Noise suppression controls
AssigneeHewlett-Packard Development Company, L.P.
InventorsAlexander Wayne Clark, Yun David Tang, Maureen Ann Aragon
Life of this patent
Estimating the term…
Abstract
Examples of noise suppression controls are described herein. In some examples, an electronic device includes a processor to classify, using a first machine learning model, an environment based on video of the environment to produce a classification. In some examples, the processor is to detect, using a second machine learning model, a situation in the environment based on the video to produce a detection. In some examples, the processor is to control noise suppression on audio captured from the environment based on the classification and the detection.
Claims19
An electronic device, comprising:
- a processor to:
- classify, using a first machine learning model that is to assign a probability to each of a set of environments, the environment with a greatest probability from the set of environments based on video of the environment to produce a classification;
- detect, using a second machine learning model, a situation in the environment based on the video to produce a detection, the situation indicating a first bounding box location and a second bounding box location and includes determining a distance between the first bounding box location and the second bounding box location using a relative distance bounding box diagonal length for each bounding box, each bounding box diagonal length being an average variation of a corresponding bounding box width and height;
- and
- control noise suppression, with a noise suppressor using frequency-based filtering, spectral subtraction, linear time invariant filtering, machine learning based noise suppression, or beamforming, on audio captured from the environment based on the classification and the detection.
An apparatus, comprising:
- a processor to:
- determine whether video from a camera indicates an office using a first machine learning model;
- determine a scenario that indicates a location relationship by establishing a distance between a first bounding box location and a second bounding box location, based on the video in response to determining whether the video indicates an office;
- determine the distance between the first bounding box location and the second bounding box location based on a relative distance bounding box diagonal length for each bounding box;
- determine, with a noise suppressor using frequency-based filtering, spectral subtraction, linear time invariant filtering, machine learning based noise suppression, or beamforming, a level of noise suppression based on the scenario;
- and
- process an audio signal from a microphone at the level of noise suppression.
A non-transitory tangible computer-readable medium comprising instructions when executed cause a processor of an electronic device to:
- select, using a first machine learning model that is to assign probabilities to each environment in a set of environments and an image, the environment with the greatest probability from the set of environments that includes an office environment and a non-office environment;
- infer, using a second machine learning model, a situation in the environment in response to the selection;
- determine a noise suppression level, with a noise suppressor, based on the situation, the situation indicating a first bounding box location and a second bounding box location, wherein a distance between the first bounding box location and the second bounding box location is based on a relative distance bounding box diagonal length for each bounding box, the bounding box diagonal length being an average variation of a corresponding bounding box width and height;
- and
- perform noise suppression on an audio signal captured from the environment with the noise suppressor using frequency-based filtering, spectral subtraction, linear time invariant filtering, machine learning based noise suppression, or beamforming based on the noise suppression level.
Dependencies are inferred from the claim text.