Abstract: A media server generates a pattern using an irregular polygon. A generative model is applied to the irregular polygon to determine padding values at seams between contacting polygons. The media server may perform inpainting using the generative model to inpaint an area of a virtual object using a pattern generated using the irregular polygon. In one example, the media server may use an einstein, an aperiodic monotile, to generate a pattern (e.g., a pattern to fill in petals of a virtual lotus, a three-dimensional forest, etc.) for presentation at a client device (e.g., an augmented reality headset). The media server may generate different patterns based on how a user is interacting with a media application (e.g., the media server may generate a different pattern to target a different breathing rate for a user interacting with a meditation exercise at the client device).

Abstract: A virtual reality application adaptively generates a virtual reality experience intended to improve a user's mood. A plurality of digital assets is stored together with associated scores. A score for a digital asset represents a predicted amount of mood improvement occurring in a cohort of users having similar profiles to a target user in response to virtual reality experiences including the digital asset. A customized virtual reality experience is generated for the target user based at least in part of the user profile data on the scores. The user's change in mood is detected through survey and/or biometric data. Scores for the digital assets may then be updated in order to further learn their respective effects on the cohort of users.

Abstract: A virtual reality application adaptively generates a virtual reality experience intended to improve a user's mood. A plurality of digital assets is stored together with associated scores. A score for a digital asset represents a predicted amount of mood improvement occurring in a cohort of users having similar profiles to a target user in response to virtual reality experiences including the digital asset. A customized virtual reality experience is generated for the target user based at least in part of the user profile data on the scores. The user's change in mood is detected through survey and/or biometric data. Scores for the digital assets may then be updated in order to further learn their respective effects on the cohort of users.

Abstract: A media system controls presentation of media content based on a detected state of a user. The user state may be detected based on biometric inputs that may be derived from IMU and/or microphone data. The biometric inputs may include a heart rate detected from IMU data representing motion of a head-mounted media processing device, a breathing rate detected from IMU data representing motion of a head-mounted media processing device, a breathing rate detected based on microphone data, or a combination thereof.