Meta Platforms, Inc.
Accounting for locations of a gaze of a user within content to select content for presentation to the user

Abstract:

An online system provides video data to client devices for presentation to users. Various client devices include a gaze tracking system that identifies locations within the video data where users' gazes were directed at different times during presentation of the video data and provides the identified locations to the online system. Based on characteristics of the users and of the video data, the online system generates a model determining likelihoods of a user's gaze being directed to different locations within the video data at different times. When the online system selects the video data for presentation to a target user, the online system applies the model to the target user and the video data. Based on the resulting likelihoods, the online system modifies resolutions of different locations of the video data when providing the video data to a client device for presentation to the target user.