Processing of data of a video sequence in order to zoom to a speaker detected in the sequence
US11076224B2 · kind B2 · utility
Assignee
Inventors
Key dates
| Filing date | Dec 4, 2018 |
| Grant date | Jul 27, 2021 |
| Priority date | — |
| Expiry date | Dec 4, 2038 |
Classification
- Technology area (CPC G)Physics
- CPC primaryG10L2021/02166
- WIPO fieldAudio-visual technology
- WIPO sectorElectrical engineering
Abstract
Method and device for processing a video sequence containing a succession of images of one or more participant speakers, captured by a wide-angle camera. The method includes: capturing sound using a microphone having a plurality of sensors for capturing a sound field; processing the audio data captured by the microphone in order to determine at least one direction of origin of sound coming from a participant, relative to an optical axis of the wide-angle camera; generating a signal including data concerning the direction of origin of the sound relative to the optical axis of the camera, for the purpose of utilizing the signal when rendering the captured images by zooming into an area around the participant emitting the sound for which the direction of origin corresponds to the data of the signal.
Source: USPTO / EPO open patent data. Objective bibliographic and citation counts.