Signal Recognition Method, System and Storage Medium Based on Micro-Nano Optical Fiber Coupler

Abstract

This application relates to the field of signal recognition, and particularly to a signal recognition method, system, and storage medium based on a micro-nano fiber coupler. The recognition method includes obtaining two optical signals collected by the micro-nano fiber coupler for characterizing the muscle movement signals of a subject when speaking; converting the optical signals into electrical signals using a photoelectric converter, and preprocessing the electrical signals to obtain preprocessed data; inputting the preprocessed data into a trained classification model for classification and recognition to obtain expression information corresponding to the preprocessed data; judging the accuracy rate of the expression information; when the accuracy rate reaches a preset value, outputting the expression information. Through the technical solution of this application, the accuracy rate of information recognition is improved, and the expression information of muscle movement signals is accurately decoded, which is beneficial to helping mute patients communicate.

Description

Technical Field

[0001] This application relates to the field of signal recognition, and particularly to a signal recognition method, system, and storage medium based on a micro-nano fiber coupler. Background Art

[0002] Losing one's voice has a great impact on the daily life of patients, especially in aspects such as social interaction, work, and education. People with voice loss need to rely on writing, gestures, or other communication methods during communication, which often makes them face problems such as estrangement, loneliness, and social exclusion. Currently, the medical community uses a variety of methods to treat voice loss, including drugs, surgery, and speech therapy. In addition, some technologies and devices can also help people with voice loss communicate, such as electronic hearing aids, speech synthesizers, etc. In recent years, with the continuous development of artificial intelligence technology, some new speech recognition and synthesis technologies have gradually been applied to the communication of people with voice loss, but the technology is not yet mature. Therefore, there is an urgent need to provide a method for efficiently recognizing signals to accurately decode the true expressions of people with voice loss.

[0003] In related technologies, optical sensors are used to recognize the communication signals of people with voice loss, but generally, sensing is carried out using loss. This method will cause a certain amount of energy loss, resulting in a lower sensing sensitivity of the optical sensor, making the recognized content inaccurate and unable to better decode the true expressions of people with voice loss. Summary of the Invention

[0004] The following is an overview of the subject matter described in detail in this article. This overview is not intended to limit the scope of protection of the claims.

[0005] This application aims to solve at least one of the technical problems existing in the prior art. To this end, embodiments of this application provide a signal recognition method, system, and storage medium based on a micro-nano fiber coupler, which helps to solve the problems of low sensing sensitivity and inaccurate recognized content. This application collects two optical signals used to characterize the muscle movement signals of a subject when speaking through a micro-nano fiber coupler, preprocesses the optical signals by the splitting ratio, improving the sensing sensitivity of the sensor; inputs the preprocessed data into a preset classification model for classification recognition to obtain the corresponding expression information and judge the accuracy rate of the expression information; when the accuracy rate reaches a preset value, the expression information is output, improving the accuracy rate of information recognition, accurately decoding the expression information of muscle movement signals, and helping voice-loss patients communicate.

[0006] In a first aspect, embodiments of this application provide a signal recognition method based on a micro-nano fiber coupler, including obtaining two optical signals collected by the micro-nano fiber coupler and used to characterize the muscle movement signals of a subject when speaking;

[0007] Convert the optical signal into an electrical signal using an optoelectronic converter, and preprocess the electrical signal to obtain preprocessed data;

[0008] Input the preprocessed data into a trained classification model for classification and recognition to obtain expression information corresponding to the preprocessed data;

[0009] Judge the accuracy rate of the expression information;

[0010] When the accuracy rate reaches a preset value, output the expression information.

[0011] At least one of the advantages or beneficial effects of the technical solution of the first aspect of the present application is as follows: Obtain two optical signals collected by a micro-nano fiber coupler for characterizing the muscle movement signals of a subject when speaking, convert the optical signals into electrical signals using an optoelectronic converter and preprocess the electrical signals to obtain preprocessed data, which improves the sensitivity of sensor sensing by processing the electrical signals; Input the preprocessed data into a trained classification model for classification and recognition to obtain expression information corresponding to the preprocessed data, which can accurately decode the expression information of muscle movement signals and is beneficial to helping mute patients communicate; And judge the accuracy rate of the expression information, and only output the expression information when the accuracy rate reaches a preset value; Improve the accuracy rate of information recognition.

[0012] Further, the electrical signal includes a first electrical signal and a second electrical signal, the optical power of the first electrical signal is P1, and the optical power of the second electrical signal is P2. The step of converting the optical signal into an electrical signal using an optoelectronic converter and preprocessing the electrical signal to obtain preprocessed data includes:

[0013] Judge whether the electrical signal is complete;

[0014] If so, perform a splitting ratio process on the electrical signal to obtain preprocessed data, so that the optical power ratio between the first electrical signal and the electrical signal in the preprocessed data is P1 / (P1 + P2);

[0015] If not, discard the electrical signal and end the recognition.

[0016] Further, after the step of converting the optical signal into an electrical signal using an optoelectronic converter and preprocessing the electrical signal to obtain preprocessed data, the method further includes:

[0017] Use a recursive method to extract features from the preprocessed data to obtain phase features;

[0018] Reconstruct the phase features to obtain a phase reconstruction diagram;

[0019] Determine the distance between adjacent points in the phase reconstruction diagram and encode it with grayscale to obtain a phase recurrence diagram.

[0020] Furthermore, the reconstructing the phase feature to obtain a phase reconstruction diagram includes:

[0021] Map the one-dimensional time series of the phase feature to a high-dimensional phase space through time delay;

[0022] Delay the collected time series and time to construct a multi-dimensional phase space;

[0023] Obtain a phase reconstruction diagram according to the multi-dimensional phase space.

[0024] Furthermore, after determining the distance between adjacent points in the phase reconstruction diagram and encoding it with grayscale to obtain a phase recurrence diagram, the method further includes:

[0025] Store the phase recurrence diagram in a database for training the classification model.

[0026] Furthermore, the trained classification model is obtained through the following training method:

[0027] Obtain two optical signals collected by a micro-nano fiber coupler for training to represent the muscle movement signals of the subject when speaking, and store the optical signals and the common sentences corresponding to the optical signals in the database;

[0028] Obtain the phase recurrence diagram corresponding to the optical signal from the database;

[0029] Put the phase recurrence diagrams into a training set and a test set for training and learning according to a preset ratio, pair the phase recurrence diagrams and the common sentences, and when the accuracy of the pairing is greater than a preset value, determine the common sentence as the expressed information;

[0030] Classify and identify the phase recurrence diagram and output the corresponding expressed information to obtain a trained classification model.

[0031] Furthermore, after classifying and identifying the phase recurrence diagram and outputting the corresponding expressed information to obtain a trained classification model, the method further includes:

[0032] Use depthwise separable convolution to extract features of the optical signal reflected in the phase recurrence diagram, where the features include the peak value and frequency of the waveform.

[0033] Furthermore, the method further includes: when the accuracy rate does not reach the preset value, end the recognition.

[0034] Second aspect, an embodiment of the present application provides a signal recognition system based on a micro-nano fiber coupler, including a micro-nano fiber coupler for collecting two optical signals used to characterize the muscle movement signals of a subject when speaking;

[0035] An optoelectronic converter for converting the optical signal into an electrical signal and preprocessing the electrical signal to obtain preprocessed data;

[0036] A controller for inputting the preprocessed data into a preset classification model for classification and recognition to obtain expression information corresponding to the preprocessed data; and for judging the accuracy rate of the expression information; when the accuracy rate reaches a preset value, output the expression information.

[0037] At least one of the technical aspects of the second aspect of the present application has the following advantages or beneficial effects: The signal recognition system based on the micro-nano fiber coupler cooperates with each other through the micro-nano fiber coupler, the optoelectronic converter and the controller. The micro-nano fiber coupler is used to obtain the muscle movement signals of the patient, and the optoelectronic converter performs a splitting ratio process on the signals, improving the sensing sensitivity of the sensor. The controller inputs the preprocessed data into a preset classification model for classification and recognition to obtain expression information corresponding to the preprocessed data; and for judging the accuracy rate of the expression information; when the accuracy rate reaches a preset value, output the expression information, which can accurately decode the expression information of the muscle movement signals, is beneficial to helping mute patients communicate; and only when the accuracy rate reaches the preset value, output the expression information; improving the accuracy rate of information recognition.

[0038] Third aspect, an embodiment of the present application further provides a computer-readable storage medium storing computer-executable instructions for executing the signal recognition method based on the micro-nano fiber coupler described in the technical solution of the first aspect above. Description of the Drawings

[0039] Figure 1 is a flowchart of the steps of a signal recognition method based on a micro-nano fiber coupler provided by an embodiment of the present application;

[0040] Figure 2 is Figure 1 the flowchart of step S200 in

[0041] Figure 3 is a flowchart of the steps of another signal recognition method based on a micro-nano fiber coupler provided by an embodiment of the present application;

[0042] Figure 4 is Figure 3 the flowchart of step S250 in

[0043] Figure 5It is a flowchart of steps of another signal recognition method based on a micro-nano fiber coupler provided by an embodiment of the present application;

[0044] Figure 6 It is a flowchart of steps of a classification model training process provided by an embodiment of the present application;

[0045] Figure 7 It is a flowchart of steps of another classification model training process provided by an embodiment of the present application;

[0046] Figure 8 It is a flowchart of steps of another signal recognition method based on a micro-nano fiber coupler provided by an embodiment of the present application;

[0047] Figure 9 It is a schematic structural diagram of a signal recognition system based on a micro-nano fiber coupler provided by an embodiment of the present application;

[0048] Figure 10 is Figure 9 a schematic structural diagram of the controller in Detailed implementation manners

[0049] In order to make the objectives, technical solutions and advantages of the present application clearer, the present application will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0050] In the description of the present application, "a plurality of" refers to more than two. If there is a description of "first" and "second", it is only for the purpose of distinguishing technical features and cannot be understood as indicating or implying relative importance or implicitly indicating the quantity of the indicated technical features or implicitly indicating the sequence of the indicated technical features.

[0051] In the related art, an optical sensor is used to identify the communication signals of mute groups, but generally, loss is used for sensing. This method will cause certain energy loss, reduce the sensing sensitivity of the optical sensor, and result in inaccurate recognized content, making it impossible to decode the true expressions of mute people well.

[0052] To this end, the embodiments of the present application provide a signal recognition method, system, and storage medium based on a micro-nano fiber coupler, which are beneficial to solving the problems of low sensing sensitivity and inaccurate recognized content. The present application collects two optical signals used to characterize the muscle movement signals of a subject when speaking through a micro-nano fiber coupler, preprocesses the optical signals by the splitting ratio, improving the sensing sensitivity of the sensor; inputs the preprocessed data into a preset classification model for classification and recognition to obtain the corresponding expression information and judge the accuracy rate of the expression information; only outputs the expression information when the accuracy rate reaches the preset value, improving the accuracy rate of information recognition, accurately decoding the expression information of the muscle movement signals, and being beneficial to helping mute patients communicate.

[0053] Referring to Figure 1 , Figure 1 is the step flowchart of a signal recognition method based on a micro-nano fiber coupler provided by the embodiments of the present application, including steps S100 to S500. Specifically,

[0054] S100: Obtain two optical signals collected by the micro-nano fiber coupler and used to characterize the muscle movement signals of a subject when speaking;

[0055] S200: Convert the optical signals into electrical signals and preprocess the electrical signals to obtain preprocessed data;

[0056] S300: Input the preprocessed data into a trained classification model for classification and recognition to obtain the expression information corresponding to the preprocessed data;

[0057] S400: Judge the accuracy rate of the expression information;

[0058] S500: When the accuracy rate reaches the preset value, output the expression information.

[0059] Obtain two optical signals collected by the micro-nano fiber coupler and used to characterize the muscle movement signals of a subject when speaking, convert the optical signals into electrical signals and preprocess the electrical signals to obtain preprocessed data, improving the sensing sensitivity of the sensor by processing the electrical signals; input the preprocessed data into a trained classification model for classification and recognition to obtain the expression information corresponding to the preprocessed data, being able to accurately decode the expression information of the muscle movement signals and being beneficial to helping mute patients communicate; and judge the accuracy rate of the expression information, and only output the expression information when the accuracy rate reaches the preset value; improving the accuracy rate of information recognition.

[0060] It should be noted that in the embodiments of the present application, the preset value can be one or more of 80%, 85%, 90%, 95%, and the preset value can be set according to the actual situation. The embodiments of the present application do not limit the size of the preset value.

[0061] In a specific embodiment, the preset value is 80%. When two optical signals of the muscle movement signal during the patient's speech are acquired, the signals are converted into electrical signals and preprocessed to obtain preprocessed data. The preprocessed data is input into a trained classification model for classification and recognition to obtain expression information corresponding to the preprocessed data. The accuracy rate of the expression information is judged. When the accuracy rate is 85%, the expression information is output. When the accuracy rate is lower than 80%, the expression information is discarded, and two optical signals of the muscle movement signal are acquired again and classified and recognized, improving the accuracy rate of information recognition and facilitating better communication for mute patients.

[0062] In one embodiment, the signal recognition method based on the micro-nano fiber coupler can be used for the daily communication of mute patients and the call for help on hospital beds. In addition, the output expression information can be emitted as sound through an external speaker, realizing the communication function for mute patients.

[0063] It should be noted that the expression information in the embodiments of the present application includes one or more of common words, common sentences, and greetings. The embodiments of the present application do not limit the content of the expression information.

[0064] Refer to Figure 2 , Figure 2 is Figure 1 The flowchart of step S200 in

[0065] S210: Judge whether the electrical signal is complete;

[0066] S220: If it is, perform a splitting ratio process on the electrical signal to obtain preprocessed data, so that the optical power ratio between the first electrical signal and the electrical signal in the preprocessed data is P1 / (P1 + P2);

[0067] S230: If not, discard the electrical signal and end the recognition.

[0068] The optical signal is converted into an electrical signal, and the electrical signal is preprocessed. Among them, the electrical signal includes a first electrical signal P1 and a second electrical signal P2. The optical power of the first electrical signal is P1, and the optical power of the second electrical signal is P2. Judge whether the electrical signal is complete. When the electrical signal is complete, perform a splitting ratio process on the electrical signal to obtain preprocessed data, so that the optical power ratio between the first electrical signal and the electrical signal in the preprocessed data is P1 / (P1 + P2). By performing a splitting ratio process on the signal, the sensing sensitivity of the sensor is improved, the muscle movement information of the patient can be better recognized, and the accuracy rate of information recognition is improved. When the electrical signal is incomplete, the electrical signal is discarded and the recognition is ended, improving the accuracy of recognition.

[0069] It should be noted that in the embodiments of the present application, determining whether an electrical signal is complete includes determining whether the signal points of the electrical signal are greater than a first preset value. When it is greater than or equal to the first preset value, the electrical signal is complete; when it is less than the first preset value, the electrical signal is incomplete.

[0070] It should be noted that the first preset value can be one or more of 1000, 1200, and 1300, and the first preset value can be set according to actual situations. The embodiments of the present application do not limit the first preset value.

[0071] Refer to Figure 3 , Figure 3 is a flowchart of the steps of another signal recognition method based on a micro-nano fiber coupler provided by the embodiments of the present application, including steps S240 to S260. Specifically,

[0072] S240: Use a recursive method to extract features from the preprocessed data to obtain phase features;

[0073] S250: Reconstruct the phase features to obtain a phase reconstruction diagram;

[0074] S260: Determine the distance between adjacent points in the phase reconstruction diagram and encode it with gray scale to obtain a phase recurrence diagram.

[0075] After converting the optical signal into an electrical signal and performing splitting ratio processing on the electrical signal to obtain preprocessed data, the signal recognition method based on the micro-nano fiber coupler further includes using a recursive method to extract phase features by converting the original time series into a new time series from the preprocessed data, and reconstructing the phase features to obtain a phase reconstruction diagram; determining the distance between adjacent points in the phase reconstruction diagram and encoding it with gray scale to obtain a phase recurrence diagram, and improving the accuracy of information recognition by extracting various phase features in the phase recurrence diagram.

[0076] Refer to Figure 4 , Figure 4 is Figure 3 a flowchart of the steps of S250 in

[0077] S251: Map the one-dimensional time series of the phase features to a high-dimensional phase space through time delay;

[0078] S252: Delay the collected time series and time to construct a multi-dimensional phase space;

[0079] S253: Obtain a phase reconstruction diagram according to the multi-dimensional phase space.

[0080] In one embodiment, the phase features are reconstructed to obtain a phase reconstruction diagram, including mapping the one-dimensional time series of the phase features to a high-dimensional phase space through time delay. For the collected time series X(n), the time τ is delayed to construct an m-dimensional phase space, where X(n) = {x(j), x(j + τ),... x(j + (m - 1)τ)}, i = 1, 2,..., N. Calculate the distance between two adjacent points in the m-dimensional phase space and encode it with color or grayscale to obtain the phase reconstruction diagram.

[0081] It should be noted that in the embodiments of the present application, by optimizing the values of τ and m, different phase features can be extracted for different values.

[0082] Refer to Figure 5 , Figure 5 is the flowchart of steps of another signal recognition method based on a micro-nano fiber coupler provided by the embodiments of the present application, including steps S240 to S270. Specifically,

[0083] S240: Use a recursive method to extract features from the preprocessed data to obtain phase features;

[0084] S250: Reconstruct the phase features to obtain a phase reconstruction diagram;

[0085] S260: Determine the distance between two adjacent points in the phase reconstruction diagram and encode it with grayscale to obtain a phase recurrence diagram;

[0086] S270: Store the phase recurrence diagram in a database for training a classification model.

[0087] After using the recursive method to extract features from the preprocessed data to obtain phase features, determining the distance between two adjacent points in the phase reconstruction diagram and encoding it with grayscale to obtain a phase recurrence diagram, the signal recognition method based on a micro-nano fiber coupler further includes storing the phase recurrence diagram in a database for training a classification model. By storing the phase recurrence diagram in the personal database of the subject for training the classification model, it is beneficial to provide a data set for the classification model to train, and at the same time improves the classification accuracy of the classification model.

[0088] Refer to Figure 6 , Figure 6 is the flowchart of steps of a classification model training process provided by the embodiments of the present application, including steps S600 to S900. Specifically,

[0089] S600: Obtain two optical signals collected by the micro-nano fiber coupler for training, which represent the muscle movement signals of the subject when speaking, and store the optical signals and the common sentences corresponding to the optical signals in the database;

[0090] S700: Obtain the phase recurrence plot corresponding to the optical signal from the database;

[0091] S800: Put the phase recurrence plots into the training set and the test set respectively according to a preset ratio for training and learning, pair the phase recurrence plots with common statements, and when the accuracy of the pairing is greater than the preset value, determine the common statement as the expressed information;

[0092] S900: Classify and identify the phase recurrence plots and output the corresponding expressed information to obtain a trained classification model.

[0093] In the embodiment of the present application, the training process of the classification model includes storing two optical signals collected by the micro-nano fiber coupler for training, which characterize the muscle movement signals of the subject when speaking, and the common statements corresponding to the optical signals in the database, obtaining the phase recurrence plots corresponding to the optical signals from the database, putting the phase recurrence plots into the training set and the test set respectively according to a preset ratio for training and learning, pairing the phase recurrence plots with common statements, and when the accuracy of the pairing is greater than the preset value, determining the common statement as the expressed information, classifying and identifying the phase recurrence plots and outputting the corresponding expressed information to obtain a trained classification model. The trained classification model can well identify and distinguish the muscle movement signals of the patient, accurately decode the expressed information of the muscle movement signals, which is beneficial to helping mute patients communicate. At the same time, the model will also judge the accuracy of the expressed information and only output the expressed information when the accuracy reaches the preset value, further improving the accuracy of information recognition.

[0094] It should be noted that in the embodiment of the present application, the preset value can be one or more of 80%, 85%, 90%, 95%, and the preset value can be set according to the actual situation. The present application does not limit the size of the preset value.

[0095] It should be noted that the preset ratio in the embodiment of the present application can be to put the phase recurrence plots into the training set and the test set respectively in one or more of 8:2 and 7:3, which can improve the accuracy of the phase recurrence plot training, reduce the number of model parameters, and reduce the computer memory.

[0096] Refer to Figure 7 , Figure 7 is the step flow chart of another classification model training process provided by the embodiment of the present application, including steps S600 to S1000. Specifically,

[0097] S600: Obtain two optical signals collected by the micro-nano fiber coupler for training, which characterize the muscle movement signals of the subject when speaking, and store the optical signals and the common statements corresponding to the optical signals in the database;

[0098] S700: Obtain the phase recurrence plot corresponding to the optical signal from the database;

[0099] S800: Put the phase recurrence diagrams into the training set and the test set respectively for training and learning according to a preset ratio, pair the phase recurrence diagrams with common statements, and when the accuracy rate of the pairing is greater than the preset value, determine the common statement as the expression information.

[0100] S900: Classify and identify the phase recurrence diagrams and output the corresponding expression information to obtain a trained classification model.

[0101] S1000: Use depthwise separable convolution to extract features from the optical signals reflected in the phase recurrence diagrams, where the features include the peak value and frequency of the waveform.

[0102] After classifying and identifying the phase recurrence diagrams and outputting the corresponding expression information to obtain a trained classification model, the signal recognition method based on a micro-nano fiber coupler further includes using depthwise separable convolution to extract features from the optical signals reflected in the phase recurrence diagrams. By extracting features from the phase recurrence diagrams in the trained classification model, the number of model parameters can be effectively reduced, the memory of the computer can be reduced, the classification efficiency of the classification model can be improved, and at the same time, the training efficiency and accuracy are not affected.

[0103] It should be noted that in the embodiments of the present application, using depthwise separable convolution to extract features from the optical signals reflected in the phase recurrence diagrams includes extracting one or more of the peak value and frequency of the waveform.

[0104] Refer to Figure 8 , Figure 8 is the flowchart of the steps of another signal recognition method based on a micro-nano fiber coupler provided by the embodiments of the present application, including steps S100 to S510. Specifically,

[0105] S100: Obtain two optical signals collected by the micro-nano fiber coupler for characterizing the muscle movement signals of the subject when speaking.

[0106] S200: Convert the optical signals into electrical signals and preprocess the electrical signals to obtain preprocessed data.

[0107] S300: Input the preprocessed data into the trained classification model for classification and identification to obtain the expression information corresponding to the preprocessed data.

[0108] S400: Judge the accuracy rate of the expression information.

[0109] S510: When the accuracy rate does not reach the preset value, end the recognition.

[0110] It should be noted that in the embodiments of the present application, the preset value can be one or more of 80%, 85%, 90%, and 95%. The preset value can be set according to the actual situation, and the size of the preset value is not limited in the embodiments of the present application.

[0111] In one embodiment, the preset value is 80%. When two optical signals of the muscle movement signal during the patient's speech are obtained, the signals are converted into electrical signals and preprocessed to obtain preprocessed data. The preprocessed data is input into a trained classification model for classification and recognition to obtain the expression information corresponding to the preprocessed data, and the accuracy rate of the expression information is judged. When the accuracy rate is 75%, since the accuracy rate does not reach the preset value, the expression information is discarded, and two optical signals of the muscle movement signal are obtained again and classified and recognized. The correct expression information is obtained through re-classification and recognition, which improves the accuracy rate of information recognition and is beneficial to helping laryngectomees communicate better.

[0112] Refer to Figure 9 , Figure 9 FIG. is a schematic structural diagram of a signal recognition system 1000 based on a micro-nano fiber coupler provided by an embodiment of the present application, including a micro-nano fiber coupler 100, a photoelectric converter 200, and a controller 300; wherein, the micro-nano fiber coupler 100 is used to collect two optical signals for characterizing the muscle movement signal during the subject's speech; the photoelectric converter 200 is used to convert the optical signal into an electrical signal and preprocess the electrical signal to obtain preprocessed data; the controller 300 is used to input the preprocessed data into a preset classification model for classification and recognition to obtain the expression information corresponding to the preprocessed data; and is used to judge the accuracy rate of the expression information; when the accuracy rate reaches the preset value, the expression information is output.

[0113] The signal recognition system 1000 based on the micro-nano fiber coupler cooperates with each other through the micro-nano fiber coupler 100, the photoelectric converter 200, and the controller 300. The micro-nano fiber coupler 100 is used to obtain the muscle movement signal of the patient, and the photoelectric converter performs splitting ratio processing on the signal, which improves the sensing sensitivity of the sensor. The controller inputs the preprocessed data into a preset classification model for classification and recognition to obtain the expression information corresponding to the preprocessed data; and is used to judge the accuracy rate of the expression information; when the accuracy rate reaches the preset value, the expression information is output, which can accurately decode the expression information of the muscle movement signal and is beneficial to helping laryngectomees communicate; and the expression information is output only when the accuracy rate reaches the preset value; the accuracy rate of information recognition is improved.

[0114] In one embodiment, the signal recognition system 1000 based on the micro-nano fiber coupler may further include an amplifying module, and the amplifying module is used to emit the output expression information through an external speaker, which can help laryngectomees communicate better.

[0115] Refer to Figure 10, Figure 10 is Figure 9 a schematic structural diagram of the controller 300 in the middle, a schematic hardware structure diagram of a controller 300, including a processor 301, which can be implemented in ways such as a general-purpose CPU (Central Processing Unit), a microprocessor, an application-specific integrated circuit (ASIC), or one or more integrated circuits, etc., and is used to execute relevant programs to implement a signal recognition method based on a micro-nano fiber coupler provided by an embodiment of the present application; a memory 302, which can be implemented in forms such as a read-only memory 302 (ROM), a static storage device, a dynamic storage device, or a random access memory 302 (RAM). The memory 302 can store an operating system and other application programs. When implementing the technical solutions provided by the embodiments of this specification through software or firmware, the relevant program codes are stored in the memory 302 and are called by the processor 301 to execute the embodiments of the present application; an input / output interface 303, which is used to implement information input and output; a communication interface 304, which is used to implement communication interaction between this device and other devices, and can implement communication through a wired method (such as USB, network cable, etc.) or through a wireless method (such as a mobile network, WIFI, Bluetooth, etc.); a bus, which transmits information between various components of the device (such as the processor 301, the memory 302, the input / output interface 303, and the communication interface 304); among which the processor 301, the memory 302, the input / output interface 303, and the communication interface 304 are communicatively connected to each other inside the device through the bus.

[0116] An embodiment of the present application also provides a storage medium, which is a computer-readable storage medium. The storage medium stores a computer program, and when the computer program is executed by a processor, it implements the flowchart of the above-mentioned signal recognition method based on a micro-nano fiber coupler. The memory, as a non-transitory computer-readable storage medium, can be used to store non-transitory software programs and non-transitory computer-executable programs. In addition, the memory can include a high-speed random access memory, and can also include a non-transitory memory, such as at least one magnetic disk storage device, a flash memory device, or other non-transitory solid-state storage devices. In some embodiments, the memory can optionally include a memory remotely set relative to the processor, and these remote memories can be connected to the processor through a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an enterprise intranet, a local area network, a mobile communication network, and combinations thereof.

[0117] The embodiments of the present application have been described in detail above with reference to the accompanying drawings. However, the present application is not limited to the above embodiments, and various changes can be made without departing from the gist of the present application within the scope of knowledge possessed by those of ordinary skill in the art.

Claims

1. A signal recognition method based on a micro-nano fiber coupler, characterized in that, Including: Obtain two optical signals collected by a micro-nano fiber coupler for characterizing the muscle movement signals of a subject when speaking; Use a photoelectric converter to convert the optical signals into electrical signals, and preprocess the electrical signals to obtain preprocessed data; Input the preprocessed data into a trained classification model for classification and recognition to obtain expression information corresponding to the preprocessed data; Judge the accuracy rate of the expression information; When the accuracy rate reaches a preset value, output the expression information; After obtaining the preprocessed data, the method further includes: Use a recursive method to extract features from the preprocessed data to obtain phase features; Reconstruct the phase features to obtain a phase reconstruction diagram; Determine the distance between adjacent two points in the phase reconstruction diagram and encode it with gray scale to obtain a phase recurrence diagram.

2. The signal recognition method based on a micro-nano fiber coupler according to claim 1, wherein The electrical signals include a first electrical signal and a second electrical signal. The optical power of the first electrical signal is P1, and the optical power of the second electrical signal is P2. Using a photoelectric converter to convert the optical signals into electrical signals and preprocessing the electrical signals to obtain preprocessed data includes: Judge whether the electrical signals are complete; If so, perform a splitting ratio process on the electrical signals to obtain preprocessed data, so that the optical power ratio between the first electrical signal and the electrical signals in the preprocessed data is P1 / (P1 + P2); If not, discard the electrical signals and end the recognition.

3. The signal recognition method based on a micro-nano fiber coupler according to claim 1, characterized in that, The reconstructing the phase features to obtain a phase reconstruction diagram includes: Map the one-dimensional time series of the phase features to a high-dimensional phase space through time delay; Delay the collected time series and time to construct a multi-dimensional phase space; Obtain a phase reconstruction diagram according to the multi-dimensional phase space.

4. The signal recognition method based on a micro-nano fiber coupler according to claim 1, characterized in that After determining the distance between adjacent two points in the phase reconstruction diagram and encoding it with gray scale to obtain a phase recurrence diagram, the method further includes: Store the phase recurrence diagram in a database for training the classification model.

5. The signal recognition method based on a micro-nano optical fiber coupler according to claim 4, wherein The trained classification model is obtained through the following training method: Obtain two optical signals collected by a micro-nano fiber coupler for training to characterize the muscle movement signals of a subject when speaking, and store the optical signals and the common sentences corresponding to the optical signals in the database; Obtain the phase recurrence diagram corresponding to the optical signals from the database; Put the phase recurrence diagrams into a training set and a test set for training and learning according to a preset ratio, pair the phase recurrence diagrams and the common sentences, and when the accuracy rate of the pairing is greater than the preset value, determine the common sentences as expression information; Classify and recognize the phase recurrence diagrams and output the corresponding expression information to obtain a trained classification model.

6. The signal recognition method based on a micro-nano fiber coupler according to claim 5, wherein After classifying and recognizing the phase recurrence diagrams and outputting the corresponding expression information to obtain a trained classification model, the method further includes: Use depthwise separable convolution to extract features from the optical signals reflected in the phase recurrence diagram, where the features include one or more of the peak value and frequency of the waveform.

7. The signal recognition method based on a micro-nano fiber coupler according to claim 1, wherein The method further includes: When the accuracy rate does not reach the preset value, end the recognition.

8. A signal recognition system based on a micro-nano fiber coupler, which is applied to the signal recognition method based on a micro-nano fiber coupler as described in any one of claims 1 to 7, and is characterized in that, The system includes: A micro-nano fiber coupler for collecting two optical signals used to characterize the muscle movement signals of a subject when speaking; An optoelectronic converter for converting the optical signal into an electrical signal and preprocessing the electrical signal to obtain preprocessed data; A controller for inputting the preprocessed data into a preset classification model for classification and recognition to obtain expression information corresponding to the preprocessed data; and for judging the accuracy rate of the expression information; when the accuracy rate reaches a preset value, outputting the expression information; The controller is further configured to, after obtaining the preprocessed data, extract features from the preprocessed data by using a recursive method to obtain phase features; reconstruct the phase features to obtain a phase reconstruction diagram; determine the distance between adjacent two points in the phase reconstruction diagram and encode it with gray scale to obtain a phase recurrence diagram.

9. A computer-readable storage medium, characterized in that: Stored with computer-executable instructions, the computer-executable instructions are used to execute the signal recognition method based on the micro-nano fiber coupler according to any one of claims 1 to 7.

Images