An apparatus for diagnosing a disease and a method for diagnosing a disease, in which: a plurality of voice signals are received to generate a first image signal and a second image signal which are image signals for each voice signal; a plurality of disease probability information for a target disease causing a voice change are extracted by using an artificial intelligence model determined according to the type of each voice signal and a generation method used to generate each image signal for the first image signal and the second image signal for each voice signal; and it is determined whether the target disease is negative or positive on the basis of the plurality of disease probability information.

A method for counting coughs is provided. The method includes acquiring a plurality of onset signals from the sound signal, wherein the onset signal has a predetermined time length; acquiring a plurality of spectrograms corresponding to each of the plurality of onset signals; determining whether each of the acquired plurality of spectrograms represents a cough using a cough determination model; and calculating a number of coughs included in the sound signal based on a time point of a cough signal. The cough signal is an onset signal corresponding to one spectrogram determined to represent the cough. When a time interval between a first time point of a first cough signal and a second time point of a second cough is within a reference time interval, the first cough signal and the second cough signal are regarded as one cough signal at the first time point.

A health condition or change in health condition of a person may be determined by processing the person's speech with a neural network. Speech from more than one time period may be processed and, in some implementations, speech from a time period may be associated with a health condition label. For each time period, a feature vector may be computed from the speech and the feature vector may be processed with a neural network to obtain a speech embedding vector. In some implementations, feature vector may include word-piece encodings and the neural network may be a transformer neural network. The speech embedding vectors may be processed with a mathematical model to determine a change in a health condition between two time periods or to determine a health condition label for a specific time period. In some implementations, the mathematical model may be a regression model or a fully-connected neural network.

A health condition or change in health condition of a person may be determined by processing the person's speech with a neural network. Speech from more than one time period may be processed and, in some implementations, speech from a time period may be associated with a health condition label. For each time period, a feature vector may be computed from the speech and the feature vector may be processed with a neural network to obtain a speech embedding vector. In some implementations, feature vector may include word-piece encodings and the neural network may be a transformer neural network. The speech embedding vectors may be processed with a mathematical model to determine a change in a health condition between two time periods or to determine a health condition label for a specific time period. In some implementations, the mathematical model may be a regression model or a fully-connected neural network.

A method and electronic device for processing a waveform are disclosed. The waveform is representative of bodily sounds. The method includes acquiring the waveform from the sound recording component and having a low-frequency component and a high-frequency component, selecting a target moving averaging filter amongst a first moving averaging filter and a second moving averaging filter for filtering the waveform. The first moving averaging filter is to be used for preserving the low-frequency component of the waveform, and the second moving averaging filter is to be used for preserving the high-frequency component of the waveform. The method includes applying the target moving averaging filter on the waveform for reducing noise in the waveform, thereby generating a second waveform.

A method and electronic device for processing a waveform are disclosed. The waveform is representative of bodily sounds. The method includes acquiring the waveform from the sound recording component and having a low-frequency component and a high-frequency component, selecting a target moving averaging filter amongst a first moving averaging filter and a second moving averaging filter for filtering the waveform. The first moving averaging filter is to be used for preserving the low-frequency component of the waveform, and the second moving averaging filter is to be used for preserving the high-frequency component of the waveform. The method includes applying the target moving averaging filter on the waveform for reducing noise in the waveform, thereby generating a second waveform.

Aspects relate to systems and methods for determining actor status according to behavioral phenomena. An exemplary system includes an eye sensor configured to detect an eye parameter as a function of an eye phenomenon, a speech sensor configured to detect a speech parameter as a function of a speech phenomenon, and a processor in communication with the eye sensor and the speech sensor; the processor is configured to receive the eye parameter and the speech parameter, determine an eye pattern as a function of the eye parameter, determine a speech pattern as a function of the speech parameter, and correlate one or more of the eye pattern and the speech pattern to a cognitive status.

A notification system includes: detection means (110) for detecting an acoustic event from voice data transmitted from a communication terminal held by a target person; and notification means (120) for sending a predetermined notification when the detection means (110) has detected the acoustic event. Accordingly, it is possible to determine the state of a target person regardless of the state of this person. Further, when the difference between an acoustic pattern of the voice data transmitted from the communication terminal and acoustic patterns registered in advance is outside a predetermined range, a management server (101) does not send a notification, whereby it is possible to prevent communication traffic from being increased based on unnecessary notifications.

A notification system includes: detection means (110) for detecting an acoustic event from voice data transmitted from a communication terminal held by a target person; and notification means (120) for sending a predetermined notification when the detection means (110) has detected the acoustic event. Accordingly, it is possible to determine the state of a target person regardless of the state of this person. Further, when the difference between an acoustic pattern of the voice data transmitted from the communication terminal and acoustic patterns registered in advance is outside a predetermined range, a management server (101) does not send a notification, whereby it is possible to prevent communication traffic from being increased based on unnecessary notifications.

In one example in accordance with the present disclosure, an electronic device is described. The example electronic device includes a microphone device to record a tracheal sound. The example electronic device also includes an oral expiratory flow sensor to measure oral expiratory flow contents. The example electronic device further includes a processor to determine a health condition based on the tracheal sound and the oral expiratory flow contents.