Back to results
US 12,156,727 B2Granted
Deep learning-based cough recognition method and device
AssigneesKorea Advanced Institute of Science and Technology, Sm Instruments Co., Ltd.
InventorsYong Hwa Park, Gyeong Tae Lee, Hyeonuk Nam, Seonghu Kim, Young Key Kim, In Kwon Kim, Kwang Hyun Lee, Jae Sun Lee, Sung Hyo Park
Life of this patent
Estimating the term…
Abstract
Provided is a cough recognition method and device, which can detect cough sounds from an audio signal, and not only can detect coughs but also can track the location at which the cough sounds are generated by calculating the location of a sound source.
Claims6
A cough recognition device comprising:
- a plurality of microphone arrays configured to receive audio signals;
- a camera module configured to generate a camera image;
- and
- at least one processor,
- wherein the at least one processor is configured to,
- collect the audio signals through the plurality of microphone arrays and display a location of a sound source on the camera image,
- determine whether the audio signals include cough sound using a deep learning model, and
- when the audio signals include the cough sound, further display a cough indication in association with the location of the sound source displayed on the camera image,
- wherein the deep learning model includes a binary classification model configured to receive and learn learning data including data augmented cough sound with background noise and to output whether the input audio signals include the cough sound,
- wherein the at least one processor is further configured to, when the input audio signals is classified as including the cough sound, displaying a location where the cough sound is generated by displaying the cough label indicating the cough mark along with the location of the sound source on the camera image.
7 more elements
A cough recognition method comprising:
- generating a camera image through a camera module;
- collecting audio signals through a plurality of microphone arrays and displaying a location of a sound source on the camera image;
- determining whether the audio signals include cough sound using a deep learning model;
- and
- when the audio signals are determined to include the cough sound, further displaying a cough indication in association with the location of the sound source displayed on the camera image,
- wherein the deep learning model includes a binary classification model configured to receive and learn learning data including data augmented cough sound with background noise and to output whether the input audio signals include the cough sound,
- wherein, when the input audio signals is classified as including the cough sound, a location where the cough sound is generated is displayed by displaying the cough label indicating the cough mark along with the location of the sound source on the camera image.
4 more elements
Dependencies are inferred from the claim text.