# US 2020/0404173 A1 · Video processing method and apparatus, terminal device, server, and storage medium > Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt - Number: US 2020/0404173 A1 - Document ID: US20200404173A1 - Status: Published application - Kind: A1 - Application number: 17/010,549 - Filed: 2020-09-02 - Published: 2020-12-24 - Earliest claimed: 2018-07-23 - Assignees: Tencent Technology (Shenzhen) Company Limited - Applicants: Tencent Technology (Shenzhen) Company Limited - Inventors: Xue LI; Wei XIONG - Web page: https://patentgrep.com/patent/US20200404173A1 ## CPC - G06V 20/41: {Higher-level, semantic clustering, classification or understanding of video scenes, e.g. detection, labelling or Markovian modelling of sport events or news items (segmenting video sequences G06V20/49)} - G06F 18/214: Generating training patterns; Bootstrap methods, e.g. bagging or boosting - G06V 20/46: {Extracting features or characteristics from the video content, e.g. video fingerprints, representative shots or key frames} - H04N 23/80: Camera processing pipelines; Components thereof - G06V: IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING - G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N) - H04N: PICTORIAL COMMUNICATION, e.g. TELEVISION ## Abstract Aspects of the disclosure can provide a method, a corresponding target video processing package may be automatically obtained through matching according to scene description information, so that an operation of manually making a selection by a user in the related art is omitted. Therefore, matching efficiency of video processing manners may be improved. In addition, a plurality of video processing manners may be obtained through matching at a time, so that the matching efficiency is further improved. In addition, because the target video processing package is dynamically and correspondingly obtained through matching based on the scene description information, the target video processing package obtained through matching can be caused to accord with actual video content as much as possible. Further, accuracy of video processing may be improved, to satisfy an actual requirement of a user as much as possible. ## Claims ### Claim 1 A video processing method, comprising: obtaining, by a terminal device, scene description information of a target shooting scene; obtaining, by the terminal device, through matching based on the scene description information a target video processing package that corresponds to the target shooting scene and includes one video processing manner for processing a video in a predetermined processing mode; and processing, by the terminal device, a target video obtained by shooting the target shooting scene based on the target video processing package. ### Claim 2 (depends on 1) The method according to claim 1, wherein the obtaining through matching based the scene description information further comprises: performing word vector representation on the scene description information, to obtain a video content feature variable; inputting the video content feature variable into a pre-established package recommendation model for performing package matching, to obtain at least one recommended video processing package matching the scene description information; and determining the target video processing package based on the at least one recommended video processing package. ### Claim 3 (depends on 2) The method according to claim 2, the determining the target video processing package according to the at least one recommended video processing package further comprises: determining a video processing package with the highest usage frequency in the at least one recommended video processing package as the target video processing package; or determining a video processing package having a maximum similarity with a priority video processing package in the at least one recommended video processing package as the target video processing package, the priority video processing package being a video processing package that is obtained through matching according to user attribute information. ### Claim 4 (depends on 2) The method according to claim 2, wherein the determining the target video processing package according to the at least one recommended video processing package further comprises: determining whether there is a recommended video processing package having a similarity with a priority video processing package greater than or equal to a predetermined similarity in the at least one recommended video processing package, the priority video processing package being a video processing package that is obtained through matching according to user attribute information; determining a recommended video processing package having a maximum similarity as the target video processing package when there is a recommended video processing package having a similarity with the priority video processing package greater than or equal to the predetermined similarity in the at least one recommended video processing package; and determining the priority video processing package as the target video processing package when there is no recommended video processing package having a similarity with the priority video processing package greater than or equal to the predetermined similarity in the at least one recommended video processing package. ### Claim 5 (depends on 1) The method according to claim 1, wherein the obtaining through matching based on the scene description information the target video processing package further comprises: inputting the scene description information into a preset set of correspondences between scenes and video processing packages for performing a matching search, to obtain a video processing package having the highest matching degree with the scene description information; and determining the video processing package having the highest matching degree with the scene description information as the target video processing package. ### Claim 6 (depends on 1) The method according to claim 1, wherein the obtaining through matching based the scene description information the target video processing package further comprises: determining separately target video processing packages before and after the shot object changes when the scene description information indicates that a shot object in the target shooting scene changes; and the processing based on the target video processing package, a target video obtained by shooting the target shooting scene further comprises: processing separately a video obtained through shooting by using the corresponding target video processing packages before and after the shot object changes. ### Claim 7 A video processing method, comprising: receiving, by a modeling server, scene description information of a target shooting scene transmitted by a terminal device; performing, by the modeling server, word vector representation on the scene description information to obtain a video content feature variable; inputting, by the modeling server, the video content feature variable into a pre-established package recommendation model for performing package matching to obtain a target video processing package matching the target shooting scene, the target video processing package including a video processing manner for processing a video in a predetermined processing mode; and transmitting, by the modeling server, the target video processing package to the terminal device or a cloud server, to enable the terminal device or the cloud server to process a target video obtained by shooting the target shooting scene based on the target video processing package. ### Claim 8 (depends on 7) The method according to claim 7, wherein the inputting the video content feature variable into a pre-established package recommendation model for performing package matching further comprises: analyzing the video content feature variable to determine at least one of an environment and categories, and a quantity of objects corresponding to the target shooting scene; obtaining through matching one video processing manner for the determined at least one of environment and different categories of objects separately; and using a video processing package including a plurality of video processing manners obtained through matching as the target video processing package. ### Claim 9 (depends on 7) The method according to claim 7, wherein the method further comprises: obtaining at least one of user attribute information and historical viewing information of a user corresponding to the terminal device; performing word vector representation on the at least one of the user attribute information and the historical viewing information separately to obtain an auxiliary feature variable; and the inputting the video content feature variable into a pre-established package recommendation model for performing package matching further comprises: inputting the video content feature variable and the auxiliary feature variable together into the pre-established package recommendation model for performing package matching to obtain the target video processing package. ### Claim 10 (depends on 7) The method according to claim 7, wherein the package recommendation model is established in the following manner: selecting a plurality of videos from videos that have been published as video training samples; marking shooting environments and/or shot objects of the video training samples based on image recognition results of video sequence frames included in the video training samples to obtain video content tags of the video training samples; extracting video processing packages used by the video training samples; and inputting the video content tags of the video training samples and the corresponding video processing packages as training features into a preset network model for performing training and learning to obtain the package recommendation model. ### Claim 11 (depends on 10) The method according to claim 10, wherein the inputting the video content tags further comprising: determining recommendation point values of the video training samples according to historical interaction data of the video training samples, the historical interaction data of the video training samples being used for indicating interaction statuses between users and the video training samples; and performing training and learning after establishing associations between the recommendation point values of the video training samples and the corresponding video processing packages according to a predetermined association rule to obtain the package recommendation model. ### Claim 12 A terminal device comprising processing circuitry configure to: obtain scene description information of a target shooting scene; obtain through matching, based on the scene description information, a target video processing package corresponding to the target shooting scene, the target video processing package including at least one video processing manner for processing a video in a predetermined processing mode; and process a target video obtained by shooting the target shooting scene based on the target video processing package. ### Claim 13 (depends on 12) The terminal device according to claim 12, wherein the processing circuitry is further configured to: perform word vector representation on the scene description information to obtain a video content feature variable; input the video content feature variable into a pre-established package recommendation model for performing package matching to obtain at least one recommended video processing package matching the scene description information; and determine the target video processing package according to the at least one recommended video processing package. ### Claim 14 (depends on 13) The terminal device according to claim 13, wherein the processing circuitry is further configured to: determine a video processing package with a highest usage frequency in the at least one recommended video processing package as the target video processing package; or determine a video processing package having a maximum similarity with a priority video processing package in the at least one recommended video processing package as the target video processing package, the priority video processing package being a video processing package that is obtained through matching based on user attribute information. ### Claim 15 (depends on 13) The terminal device according to claim 13, wherein the processing circuitry is further configured to: determine whether there is a recommended video processing package having a similarity with a priority video processing package greater than or equal to a predetermined similarity in the at least one recommended video processing package, the priority video processing package being a video processing package that is obtained through matching according to user attribute information; determine a recommended video processing package having a maximum similarity as the target video processing package when there is a recommended video processing package having a similarity with the priority video processing package greater than or equal to the predetermined similarity in the at least one recommended video processing package; and determine the priority video processing package as the target video processing package when there is no recommended video processing package having a similarity with the priority video processing package greater than or equal to the predetermined similarity in the at least one recommended video processing package. ### Claim 16 (depends on 12) The terminal device according to claim 12, wherein the processing circuitry is further configured to: input the scene description information into a preset set of correspondences between scenes and video processing packages for performing a matching search to obtain a video processing package having the highest matching degree with the scene description information; and determine the video processing package having the highest matching degree with the scene description information as the target video processing package. ### Claim 17 (depends on 13) The terminal device according to claim 13, wherein the processing circuitry is further configured to: determine separately target video processing packages before and after the shot object changes when the scene description information indicates that a shot object in the target shooting scene changes; and process separately a video obtained through shooting by using the corresponding target video processing packages before and after the shot object changes. ### Claim 18 (depends on 12) The terminal device according to claim 12, wherein the processing circuitry is further configured to: obtain a preview video of the target shooting scene, or obtain the target video actually shot in the target shooting scene; perform image recognition on a video sequence frame of the preview video or a video sequence frame of the target video to obtain key feature information of each frame, the key feature information being feature information of a shot object occupying a largest area and/or visually presented in the foremost location in each frame; and determine the scene description information according to key feature information of all frames. ### Claim 19 (depends on 7) A server comprising a processor for causing the server to be configured to perform the method of claim 7. ### Claim 20 (depends on 19) The server according to claim 19, wherein the server is further configured to: analyze the video content feature variable to determine an environment and/or categories and a quantity of objects corresponding to the target shooting scene; obtain through matching one video processing manner for the determined environment and/or different categories of objects separately; and use a video processing package comprising a plurality of video processing manners obtained through matching as the target video processing package. ### Claim 21 (depends on 19) The server according to claim 19, wherein the server is further configured to: obtain at least on of user attribute information and historical viewing information of a user corresponding to the terminal device; perform word vector representation on the at least one of the user attribute information and the historical viewing information separately to obtain an auxiliary feature variable; and input the video content feature variable and the auxiliary feature variable together into the package recommendation model for performing package matching to obtain the target video processing package. ### Claim 22 (depends on 19) The server according to claim 19, wherein the server is further configured to: select a plurality of videos from videos that have been published as video training samples; mark at least one of shooting environments and shot objects of the video training samples based on image recognition results of video sequence frames included in the video training samples to obtain video content tags of the video training samples; extract video processing packages used by the video training samples; and input the video content tags of the video training samples and the corresponding video processing packages as training features into a preset network model for performing training and learning to obtain the package recommendation model. ### Claim 23 (depends on 22) The server according to claim 22, wherein the server is configured to: determine recommendation point values of the video training samples according to historical interaction data of the video training samples, the historical interaction data of the video training samples being used for indicating interaction statuses between users and the video training samples; and perform training and learning after establishing associations between the recommendation point values of the video training samples and the corresponding video processing packages according to a predetermined association rule, to obtain the package recommendation model. ### Claim 24 (depends on 23) The server according to claim 23, wherein the server is further configured to: perform association training on the video content tags and the corresponding video processing packages in the preset network model according to a principle that a training weight of a video processing package corresponding to a larger recommendation point value is larger to obtain the package recommendation model; or determine target video content tags whose recommendation point values are greater than or equal to a predetermined point value, and then perform association training on the target video content tags and corresponding video processing packages in the preset network model according to a principle that a training weight of a video processing package corresponding to a larger recommendation point value is larger to obtain the package recommendation model. ### Claim 25 (depends on 1) A video processing apparatus, comprising: a memory that is configured to store program instructions; a processor that is configured to invoke the program instructions stored in the memory, to perform, according to the obtained program instructions, operations comprised in the method according to claim 1. ### Claim 26 (depends on 1) A non-transitory storage medium, storing computer-executable instructions, the computer-executable instructions being used for causing a computer to perform operations comprised in the method according to claim 1.