Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for implementing a machine-learned object tracking policy. One of the methods includes receiving a current video frame by a user device having a plurality of installed object trackers, wherein each object tracker is configured to perform a different object tracking procedure on the current video frame rent video frame. The current video frame and one or more object tracks previously generated by the one or more object trackers are provided as input to a trained policy engine that implements a reinforcement learning model to generate a particular object tracking plan. A particular object tracking plan is selected based on the output of the reinforcement learning model, and the selected object tracking plan is performed on the current video frame to generate one or more updated object tracks for the current video frame.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for implementing a machine-learned object tracking policy. One of the methods includes receiving a current video frame by a user device having a plurality of installed object trackers, wherein each object tracker is configured to perform a different object tracking procedure on the current video frame rent video frame. The current video frame and one or more object tracks previously generated by the one or more object trackers are provided as input to a trained policy engine that implements a reinforcement learning model to generate a particular object tracking plan. A particular object tracking plan is selected based on the output of the reinforcement learning model, and the selected object tracking plan is performed on the current video frame to generate one or more updated object tracks for the current video frame.

Abstract: A method for personalizing a content item using captured footage is disclosed. The method includes receiving a first video feed from a first camera, wherein the first camera is designated as a source camera for capturing an event during a first time duration. The method also includes receiving data from a second camera, and determining, based on the received data from the second camera, that an action was performed using the second camera, the action being indicative of a region of interest (ROI) of the user of the second camera occurring within a second time duration. The method further includes designating the second camera as the source camera for capturing the event during the second time duration.

Abstract: A system and method are disclosed for motion-based feature correspondence. A method may include detecting a first motion of a first feature across two or more first frames of a first video clip captured by a first video camera and a second motion of a second feature across two or more second frames of a second video clip captured by a second video camera. The method may further include determining, based on the first motion in the first video clip and the second motion in the second video clip, that the first feature and the second feature correspond to a common entity, the first motion in the first video clip and the second motion in the second video clip corresponding to one or more common points in time in the first video clip and the second video clip.

Abstract: A plurality of videos is analyzed (in real time or after the videos are generated) to identify interesting portions of the videos. The interesting portions are identified based on one or more of the people depicted in the videos, the objects depicted in the videos, the motion of objects and/or people in the videos, and the locations where people depicted in the videos are looking. The interesting portions are combined to generate a content item.

Abstract: A method and system include identifying a plurality of media clips pertaining to an event, for each pair of the plurality of media clips, calculating pairwise relative time offsets for pairs of media clips from the plurality of media clips, calculating, by a processing device, an inconsistency measure, in which the inconsistency measure of a pair of media clips is calculated using at least three relative time offsets and one of the at least three relative time offsets is a relative time offset of the pair of media clips, and synchronizing the plurality of media clips based on resulting calculated inconsistency measures.

Abstract: A method for localizing the attention of a user of a first-person point-of-view (FPPOV) device is disclosed. The method includes receiving data from an FPPOV device, the data being indicative of a first region-of-interest (ROI) of an event for a first time duration and a second ROI of the event for a second time duration. The method further include determining that a first camera from a plurality of cameras best captures the first ROI during the first time duration, and determining that a second camera from the plurality of cameras best captures the second ROI during the second time duration.

Abstract: A system and method are disclosed for identifying feature correspondences among a plurality of video clips of a dynamic scene. In one implementation, a computer system identifies a first feature in a first video clip of a dynamic scene that is captured by a first video camera, and a second feature in a second video clip of the dynamic scene that is captured by a second video camera. The computer system determines, based on motion in the first video clip and motion in the second video clip, that the first feature and the second feature do not correspond to a common entity.

Abstract: A method for localizing the attention of a user of a first-person point-of-view (FPPOV) device is disclosed. The method includes receiving a plurality of images of an event, each image having been captured by one of a plurality of reference cameras during a first time duration. The method further includes receiving a first user-captured image captured by the FPPOV device during the first time duration. A first image of the plurality of images is selected as a best-matched image, based on the first user-captured image, for capturing a region-of-interest.