# US 2022/0093089 A1 · Model constructing method for audio recognition

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2022/0093089 A1
- Document ID: US20220093089A1
- Status: Published application
- Kind: A1
- Application number: 17/197,050
- Filed: 2021-03-10
- Published: 2022-03-24
- Earliest claimed: 2020-09-21
- Assignees: Askey Computer Corp.; Askey Technology (Jiangsu) Ltd.
- Applicants: Askey Computer Corp.; Askey Technology (Jiangsu) Ltd.
- Inventors: Chien-Fang Chen; SETYA WIDYAWAN PRAKOSA; Huan-Ruei Shiu; Chien-Ming Lee
- Web page: https://patentgrep.com/patent/US20220093089A1

## CPC

- G10L 15/197: Probabilistic grammars, e.g. word n-grams
- G10L 25/54: for retrieval
- G06N 3/044: Recurrent networks, e.g. Hopfield networks
- G06N 3/0442: characterised by memory or gating, e.g. long short-term memory [LSTM] or gated recurrent units [GRU]
- G06N 3/0499: Feedforward networks
- G06N 3/08: Learning methods
- G06N 3/09: Supervised learning
- G06N 3/091: Active learning
- G10L 15/02: Feature extraction for speech recognition; Selection of recognition unit
- G10L 15/05: Word boundary detection
- G10L 15/063: {Training}
- G10L 15/22: Procedures used during a speech recognition process, e.g. man-machine dialogue
- G10L 15/30: Distributed recognition, e.g. in client-server systems, for mobile phones or network applications
- G10L 21/0232: Processing in the frequency domain
- G10L 25/51: for comparison or discrimination
- G10L 21/0216: characterised by the method used for estimating noise
- G10L 25/09: the extracted parameters being zero crossing rates
- G10L 25/21: the extracted parameters being power information
- G10L 25/30: using neural networks
- G10L 25/81: for discriminating voice from music
- G10L 25/84: for discriminating voice from noise
- G10L 2015/0635: {updating or merging of old and new templates; Mean values; Weighting}
- G10L 2025/783: {based on threshold decision}
- G10L: SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
- G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS

## Abstract

A model constructing method for audio recognition is provided. In the method, audio data is obtained. A predicted result of the audio data is determined by using the classification model which is trained by machine learning algorithm. The predicted result includes a label defined by the classification model. A prompt message is provided according to a loss level of the predicted result. The loss level is related to a difference between the predicted result and a corresponding actual result. The prompt message is used to query a correlation between the audio data and the label. The classification model is modified according to a confirmation response of the prompt message, and the confirmation response is related to a confirmation of the correlation between the audio data and the label. Accordingly, the labeling efficiency and predicting correctness can be improved.

## Claims

### Claim 1

A model construction method for audio recognition, comprising:
obtaining an audio data;
determining a predicted result of the audio data by using a classification model, wherein the classification model is trained based on a machine learning algorithm, and the predicted result comprises a label defined by the classification model;
providing a prompt message according to a loss level of the predicted result, wherein the loss level is related to a difference between the predicted result and a corresponding actual result, and the prompt message is provided to query a correlation between the audio data and the label; and
modifying the classification model according to a confirmation response of the prompt message, wherein the confirmation response is related to a confirmation of the correlation between the audio data and the label.

### Claim 2 (depends on 1)

The model construction method for audio recognition according to claim 1, wherein the prompt message comprises the audio data and an inquiry content, the inquiry content is to query whether the audio data belongs to the label, and the steps of providing the prompt message comprises:
playing the audio data and providing the inquiry content.

### Claim 3 (depends on 2)

The model construction method for audio recognition according to claim 2, wherein the step of modifying the classification model according to the confirmation response of the prompt message comprises:
receiving an input operation, wherein the input operation corresponds to an option of the inquiry content, and the option is that the audio data belongs to the label or the audio data does not belong to the label; and
determining the confirmation response based on the input operation.

### Claim 4 (depends on 1)

The model construction method for audio recognition according to claim 1, wherein the step of modifying the classification model according to the confirmation response of the prompt message comprises:
adopting a label and the audio data corresponding to the confirmation response as training data of the classification model, and the classification model is retrained accordingly.

### Claim 5 (depends on 1)

The model construction method for audio recognition according to claim 1, wherein the step of obtaining the audio data comprises:
analyzing properties of an original audio data to determine a noise component of the original audio data; and
eliminating the noise component from the original audio data to generate the audio data.

### Claim 6 (depends on 5)

The model construction method for audio recognition according to claim 5, wherein the properties comprise a plurality of intrinsic mode functions (IMF), and the step of determining the noise component of the audio data comprises:
decomposing the original audio data to generate a plurality of mode components of the original audio data, wherein each of the mode components corresponds to an intrinsic mode function;
determining an autocorrelation of each of the mode components; and
selecting one of the mode components as the noise component according to the autocorrelation of the mode components.

### Claim 7 (depends on 1)

The model construction method for audio recognition according to claim 1, wherein the step of obtaining the audio data comprises:
extracting a sound feature from the audio data;
determining a target segment and a non-target segment in the audio data according to the sound feature; and
retaining the target segment, and removing the non-target segment.

### Claim 8 (depends on 5)

The model construction method for audio recognition according to claim 5, wherein the step of obtaining the audio data comprises:
extracting a sound feature from the audio data;
determining a target segment and a non-target segment in the audio data according to the sound feature; and
retaining the target segment, and removing the non-target segment.

### Claim 9 (depends on 7)

The model construction method for audio recognition according to claim 7, wherein the target segment is a voice content, the non-target segment is not the voice content, the voice features comprises a short time energy and a zero crossing rate, and the step of extracting the sound feature from the audio data comprises:
determining two end points of the target segment in the audio data according to the short time energy and the zero crossing rate of the audio data, wherein the two end points are related to a boundary of the target segment in a time domain.

### Claim 10 (depends on 7)

The model construction method for audio recognition according to claim 7, further comprising:
providing a second prompt message according to the target segment, wherein the second prompt message is provided to request the label be assigned to the target segment; and
training the classification model according to a second confirmation response of the second prompt message, wherein the second confirmation response comprises the label corresponding to the target segment.

### Claim 11 (depends on 1)

The model construction method for audio recognition according to claim 1, further comprising:
providing the classification model that is transmitted through a network;
loading the classification model obtained through the network to recognize a voice input; and
providing an event notification based on a recognition result of the voice input.
