Modifying capture of video data by an image capture device based on video data previously captured by the image capture device

Abstract

Various client devices include displays and one or more image capture devices configured to capture video data. Different users of an online system may authorize client devices to exchange information captured by their respective image capture devices. Additionally, a client device modifies captured video data based on users identified in the video data. For example, the client device changes parameters of the image capture device to more prominently display a user identified in the video data and may further change parameters of the image capture device based on gestures or movement of the user identified in the video data. The client device may apply multiple models to captured video data to modify the captured video data or subsequent capturing of video data by the image capture device.

What is claimed is: 
     
         1 . A method comprising:
 capturing, by an image capture device of a client device, video data of a local area within a field of view of the image capture device;   locating a user in the captured video data by applying a first machine-learning model to the captured video data;   determining a three-dimensional pose of the user in the captured video data by applying a three-dimensional pose tracking model to the captured video data;   identifying a portion of a body of the user based on the three-dimensional pose of the user;   modifying the captured video data by changing a focal point of the image capture device to focus on the identified portion of the body of the user; and   transmitting the modified video data to an online system.   
     
     
         2 . The method of  claim 1 , further comprising:
 responsive to locating the user in the captured video data, enforcing a privacy setting established by the user.   
     
     
         3 . The method of  claim 1 , wherein determining a three-dimensional pose of the user comprises:
 identifying a plurality of joints of the user in the captured video data.   
     
     
         4 . The method of  claim 1 , wherein identifying a portion of the body of the user comprises:
 identifying a face, a hand, or a torso of the user.   
     
     
         5 . The method of  claim 1 , wherein further comprising:
 locating a plurality of users in the captured video data by applying the first machine-learning model to the captured video data;   determining a three-dimensional pose of each user of the plurality of users by applying the three-dimensional pose tracking model to the captured video data;   identifying a portion of a body of each user of the plurality of users based on the three-dimensional pose of each user; and   modifying the captured video data such that different portions of the video data show different users of the plurality of users.   
     
     
         6 . The method of  claim 5 , wherein modifying the captured video data such that different portions of the video data show different users of the plurality of users comprises:
 modifying the captured video data to display each user of the plurality of users in a grid region of a plurality of grid regions.   
     
     
         7 . The method of  claim 1 , wherein modifying the captured video data comprises:
 modifying a parameter of the image capture device.   
     
     
         8 . The method of  claim 1 , wherein modifying the captured video data comprises:
 applying a zoom or a crop to the captured video data to focus on the identified portion of the body of the user.   
     
     
         9 . The method of  claim 1 , wherein modifying the captured video data comprises:
 determining whether to modify the captured video data to focus on the identified portion by applying a plurality of rules to the captured video data.   
     
     
         10 . The method of  claim 9 , wherein applying the plurality of rules to the captured video data comprises:
 applying the plurality of rules to content in previously captured video data or to movement in previously captured video data.   
     
     
         11 . A non-transitory computer-readable medium storing instructions that, when executed by a processor, cause the processor to:
 capture, by an image capture device of a client device, video data of a local area within a field of view of the image capture device;   locate a user in the captured video data by applying a first machine-learning model to the captured video data;   determine a three-dimensional pose of the user in the captured video data by applying a three-dimensional pose tracking model to the captured video data;   identify a portion of a body of the user based on the three-dimensional pose of the user;   modify the captured video data by changing a focal point of the image capture device to focus on the identified portion of the body of the user; and   transmit the modified video data to an online system.   
     
     
         12 . The computer-readable medium of  claim 11 , further storing instructions that, when executed by a processor, cause the processor to:
 responsive to locating the user in the captured video data, enforce a privacy setting established by the user.   
     
     
         13 . The computer-readable medium of  claim 11 , wherein the instructions for determining a three-dimensional pose of the user comprise instructions that, when executed by a processor, cause the processor to:
 identify a plurality of joints of the user in the captured video data.   
     
     
         14 . The computer-readable medium of  claim 11 , wherein the instructions for identifying a portion of the body of the user comprise instructions that, when executed by a processor, cause the processor to:
 identify a face, a hand, or a torso of the user.   
     
     
         15 . The computer-readable medium of  claim 11 , wherein further storing instructions that, when executed by a processor, cause the processor to:
 locate a plurality of users in the captured video data by applying the first machine-learning model to the captured video data;   determine a three-dimensional pose of each user of the plurality of users by applying the three-dimensional pose tracking model to the captured video data;   identify a portion of a body of each user of the plurality of users based on the three-dimensional pose of each user; and   modify the captured video data such that different portions of the video data show different users of the plurality of users.   
     
     
         16 . The computer-readable medium of  claim 15 , wherein the instructions for modifying the captured video data such that different portions of the video data show different users of the plurality of users comprise instructions that, when executed by a processor, cause the processor to:
 modify the captured video data to display each user of the plurality of users in a grid region of a plurality of grid regions.   
     
     
         17 . The computer-readable medium of  claim 11 , wherein the instructions for modifying the captured video data comprise instructions that, when executed by a processor, cause the processor to:
 modify a parameter of the image capture device.   
     
     
         18 . The computer-readable medium of  claim 11 , wherein the instructions for modifying the captured video data comprise instructions that, when executed by a processor, cause the processor to:
 apply a zoom or a crop to the captured video data to focus on the identified portion of the body of the user.   
     
     
         19 . The computer-readable medium of  claim 11 , wherein the instructions for modifying the captured video data comprise instructions that, when executed by a processor, cause the processor to:
 determining whether to modify the captured video data to focus on the identified portion by applying a plurality of rules to the captured video data.   
     
     
         20 . The computer-readable medium of  claim 19 , wherein the instructions for applying the plurality of rules to the captured video data comprise instructions that, when executed by a processor, cause the processor to:
 apply the plurality of rules to content in previously captured video data or to movement in previously captured video data.