Automatic selection of representative media clips
AssigneeGoogle Inc.
InventorVictor Bennett
Life of this patent
Estimating the term…
Abstract
A system determines human recognizable labels for portions of an electronic media stream, gathers data associated with the electronic media stream from a number of media players, and determines at least one section of the electronic media stream with a particular media feature. The system selects a representative clip for the electronic media stream based on information regarding the labeled portions, the gathered data, and the at least one section.
Claims20
A method, comprising:
- determining labels associated with portions of an audio stream;
- determining at least one audio feature associated with at least one section of the audio stream, wherein a respective one of the portions of the audio stream overlaps in time with a respective one of the at least one section of the audio stream;
- generating scores for the determined labels associated with the portions and at least one score for the at least one audio feature associated with the at least one section of the audio stream;
- and
- identifying, by a computing device, a representative clip for the audio stream in view of a combination of the scores for determined labels associated with the portions and the at least one score for the at least one section.
A system, comprising:
- a memory;
- and
- a processing device, coupled to the memory, to:
- determine labels associated with portions of an audio stream;
- determine at least one audio feature associated with at least one section of the audio stream, wherein a respective one of the portions of the audio stream overlaps in time with a respective one of the at least one section of the audio stream;
- generate scores for the determined labels associated with the portions and at least one score for the at least one audio feature associated with the at least one section of the audio stream;
- and
- identify a representative clip for the audio stream in view of a combination of the scores for determined labels associated with the portions and the at least one score for the at least one section.
A non-transitory computer-readable medium to store instructions that, when executed by a processing device, cause the processing device to perform operations comprising:
- determining labels associated with portions of an audio stream;
- determining at least one audio feature associated with at least one section of the audio stream, wherein a respective one of the portions of the audio stream overlaps in time with a respective one of the at least one section of the audio stream;
- generating scores for the determined labels associated with the portions and at least one score for the at least one audio feature associated with the at least one section of the audio stream;
- and
- identifying, by the processing device, a representative clip for the audio stream in view of a combination of the scores for determined labels associated with the portions and the at least one score for the at least one section.
Dependencies are inferred from the claim text.