Silent command earphone audio system based on bone voiceprint microphone
By collecting bone conduction signals through the bone voiceprint microphone system and combining it with adaptive noise suppression and deep neural networks, the problem of low recognition accuracy of traditional voice recognition in noisy environments is solved, silent operation and high-precision command recognition are achieved, which is suitable for a variety of scenarios and protects user privacy.
Patent Information
- Application Number
- CN202510931255.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-07-07
- Publication Date
- 2025-10-14
AI Technical Summary
Traditional speech recognition systems are severely affected by noise in noisy environments and cannot operate silently in quiet scenes, resulting in reduced recognition accuracy or even failure.
A silent command headset audio system based on a bone voiceprint microphone is adopted. The bone voiceprint microphone unit is used to collect bone conduction signals. Combined with the noise suppression, feature extraction and bone voiceprint recognition modules of the signal processing unit, it supports personalized optimization and offline encryption processing to achieve high-precision command recognition.
Achieve high-precision command recognition in noisy environments, avoid noise interference, support silent operation, suitable for quiet scenes, improve recognition accuracy and ensure privacy security.
Smart Images
Figure CN120786221A_ABST
Abstract
Description
TECHNICAL FIELD
[0001] The present application belongs to the technical field of intelligent audio devices, and particularly relates to a bone voiceprint microphone-based silent command earphone audio system. BACKGROUND
[0002] With the rapid development of artificial intelligence and Internet of Things technologies, voice interaction, as a natural and convenient human-computer interaction method, has been widely applied in many fields such as smart home, vehicle-mounted system, mobile terminal and wearable device. Traditional voice recognition systems usually collect user's voice signals through air conduction, that is, using microphones and other acoustic sensors to receive sound propagating in the air and converting it into electrical signals for processing and recognition.
[0003] However, in actual application, traditional voice recognition technology faces many challenges. First, the air conduction method is extremely susceptible to environmental noise interference, especially in high-noise environments (such as streets, factories, vehicle interiors, etc.), background noise often significantly reduces the signal-to-noise ratio of voice signals, resulting in a decrease in voice recognition accuracy, and even misrecognition or inability to recognize; second, in some special scenarios that need to be quiet (such as libraries, conference rooms, hospitals, etc.), users often cannot use voice control by making sounds, and traditional voice recognition systems based on air sound waves cannot work normally without voice input, limiting their applicability in silent or low-voice interaction scenarios. SUMMARY
[0004] The present application aims to provide a bone voiceprint microphone-based silent command earphone audio system to solve the technical problem that existing microphones cannot recognize commands in the absence of sound or under background noise interference.
[0005] To solve the above technical problems, the present application adopts the following technical solutions: The bone voiceprint microphone-based silent command earphone audio system comprises: A bone voiceprint microphone unit for collecting user's voice vibration signals transmitted through the skeleton and converting them into electrical signals; A signal processing unit connected to the bone voiceprint microphone unit, including a noise suppression module, a feature extraction module and a bone voiceprint recognition module; A command recognition module connected to the bone voiceprint recognition module for controlling the operation instructions of the earphone audio system according to the recognition results of the bone voiceprint recognition module; A feedback and display module connected to the command recognition module for feeding back the command execution status to the user through audio, vibration or visual means.
[0006] Furthermore, the noise suppression module detects and filters environmental interference signals in real time through an adaptive noise suppression algorithm; the feature extraction module is used to extract spectral features and time domain features from the bone voiceprint signal; the bone voiceprint recognition module uses a deep neural network or a convolutional neural network to match the extracted features with a pre-stored command database to identify the commands issued by the user.
[0007] Furthermore, the bone voiceprint microphone-based silent command headset audio system also includes a wireless communication module for establishing a Bluetooth or Wi-Fi connection with an external device to receive remote control instructions.
[0008] Furthermore, the silent command headset audio system based on bone voiceprint microphone also includes a personalized optimization module for storing user bone voiceprint training data and generating a personalized command recognition model to improve recognition accuracy.
[0009] Furthermore, the bone voiceprint recognition module supports an offline processing mode and is equipped with a data encryption unit to encrypt the transmitted bone voiceprint biometric data.
[0010] Compared with the existing technology, the beneficial effects of the present invention are: the silent command headset audio system based on the bone voiceprint microphone provided by the present invention realizes high-precision command recognition in a noisy environment through the bone voiceprint signal conduction mechanism, and completely avoids environmental noise interference; at the same time, it supports silent operation to avoid voice privacy leakage, and is suitable for quiet scenes such as libraries; its localized signal processing process (noise suppression → feature extraction → DNN recognition) reduces latency and improves the accuracy of command recognition; bone voiceprint biological data is encrypted and processed offline to eliminate cloud transmission risks, and combined with personalized model training to further ensure security and recognition accuracy, fundamentally solving the failure problem of traditional voice control in complex environments, and taking into account response speed, energy efficiency optimization and privacy protection requirements. BRIEF DESCRIPTION OF THE DRAWINGS
[0011] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings: Figure 1 This is a system architecture diagram of the silent command headset audio system based on bone voiceprint microphone of the present invention; Figure 2 This is a schematic diagram of the bone voiceprint signal collection and processing process; Figure 3 This is a schematic diagram of the bone voiceprint command recognition and feedback mechanism; Figure 4 This is a hardware carrier design diagram of the silent command headset audio system based on bone voiceprint microphone of the present invention. DETAILED DESCRIPTION
[0012] The technical solutions in the embodiments of the present application will be apparently and completely described below with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all the other embodiments obtained by those skilled in the art without creative work belong to the protection scope of the present application.
[0013] The present application will be further described in detail below with the embodiments.
[0014] A specific embodiment of a bone voiceprint microphone-based silent command earphone audio system provided by the present application is as follows: As shown in Figure 1 The bone voiceprint microphone-based silent command earphone audio system mainly comprises a bone voiceprint microphone unit, a signal processing unit, a command recognition module, a feedback and display module and a wireless communication module.
[0015] The bone voiceprint microphone unit mainly adopts a high-sensitivity bone voiceprint sensor VPU, which is attached to the inner side of the ear hook of the earphone, directly contacts the user, and the sensor collects the voice vibration signals conducted by the skull or other bone structures of the user and converts them into electrical signals.
[0016] The signal processing unit integrates a DSP chip, receives the electrical signals of the bone voiceprint microphone, and executes noise suppression, feature extraction and bone voiceprint recognition algorithm.
[0017] The command recognition module is connected with the signal processing unit, receives the command classification results output by the bone voiceprint recognition module, and generates corresponding audio system control instructions (such as playing / pausing, volume adjustment, telephone control).
[0018] The feedback and display module is connected with the command recognition module, and is used for feeding back the command execution state to the user in the form of audio, vibration or vision.
[0019] The wireless communication module supports Bluetooth protocol and is used for connecting external devices such as mobile phones.
[0020] As shown in Figure 2 The signal processing and command control flow chart is shown in the figure. The collected bone voiceprint electrical signals are input into the noise suppression module, and an adaptive noise suppression algorithm (such as spectral subtraction) is used to filter environmental noise. The purified signals are input into the feature extraction module to extract time domain features (short-time energy, zero-crossing rate) and frequency domain features (MFCC). The feature vectors are input into the bone voiceprint recognition module, matched with the local command database through the pre-trained DNN model, and the command classification results (such as “play”) are output. The command recognition module receives the classification results and converts them into device executable instructions.
[0021] As shown in Figure 3As shown, after the command recognition module triggers the operation, the feedback and display module is activated synchronously. The feedback and display module provides multiple feedback methods to ensure that the user can confirm the execution status of the command in a timely manner. The feedback method can be adjusted according to the user's needs and usage scenarios. For example, audio feedback: after executing the "play" command, the speaker plays a prompt tone; vibration feedback: after executing the "reject call" command, the vibration motor generates a specific vibration pattern; visual feedback: after executing the "noise reduction on" command, the LED display is always blue.
[0022] The audio system can also be further optimized. For example, a personalized optimization module can be set up to provide a personalized command recognition model through the user's bone voiceprint training data. The user can perform simple voice training when using it for the first time to make the system adapt to the individual's bone voiceprint characteristics, thereby improving recognition accuracy.
[0023] The bone voiceprint recognition module supports offline processing mode, avoiding uploading the user's bone voiceprint data to the cloud, further improving data privacy and security; and is equipped with a data encryption unit to encrypt the transmitted bone voiceprint biometric data to ensure that the user's biological data is not leaked.
[0024] It should be noted that, in this document, terms such as "comprises", "includes" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, article or apparatus that includes a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements that are inherent to such process, method, article or apparatus.
[0025] While embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to these embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the appended claims and their equivalents.
Claims
1. The silent command headset audio system based on bone voiceprint microphone is characterized by: include: Bone voiceprint microphone unit, used to collect the user's voice vibration signals transmitted through the bones and convert them into electrical signals; The signal processing unit is connected to the bone voiceprint microphone unit and includes a noise suppression module, a feature extraction module, and a bone voiceprint recognition module; A command recognition module, connected to the bone voiceprint recognition module, for controlling the operation instructions of the headphone audio system according to the recognition result of the bone voiceprint recognition module; The feedback and display module is connected to the command recognition module and is used to provide feedback of the command execution status to the user through audio, vibration or visual means.
2. The bone voiceprint microphone-based silent command headset audio system according to claim 1, characterized in that: The noise suppression module detects and filters environmental interference signals in real time through an adaptive noise suppression algorithm; the feature extraction module is used to extract spectral features and time domain features from bone voiceprint signals; The bone voiceprint recognition module uses a deep neural network or a convolutional neural network to match the extracted features with a pre-stored command database to identify the commands issued by the user.
3. The bone voiceprint microphone-based silent command headset audio system according to claim 1, characterized in that: It also includes a wireless communication module for establishing a Bluetooth or Wi-Fi connection with an external device to receive remote control instructions.
4. The bone voiceprint microphone-based silent command headset audio system according to claim 1, characterized in that: It also includes a personalized optimization module for storing user bone voiceprint training data and generating a personalized command recognition model to improve recognition accuracy.
5. The bone voiceprint microphone-based silent command headset audio system according to claim 1, characterized in that: The bone voiceprint recognition module supports an offline processing mode and is equipped with a data encryption unit to encrypt the transmitted bone voiceprint biometric data.