scene features, prompting the viewer to experience more varied and intense dynamic 3D scenes.

[0166] Those skilled in the art will appreciate that all or part of the processes in the above-described method embodiments can be implemented by instructing the relevant hardware through a computer program. The program can be stored in a computer-readable storage medium. When executed, the program can include the processes of the above-described method embodiments. The storage medium can be a magnetic disk, an optical disk, a read-only memory (ROM), or a random access memory (RAM). The computer-readable storage medium can primarily include a program storage area and a data storage area. The program storage area can store an operating system, at least one application required for a function, and the like; the data storage area can store data created based on the use of the blockchain node.

[0167] The blockchain referred to in this application refers to a new application model for computer technologies such as distributed data storage, peer-to-peer transmission, consensus mechanisms, and encryption algorithms. Blockchain is essentially a decentralized database, a series of data blocks generated using cryptographic methods. Each data block contains information about a batch of network transactions, which is used to verify the validity of the information (to prevent counterfeiting) and generate the next block. Blockchain can include the underlying blockchain platform, the platform product service layer, and the application service layer.

[0168] The above disclosure is only part of the embodiments of the present application, and it is certainly not intended to limit the scope of the rights of the present application. A person skilled in the art can understand that all or part of the processes of the above embodiments and equivalent changes made in accordance with the claims of the present application are still within the scope of the invention.

Claims

1. A method for obtaining a motion trajectory, characterized in that: include: Acquire 3D video data and first tag information of the 3D video data, where the first tag information includes: a first object identifier of a main moving object, a second object identifier of an auxiliary moving object, and first scene information in the 3D video data; performing motion trajectory identification on a main moving object corresponding to the first object identifier to obtain the motion trajectory of the main moving object; Processing the first scene information, the second object identifier, and the motion trajectory of the main moving object using the trained composition model to obtain the motion trajectory of the auxiliary moving object corresponding to the second object identifier; placing the auxiliary moving object based on the motion trajectory suggestion of the auxiliary moving object given by the trained composition model, setting the motion mode of the auxiliary moving object based on the motion trajectory suggestion of the auxiliary moving object, and photographing the auxiliary moving object based on the motion mode of the auxiliary moving object; or The 3D video data is adjusted based on the suggestion of the motion trajectory of the auxiliary moving object, so that the motion trajectory of the auxiliary moving object in the adjusted 3D video data matches the motion trajectory of the auxiliary moving object given based on the trained composition model.

2. The method according to claim 1, wherein The first object identifier and the first scene information are obtained by user annotation; The obtaining of the first tag information of the 3D video data includes: Identifying the 3D video data according to the first object identifier to obtain multiple frames of foreground images and background images, wherein the multiple frames of foreground images include a main moving object corresponding to the first object identifier; performing auxiliary moving object recognition on the background image to obtain a second object identifier of the auxiliary moving object; First tag information including the first object identifier, the second object identifier, and first scene information is generated, where the first tag information is first tag information of the 3D video data.

3. The method according to claim 1, wherein The obtaining of the first tag information of the 3D video data includes: Identifying the 3D video data to obtain multiple frames of foreground images and background images; Performing main moving object recognition on the multiple frames of foreground images to obtain a first object identifier of the main moving object; Performing scene recognition on the background image to obtain the first scene information; performing auxiliary moving object recognition on the background image to obtain a second object identifier of the auxiliary moving object; First tag information including the first object identifier, the second object identifier, and first scene information is generated, where the first tag information is first tag information of the 3D video data.

4. The method according to claim 2 or 3, wherein: After performing auxiliary moving object recognition on the background image to obtain the second object identifier of the auxiliary moving object, the method further includes: outputting first prompt information, where the first prompt information is used to prompt the user whether to use the object corresponding to the second object identifier as an auxiliary moving object; If the confirmation information is received, the process triggers the generation of first tag information including the first object identifier, the second object identifier and the first scene information; If a non-confirmation message is received, a second object identification submission interface is output; Responding to a user operation on the second object identifier submission interface, obtaining a target second object identifier; First label information including the first object identifier, the target second object identifier, and first scene information is generated, where the first label information is first label information of the 3D video data.

5. The method according to claim 2 or 3, wherein: The method further comprises: Output recommendation information, where the recommendation information includes at least the motion trajectory of the auxiliary moving object, and the recommendation information further includes one or more of the following: the second object identifier, a target video segment in the 3D video data, each frame image in the target video segment includes the auxiliary moving object, and the target video segment matches the motion trajectory of the auxiliary moving object.

6. The method according to claim 1, wherein The method further comprises: Based on the motion trajectory of the auxiliary moving object, segmenting the 3D video data to obtain a target video segment, wherein each frame of the target video segment includes the auxiliary moving object, and the target video segment matches the motion trajectory of the auxiliary moving object; performing image processing on the target video segment based on the motion trajectory of the auxiliary moving object to obtain a processed video segment, wherein the motion trajectory of the auxiliary moving object in the processed video segment matches the motion trajectory of the auxiliary moving object; splicing the processed video segment with video segments other than the target video segment in the 3D video data to obtain target 3D video data; The target 3D video data is output.

7. The method according to claim 1, wherein The method further comprises: Acquire a training sample, the training sample including training video data and second label information of the training video data, the second label information including: a third object identifier of a main moving object in the training video data, a fourth object identifier of an auxiliary moving object, and second scene information; performing motion trajectory recognition on the auxiliary moving object corresponding to the fourth object identifier to obtain the motion trajectory of the auxiliary moving object corresponding to the fourth object identifier; performing motion trajectory recognition on the main moving object corresponding to the third object identifier to obtain the motion trajectory of the main moving object corresponding to the third object identifier; Processing the second scene information, the fourth object identifier, and the motion trajectory of the main moving object corresponding to the third object identifier through a composition model to obtain the motion trajectory of the auxiliary moving object corresponding to the fourth object identifier; The composition model is trained according to the motion trajectory of the auxiliary moving object corresponding to the fourth object identification obtained by the motion trajectory recognition and the motion trajectory of the auxiliary moving object corresponding to the fourth object identification obtained by processing the composition model to obtain the trained composition model.

8. A motion trajectory acquisition device, characterized in that: The device comprises: an acquiring unit, configured to acquire 3D video data and first tag information of the 3D video data, wherein the first tag information includes: a first object identifier of a main moving object, a second object identifier of an auxiliary moving object, and first scene information in the 3D video data; a processing unit, configured to identify a motion trajectory of a main moving object corresponding to the first object identifier to obtain a motion trajectory of the main moving object; The processing unit is further configured to process the first scene information, the second object identifier, and the motion trajectory of the main moving object using the trained composition model to obtain the motion trajectory of the auxiliary moving object corresponding to the second object identifier; The processing unit is further used to place the auxiliary moving object based on the suggestion of the motion trajectory of the auxiliary moving object given by the trained composition model, set the motion mode of the auxiliary object based on the suggestion of the motion trajectory of the auxiliary moving object, and shoot the auxiliary moving object based on the motion mode of the auxiliary object; or adjust the 3D video data based on the suggestion of the motion trajectory of the auxiliary moving object, so that the motion trajectory of the auxiliary moving object in the adjusted 3D video data matches the motion trajectory of the auxiliary moving object given based on the trained composition model.

9. An electronic device, characterized in that: The method comprises a processor, a memory and a communication interface, wherein the processor, the memory and the communication interface are interconnected, wherein the memory is used to store computer program instructions, and the processor is configured to execute the program instructions to implement the method for obtaining the motion trajectory according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that The computer-readable storage medium stores computer program instructions, and when the computer program instructions are executed by a processor, they are used to execute the motion trajectory acquisition method according to any one of claims 1 to 7.

Images