Voiceprint recognition voice control method of handheld ultrasonic equipment

Abstract

The invention discloses a voiceprint recognition voice control method for handheld ultrasonic equipment, and belongs to the technical field of voice control. Comprising the following steps: S1, a handheld ultrasonic device is in communication connection with a mobile terminal device; s2, collecting and analyzing the recorded voice through a voiceprint recording module to extract voiceprint features; s3, a control instruction voice template is designed according to the operation requirement of the handheld ultrasonic equipment, and template preprocessing is carried out; s4, collecting and processing a user voice signal, and transmitting the user voice signal to the mobile terminal equipment; and S5, performing recognition through the voiceprint recognition model, judging whether the processed voice signal is matched with the processed voice instruction, if not, triggering a safety mechanism, and if yes, executing corresponding operation by the handheld ultrasonic equipment according to a control instruction obtained by analysis. The technical scheme has the beneficial effects that the voice instruction of the user can be accurately recognized and responded in a noisy environment, the ultrasonic equipment is accurately controlled to execute specialized operation, and the operation safety is ensured.

Description

Technical Field

[0001] The present invention relates to the technical field of voice control, and in particular to a voiceprint recognition voice control method for a handheld ultrasonic device. Background Art

[0002] With the continuous advancement of medical technology, portable ultrasound equipment is increasingly used in medical diagnosis, especially in emergency, field rescue and home medical scenarios. However, in actual use, when operators perform ultrasound scans in noisy environments, the traditional manual operation method is not only inefficient, but may also cause data loss due to accidental touches. In addition, frequent manual operations can easily distract operators.

[0003] Most existing ultrasound devices rely on manual operation, and operators need to free up one hand to adjust device parameters or perform image processing, which is very inconvenient in a busy clinical environment. Although some voice control systems have been applied to medical equipment, these systems still have significant problems with voice recognition accuracy and reliability in noisy environments. Especially when multiple operators are working at the same time, the system may not be able to accurately recognize the instructions of a specific person, thus affecting operational efficiency. Traditional voice recognition systems have the following shortcomings:

[0004] 1. Poor environmental adaptability. Traditional speech recognition systems are easily disturbed by environmental noise, resulting in low recognition accuracy.

[0005] 2. Insufficient security, unable to effectively distinguish user identities, and there is a risk of unauthorized operations.

[0006] 3. Single function: Existing voice control systems often only support basic commands and lack specialized functions for ultrasonic scanning. Summary of the invention

[0007] The purpose of the present invention is to provide a voiceprint recognition voice control method for a handheld ultrasonic device to solve the above technical problems;

[0008] A voiceprint recognition voice control method for a handheld ultrasonic device, comprising:

[0009] Step S1, starting the handheld ultrasound device so that the handheld ultrasound device is in communication connection with the mobile terminal device;

[0010] Step S2, collecting and analyzing the recorded voice through the voiceprint recording module on the mobile terminal device to extract voiceprint features;

[0011] Step S3, designing a control command voice template according to the operation requirements of the handheld ultrasound device and performing template preprocessing to obtain a processed voice command;

[0012] Step S4, collecting and processing the user's voice signal through the microphone, and transmitting the processed voice signal to the mobile terminal device;

[0013] Step S5, the mobile terminal device identifies the processed voice signal through the voiceprint recognition model, and determines whether the processed voice signal matches the processed voice instruction. If not, the security mechanism is triggered. If yes, the processed voice signal is parsed to obtain the control instruction, and the handheld ultrasound device is controlled according to the control instruction.

[0014] Preferably, step S2 comprises,

[0015] Step S21, starting the voiceprint recording module on the mobile terminal device, and pre-emphasizes the recorded voice signal through a high-pass filter to obtain an emphasized signal;

[0016] Step S22, performing frame processing on the emphasized signal to obtain a signal frame;

[0017] Step S23, windowing the signal frame using a window function to obtain a smoothed signal frame;

[0018] Step S24, converting the smoothed signal frame from the time domain to the frequency domain by fast Fourier transform to obtain a spectrum;

[0019] Step S25, mapping the spectrum to the Mel scale, and simulating the nonlinear perception characteristics of the human ear to the sound frequency through the Mel filter bank;

[0020] Step S26, taking the logarithm of each output of the Mel filter bank to obtain a logarithmic energy spectrum;

[0021] Step S27, performing discrete cosine transform on the logarithmic energy spectrum and retaining the first 12-20 coefficients after the transform as the Mel-frequency cepstral coefficient feature vector;

[0022] Step S28, normalizing the Mel-frequency cepstral coefficient feature vector to obtain the voiceprint feature;

[0023] Step S29, encrypt the voiceprint feature and store it in the cloud server.

[0024] Preferably, the filtering formula of the emphasized signal in step S21 is:

[0025] y(t)=x(t)-αx(t-1);

[0026] Wherein, y(t) represents the emphasized signal at time point t, x(t) represents the original speech signal value at time point t, α represents the pre-emphasis coefficient, and α is 0.95 or 0.97.

[0027] Preferably, the formula of the window function in step S23 is:

[0028] w[n]=0.54-0.46cos(2πn / (N-1));

[0029] Wherein, w[n] represents the window function, n represents the index of the sample point in the current window, and N represents the window length.

[0030] Preferably, step S1 comprises,

[0031] Step S11, starting the handheld ultrasound device and the mobile terminal device;

[0032] Step S12, the handheld ultrasound device sends a wireless signal;

[0033] Step S13, selecting the handheld ultrasound device on the mobile terminal device and performing communication configuration.

[0034] Preferably, the control instruction voice template in step S3 includes a start recording instruction, a pause instruction, a screenshot instruction, an end recording instruction and a re-recording instruction.

[0035] Preferably, the voiceprint feature in step S2 is associated with a user account registered by the user in the voiceprint recording module.

[0036] Preferably, the method further includes step S6, feeding back the operation result to the user through screen display, sound prompt or vibration feedback.

[0037] Preferably, the window function in step S23 includes a square window, a Hamming window and a Hanning window.

[0038] Preferably, the safety mechanism in step S5 includes:

[0039] Prompt the user to re-enter the voiceprint; or

[0040] Switch the handheld ultrasound device to manual operation mode.

[0041] The beneficial effects of the present invention are: being able to accurately identify and respond to the user's voice commands in a noisy environment, precisely controlling the ultrasonic equipment to perform professional operations, and at the same time using voiceprint recognition technology to ensure the safety of operations and improve the working efficiency and accuracy of the equipment. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] Figure 1 is a step diagram of a voiceprint recognition voice control method for a handheld ultrasonic device of the present invention;

[0043] Figure 2 is a schematic diagram of step S1 of the present invention;

[0044] Figure 3 It is a schematic diagram of step S2 of the present invention. DETAILED DESCRIPTION

[0045] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0046] It should be noted that, in the absence of conflict, the embodiments of the present invention and the features in the embodiments may be combined with each other.

[0047] The present invention will be further described below in conjunction with the accompanying drawings and specific embodiments, but they are not intended to limit the present invention.

[0048] A voiceprint recognition voice control method for a handheld ultrasonic device, such as Figure 1 As shown, including,

[0049] Step S1, starting the handheld ultrasound device so that the handheld ultrasound device is connected to the mobile terminal device for communication;

[0050] Step S2, collecting and analyzing the recorded voice through the voiceprint recording module on the mobile terminal device to extract voiceprint features;

[0051] Step S3, designing a control command voice template according to the operation requirements of the handheld ultrasound device and performing template preprocessing to obtain a processed voice command;

[0052] Step S4, collecting and processing the user's voice signal through the microphone, and transmitting the processed voice signal to the mobile terminal device;

[0053] Step S5, the mobile terminal device identifies the processed voice signal through the voiceprint recognition model, and determines whether the processed voice signal matches the processed voice command. If not, the security mechanism is triggered. If yes, the processed voice signal is analyzed to obtain the control command, and the handheld ultrasound device is controlled according to the control command.

[0054] Specifically, the present invention provides a voiceprint recognition speech control method for a handheld ultrasound device, which compares real-time voice commands with voiceprints recorded in advance. On the one hand, it can accurately identify and respond to the user's voice commands to avoid environmental interference; on the other hand, it makes the operation safer and avoids the risk of unauthorized operation. The control command voice template can be customized as needed to realize the professional function operation of ultrasound scanning.

[0055] In a preferred embodiment, referring to Figure 3 , step S2 comprises,

[0056] Step S21, starting the voiceprint recording module on the mobile terminal device, and pre-emphasizes the recorded voice signal through a high-pass filter to obtain an emphasized signal;

[0057] Step S22, performing frame processing on the emphasized signal to obtain a signal frame;

[0058] Step S23, adding a window to the signal frame by using a window function to obtain a smoothed signal frame;

[0059] Step S24, converting the smoothed signal frame from the time domain to the frequency domain by fast Fourier transform to obtain a spectrum;

[0060] Step S25, mapping the spectrum to the Mel scale, and simulating the nonlinear perception characteristics of the human ear to the sound frequency through the Mel filter bank;

[0061] Step S26, taking the logarithm of each output of the Mel filter bank to obtain a logarithmic energy spectrum;

[0062] Step S27, performing discrete cosine transform on the logarithmic energy spectrum and retaining the first 12-20 coefficients after the transform as the Mel-frequency cepstral coefficient feature vector;

[0063] Step S28, normalizing the Mel-frequency cepstral coefficient feature vector to obtain the voiceprint feature;

[0064] Step S29, encrypting the voiceprint feature and storing it in the cloud server;

[0065] The window functions in step S23 include square window, Hamming window and Hanning window.

[0066] Specifically, the purpose of the pre-emphasis in step S21 is to enhance the high frequency part of the speech signal and make the spectrum of the signal flat for convenience of subsequent processing.

[0067] The purpose of the frame division in step S22 is that the speech signal can be regarded as stable in a short period of time (such as 10ms-30ms), so the signal is divided into multiple short frames for processing. Usually the frame length is set to 20ms-40ms, the commonly used frame length is 25ms, and the frame shift (or step length) is 10ms (that is, there is a 15ms overlap between frames).

[0068] The purpose of adding a window in step S23 is to eliminate the discontinuity of the frame edge and reduce spectrum leakage. Commonly used window functions include square window, Hamming window and Hanning window, among which the Hamming window is often used due to its good frequency domain characteristics.

[0069] The purpose of step S24 is to convert each frame of signal from the time domain to the frequency domain to obtain spectrum information. Perform FFT transformation on each frame of signal after windowing to obtain spectrum. The number of points N of FFT is usually 256 or 512, depending on the specific implementation.

[0070] Step S25 Mel filter banks are designed to simulate the nonlinear perception of sound frequency by the human ear. The spectrum is mapped to the Mel scale, and a set of Mel-scale triangular filter banks are designed. Each filter in the filter bank covers a frequency range on the Mel scale, and the center frequencies of adjacent filters are equally spaced on the Mel scale.

[0071] The purpose of the logarithmic operation in step S26 is that since the human ear's perception of sound is nonlinear, the output of the filter bank is subjected to a logarithmic operation to better simulate the auditory characteristics of the human ear.

[0072] The purpose of discrete cosine transform (DCT) in step S27 is to further reduce the dimension and decorrelate, and extract features useful for speech recognition.

[0073] The purpose of normalization in step S28 is to make the feature vectors more consistent and comparable and improve recognition performance. Commonly used normalization methods include cepstral mean normalization (CMN) and global feature normalization.

[0074] In a preferred embodiment, the filtering formula of the emphasized signal in step S21 is:

[0075] y(t)=x(t)-ax(t-1);

[0076] Wherein, y(t) represents the emphasized signal at time point t, x(t) represents the original speech signal value at time point t, α represents the pre-emphasis coefficient, and α is 0.95 or 0.97.

[0077] In a preferred embodiment, the formula of the window function in step S23 is:

[0078] w[n]=0.54-0.46cos(2πn / (N-1));

[0079] Among them, w[n] represents the window function, n represents the index of the sample point in the current window, and N represents the window length.

[0080] In a preferred embodiment, referring to Figure 2 , step S1 comprises,

[0081] Step S11, starting the handheld ultrasound device and the mobile terminal device;

[0082] Step S12, the handheld ultrasound device sends a wireless signal;

[0083] Step S13, selecting a handheld ultrasound device on the mobile terminal device and performing communication configuration.

[0084] Specifically, the user first ensures that both the handheld ultrasound device and the mobile terminal device (such as a smartphone or tablet) are turned on and that the application (voiceprint recording module) has been installed. The handheld ultrasound device will send out a WiFi signal for the mobile terminal device to discover. The user opens the corresponding connection application or setting on the mobile terminal device, searches for and selects the handheld ultrasound device for pairing. According to the prompts, enter the pairing code to confirm the pairing request and complete the connection between the devices. After the connection is successful, the mobile terminal device will display the connection status and perform further configuration, such as setting the device name, selecting the connection method, etc. The user completes the configuration according to the prompts to ensure unimpeded communication between the handheld ultrasound device and the mobile terminal device.

[0085] In a preferred embodiment, the control instruction voice template in step S3 includes a start recording instruction, a pause instruction, a screenshot instruction, an end recording instruction and a re-recording instruction.

[0086] Specifically, a series of concise and clear control command voice templates are designed according to the operation requirements of the handheld ultrasound device, such as start recording, pause, screenshot, end recording, re-record, etc. The designed command templates are preprocessed, including text-to-speech conversion, denoising, feature extraction, etc., so as to be matched with the user's real-time voice signal later.

[0087] In a preferred embodiment, the voiceprint feature in step S2 is associated with the user account registered by the user in the voiceprint recording module.

[0088] Specifically, users need to log in or register on the application to create a unique user account. This account will be used to associate the user's voiceprint features and subsequent device operations. The user opens the application voiceprint registration page and starts voiceprint recording according to the page prompts. The application will show the user one or more preset text or number sequences. These sequences are usually designed to contain sufficient phoneme changes to accurately extract voiceprint features. Users need to read these sequences according to the prompts. The application uses the microphone of the mobile terminal device to record the user's voice. During the recording process, the application will automatically adjust the recording parameters (such as gain, noise reduction, etc.) to optimize the recording quality.

[0089] After the recording is completed, the application processes the recording file and extracts the user's voiceprint features. This process includes steps such as signal preprocessing (such as denoising, endpoint detection), feature parameter calculation (Mel-frequency cepstral coefficient MFCC) and feature vectorization. The extracted voiceprint features will be encrypted to ensure data security. Then, these encrypted voiceprint features will be stored in the cloud server. When stored, the voiceprint features will be associated with the user's unique identifier (user ID).

[0090] In a preferred embodiment, step S6 is also included, in which the operation results are fed back to the user through screen display, sound prompts or vibration feedback, so as to enhance the user experience and make the operation results clearer; improve the immediacy and intuitiveness of the interaction; reduce operation errors and help users quickly confirm or adjust their behavior.

[0091] In a preferred embodiment, the safety mechanism in step S5 includes:

[0092] Prompt the user to re-enter the voiceprint; or

[0093] Switch the handheld ultrasound device to manual operation mode.

[0094] Specifically, if voiceprint recognition fails (i.e., it cannot be confirmed that it is the voice of a registered user), the application can take a variety of security measures, such as requiring the user to re-enter the voiceprint, switching to manual operation mode, or limiting device functions, etc. Based on user feedback and actual needs, the accuracy and robustness of the voiceprint recognition model, as well as the friendliness and timeliness of the feedback mechanism, are continuously optimized.

[0095] The present invention continuously monitors the sounds of the surrounding environment through the built-in microphone of the handheld ultrasonic device, and starts to collect voice signals when sounds are detected. The collected voice signals are preprocessed, including denoising, gain adjustment, endpoint detection, etc., to improve the accuracy of subsequent voiceprint recognition.

[0096] The pre-processed voice signal is transmitted to the mobile terminal device in real time through wireless communication technologies such as WiFi or Bluetooth.

[0097] On the mobile terminal device, the trained voiceprint recognition model is used to perform real-time recognition of the received voice signal. The recognition process includes comparing the features of the voice signal with the stored voiceprint features, calculating the similarity, etc. If the voiceprint recognition is successful (that is, it is confirmed to be the voice of the registered user), the voice content is further parsed. The parsing process includes voice recognition (converting voice into text) and semantic understanding (parsing text into specific control instructions). Finally, the parsed control instructions are sent to the handheld ultrasound device via WiFi. After receiving the instructions, the handheld ultrasound device performs the corresponding operations, such as starting recording, pausing, taking screenshots, etc. After the execution is completed, the handheld ultrasound device or mobile terminal device feedbacks the operation results to the user. The feedback method can be screen display, sound prompt or vibration feedback.

[0098] In summary, this application provides a voiceprint recognition voice control method for a handheld ultrasound device. By introducing voiceprint recognition technology, accurate voice control in a noisy environment is achieved, greatly improving the convenience and efficiency of user operation. At the same time, the personalized voiceprint recognition mechanism ensures that only registered users can control the device, enhancing the security and privacy protection of the system. In addition, the system is highly scalable and flexible, and the control instruction set can be adjusted according to actual needs to meet the usage needs in different scenarios.

[0099] The above description is only a preferred embodiment of the present invention, and does not limit the implementation mode and protection scope of the present invention. For those skilled in the art, it should be aware that all solutions obtained by equivalent substitutions and obvious changes made using the description and illustrations of the present invention should be included in the protection scope of the present invention.

Claims

1. A voiceprint recognition voice control method for a handheld ultrasonic device, characterized in that: include, Step S1, starting the handheld ultrasound device so that the handheld ultrasound device is in communication connection with the mobile terminal device; Step S2, collecting and analyzing the recorded voice through the voiceprint recording module on the mobile terminal device to extract voiceprint features; Step S3, designing a control command voice template according to the operation requirements of the handheld ultrasound device and performing template preprocessing to obtain a processed voice command; Step S4, collecting and processing the user's voice signal through the microphone, and transmitting the processed voice signal to the mobile terminal device; Step S5, the mobile terminal device identifies the processed voice signal through the voiceprint recognition model, and determines whether the processed voice signal matches the processed voice instruction. If not, the security mechanism is triggered. If yes, the processed voice signal is parsed to obtain the control instruction, and the handheld ultrasound device is controlled according to the control instruction.

2. The voiceprint recognition voice control method of a handheld ultrasonic device according to claim 1, characterized in that: Step S2 comprises, Step S21, starting the voiceprint recording module on the mobile terminal device, and pre-emphasizes the recorded voice signal through a high-pass filter to obtain an emphasized signal; Step S22, performing frame processing on the emphasized signal to obtain a signal frame; Step S23, windowing the signal frame using a window function to obtain a smoothed signal frame; Step S24, converting the smoothed signal frame from the time domain to the frequency domain by fast Fourier transform to obtain a spectrum; Step S25, mapping the spectrum to the Mel scale, and simulating the nonlinear perception characteristics of the human ear to the sound frequency through the Mel filter bank; Step S26, taking the logarithm of each output of the Mel filter bank to obtain a logarithmic energy spectrum; Step S27, performing discrete cosine transform on the logarithmic energy spectrum and retaining the first 12-20 coefficients after the transform as the Mel-frequency cepstral coefficient feature vector; Step S28, normalizing the Mel-frequency cepstral coefficient feature vector to obtain the voiceprint feature; Step S29, encrypt the voiceprint feature and store it in the cloud server.

3. The voiceprint recognition voice control method of a handheld ultrasonic device according to claim 2, characterized in that: The filtering formula of the emphasized signal in step S21 is: y(t)=x(t)-αx(t-1); Wherein, y(t) represents the emphasized signal at time point t, x(t) represents the original speech signal value at time point t, α represents the pre-emphasis coefficient, and α is 0.95 or 0.

97.

4. The voiceprint recognition voice control method of a handheld ultrasonic device according to claim 2, characterized in that: The formula of the window function in step S23 is: w[n]=0.54-0.46cos(2πn / (N-1)); Wherein, w[n] represents the window function, n represents the index of the sample point in the current window, and N represents the window length.

5. The voiceprint recognition voice control method of a handheld ultrasonic device according to claim 1, characterized in that: Step S1 comprises, Step S11, starting the handheld ultrasound device and the mobile terminal device; Step S12, the handheld ultrasound device sends a wireless signal; Step S13, selecting the handheld ultrasound device on the mobile terminal device and performing communication configuration.

6. The voiceprint recognition voice control method of a handheld ultrasonic device according to claim 1, characterized in that: The control command voice template in step S3 includes a start recording command, a pause command, a screenshot command, an end recording command and a re-recording command.

7. The voiceprint recognition voice control method of a handheld ultrasonic device according to claim 1, characterized in that: In step S2, the voiceprint feature is associated with the user account registered by the user in the voiceprint recording module.

8. The voiceprint recognition voice control method of a handheld ultrasonic device according to claim 1, characterized in that: The method further includes step S6, wherein the operation result is fed back to the user through screen display, sound prompt or vibration feedback.

9. The voiceprint recognition voice control method of a handheld ultrasonic device according to claim 1, characterized in that: The window function in step S23 includes a square window, a Hamming window and a Hanning window.

10. The voiceprint recognition voice control method of a handheld ultrasonic device according to claim 1, characterized in that: The safety mechanism in step S5 includes: Prompt the user to re-enter the voiceprint; or Switch the handheld ultrasound device to manual operation mode.

Images