Systems and methods for processing content
AssigneeMeta Platforms, Inc.
InventorsThomas Frederick Dimson, Ian McIntyre Silber, Ryan Keenan Olson
Life of this patent
Estimating the term…
Abstract
Systems, methods, and non-transitory computer-readable media can determine a landscape video to be presented in a vertical orientation through a display screen of a computing device. Frames of the landscape video can be cropped to allow for full screen presentation of the landscape video in the vertical orientation. The cropped frames of the landscape video can be provided for presentation through the display screen of the computing device in the vertical orientation.
Claims20
A computer-implemented method comprising:
- aggregating, by a computing system, heat map data for a plurality of content items, wherein the aggregated heat map data is generated based on view tracking data associated with a plurality of users who viewed the plurality of content items, wherein the aggregated heat map data generated based on view tracking data is associated with salient regions of the plurality of content items;
- generating, by the computing system, a plurality of training examples based on the aggregated heat map data, wherein a training example labels a frame of a content item based on the aggregated heat map data;
- and
- training, by the computing system, a saliency prediction model to identify salient regions in content items based on the plurality of training examples.
A system comprising:
- at least one processor;
- and
- a memory storing instructions that, when executed by the at least one processor, cause the system to perform:
- aggregating heat map data for a plurality of content items, wherein the aggregated heat map data is generated based on view tracking data associated with a plurality of users who viewed the plurality of content items, wherein the aggregated heat map data generated based on view tracking data is associated with salient regions of the plurality of content items;
- generating a plurality of training examples based on the aggregated heat map data, wherein a training example labels a frame of a content item based on the aggregated heat map data;
- and
- training a saliency prediction model to identify salient regions in content items based on the plurality of training examples.
A non-transitory computer-readable storage medium including instructions that, when executed by at least one processor of a computing system, cause the computing system to perform a method comprising:
- aggregating heat map data for a plurality of content items, wherein the aggregated heat map data is generated based on view tracking data associated with a plurality of users who viewed the plurality of content items, wherein the aggregated heat map data generated based on view tracking data is associated with salient regions of the plurality of content items;
- generating a plurality of training examples based on the aggregated heat map data, wherein a training example labels a frame of a content item based on the aggregated heat map data;
- and
- training a saliency prediction model to identify salient regions in content items based on the plurality of training examples.
Dependencies are inferred from the claim text.