Performing utterance detection using convolution

    公开(公告)号:US11769491B1

    公开(公告)日:2023-09-26

    申请号:US17036091

    申请日:2020-09-29

    CPC classification number: G10L15/16 G06N3/048 G06N3/08 G10L15/02 G10L2015/088

    Abstract: A system configured to perform utterance detection using data processing techniques that are similar to those used for object detection is provided. For example, the system may treat utterances within audio data as analogous to an object represented within an image and employ techniques to separate and identify individual utterances. The system may include one or more trained models that are trained to perform utterance detection. For example, the system may include a first module to process input audio data and identify whether speech is represented in the input audio data, a second module to apply convolution filters, and a third module configured to determine a boundary identifying a beginning and ending of a portion of the input audio data along with an utterance score indicating how closely the portion of the input audio data represents an utterance.

    VOLUME AND INTENSITY-BASED ACTIVITY EVALUATIONS FOR DEVICES

    公开(公告)号:US20210386328A1

    公开(公告)日:2021-12-16

    申请号:US16899464

    申请日:2020-06-11

    Abstract: Devices, systems, and methods are provided for performing volume and intensity-based activity evaluations. A method may include determining, by a device, a heart rate. The method may include determining, based on the heart rate, a motion threshold amount for a time period. The method may include determining motion data. The method may include comparing the motion data to the motion threshold for the time period. The method may include determining, based on the comparison of the motion data to the motion threshold for the time period, an activity intensity level associated with the heart rate and the motion data. The method may include determining, based on the activity intensity level, an activity score. The method may include sending data indicating the activity score for presentation at a second device.

    Non-intrusive sleep efficiency analysis and recommendations using heart data

    公开(公告)号:US12121362B1

    公开(公告)日:2024-10-22

    申请号:US17018796

    申请日:2020-09-11

    Abstract: Devices, systems, and methods are provided for non-intrusive sleep efficiency analysis and recommendations using heart data. A method may include receiving, by a device, photoplethysmography (PPG) data associated with a heart signal. The method may include determining a first portion of the PPG data that fails to exceed a threshold frequency and a second portion of the PPG data that exceeds the threshold frequency. The method may include determining a ratio of the first portion to the second portion. The method may include determining, based on the PPG data, an inspiratory flow, and determining, based on the ratio and the inspiratory flow, a forecasted sleep efficiency score. The method may include determining a recommended activity associated with increasing the forecasted sleep efficiency score. The method may include presenting the forecasted sleep efficiency score and the recommended activity.

    Sleep scores
    7.
    发明授权

    公开(公告)号:US11771367B1

    公开(公告)日:2023-10-03

    申请号:US16677526

    申请日:2019-11-07

    Abstract: Described are systems and methods for generating sleep scores based at least in part on the visual appearance of hypnograms. The disclosed implementations generate an embedding vector representative of the visual appearance of a hypnogram and utilize that embedding to determine a sleep score of the sleep session represented by the hypnogram. For example, a deep neural network, such as a convolutional neural network (“CNN”), may be trained with input data that includes hypnograms that have been labeled according to their sleep quality, or sleep score. The trained network may then be provided with an unlabeled hypnogram, generate an embedding representative of that hypnogram, and determine a sleep score corresponding to that hypnogram.

Patent Agency Ranking