Network routing of media streams based upon semantic contents
AssigneeAT&T Intellectual Property I, L.P.
InventorsBehzad Shahraray, Eric Zavesky, Bernard S. Renger, David Crawford Gibbon
Life of this patent
Estimating the term…
Abstract
Methods, computer-readable media, and devices are disclosed for routing media streams to destination devices based upon semantic contents detected in the media streams. For example, a method may include a processing system including at least one processor detecting a first semantic content in a media stream of a media source device in accordance with a machine learning model for detecting the first semantic content, selecting a first destination device for the media stream based upon the first semantic content, and sending the media stream to the first destination device that is selected.
Claims20
A method comprising:
- detecting, by a processing system including at least one processor, a first semantic content in a media stream of a media source device in accordance with a first machine learning model for detecting the first semantic content;
- detecting, by the processing system, a second semantic content in the media stream in accordance with a second machine learning model for detecting the second semantic content, wherein the second machine learning model for detecting the second semantic content is activated in response to the detecting the first semantic content in the media stream, and wherein the second machine learning model for detecting the second semantic content is applied to a look-back time period of the media stream that is prior to a time of the media stream at which the first semantic content is detected, wherein the second semantic content is different from the first semantic content;
- selecting, by the processing system, a first destination device for the media stream based upon the first semantic content and the second semantic content that are detected;
- and
- sending, by the processing system, the media stream to the first destination device that is selected.
A non-transitory computer-readable medium storing instructions which, when executed by a processing system including at least one processor, cause the processing system to perform operations, the operations comprising:
- detecting a first semantic content in a media stream of a media source device in accordance with a first machine learning model for detecting the first semantic content;
- detecting a second semantic content in the media stream in accordance with a second machine learning model for detecting the second semantic content, wherein the second machine learning model for detecting the second semantic content is activated in response to the detecting the first semantic content in the media stream, and wherein the second machine learning model for detecting the second semantic content is applied to a look-back time period of the media stream that is prior to a time of the media stream at which the first semantic content is detected, wherein the second semantic content is different from the first semantic content;
- selecting a first destination device for the media stream based upon the first semantic content and the second semantic content that are detected;
- and
- sending the media stream to the first destination device that is selected.
A device comprising:
- a processing system including at least one processor;
- and
- a computer-readable medium storing instructions which, when executed by the processing system, cause the processing system to perform operations, the operations comprising:
- detecting a first semantic content in a media stream of a media source device in accordance with a first machine learning model for detecting the first semantic content;
- detecting a second semantic content in the media stream in accordance with a second machine learning model for detecting the second semantic content, wherein the second machine learning model for detecting the second semantic content is activated in response to the detecting the first semantic content in the media stream, and wherein the second machine learning model for detecting the second semantic content is applied to a look-back time period of the media stream that is prior to a time of the media stream at which the first semantic content is detected, wherein the second semantic content is different from the first semantic content;
- selecting a first destination device for the media stream based upon the first semantic content and the second semantic content that are detected;
- and
- sending the media stream to the first destination device that is selected.
Dependencies are inferred from the claim text.