EEG-based language recognition methods, devices, and computer equipment

Abstract

This application relates to a language recognition method, apparatus, and computer device based on electroencephalogram (EEG) signals. The method includes: extracting features from the EEG signal to be recognized to obtain feature information of the EEG signal; inputting the feature information into a trained recognition model to obtain encoded information corresponding to the EEG signal; and determining the audio signal corresponding to the EEG signal based on the encoded information. Since determining the encoded information of the EEG signal to be recognized using its feature information has high accuracy, converting the encoded information to obtain the audio signal further improves the accuracy of the audio signal. Furthermore, determining the EEG signal based on the labeled audio signal improves the security of information transmission compared to directly acquiring the original speech audio.

Description

Technical Field

[0001] This application relates to the field of signal recognition technology, and in particular to a language recognition method, apparatus and computer device based on electroencephalogram (EEG) signals. Background Technology

[0002] Language Brain-Computer Interface (LBCI) is a technology that converts brain activity into language output. Currently, language prostheses based on invasive neural signal acquisition technology can achieve relatively fast communication. Decoding Chinese language can generate audio or achieve end-to-end output of brain activity into sentences.

[0003] In traditional technologies, machine learning algorithms, such as deep learning models, are used to analyze and decode the collected neural signals, converting them into understandable language information.

[0004] However, when using traditional methods for recognizing electroencephalography (EEG) signals, there is a problem of low recognition accuracy. Summary of the Invention

[0005] Therefore, it is necessary to provide a language recognition method, device, and computer equipment based on EEG signals that can improve the accuracy of EEG recognition, addressing the aforementioned technical problems.

[0006] In a first aspect, this application provides a language recognition method based on electroencephalogram (EEG) signals, the method comprising:

[0007] Feature extraction is performed on the EEG signal to be identified to obtain the feature information of the EEG signal to be identified; the EEG signal to be identified is obtained based on the labeled audio signal.

[0008] The feature information is input into the trained recognition model to obtain the encoded information corresponding to the EEG signal to be recognized;

[0009] Based on the encoded information, the audio signal corresponding to the EEG signal to be identified is determined.

[0010] In one embodiment, determining the audio signal corresponding to the EEG signal to be identified based on the encoded information includes:

[0011] Based on the encoding information and the encoding pattern corresponding to the encoding information, the text information corresponding to the encoding information is determined;

[0012] Based on the text information corresponding to the encoded information, an audio signal corresponding to the EEG signal to be identified is generated.

[0013] In one embodiment, determining the text information corresponding to the encoding information based on the encoding information and the encoding pattern corresponding to the encoding information includes:

[0014] The encoded information is divided into multiple fields;

[0015] The multiple fields are converted according to the encoding pattern to obtain multiple text symbols;

[0016] The multiple text symbols are combined to obtain the text information corresponding to the encoded information.

[0017] In one embodiment, the step of extracting features from the EEG signal to be identified to obtain feature information of the EEG signal includes:

[0018] The EEG signal to be identified is subjected to spatial filtering to obtain the filtered EEG signal;

[0019] Event-related spectral perturbation (ERSP) features are extracted from the filtered EEG signal to obtain the feature information of the EEG signal to be identified.

[0020] In one embodiment, the method further includes:

[0021] The labeled audio signal is received, and the EEG signal to be identified is obtained based on the labeled audio signal; the labeled audio signal is generated by converting the original language audio.

[0022] In one embodiment, the marked audio signal is any one of a frequency-modulated continuous wave (FM) chrip signal, a baseband signal, an amplitude-modulated (AM) signal, and a frequency-modulated (FM) signal.

[0023] Secondly, this application also provides a speech recognition device based on electroencephalogram (EEG) signals, the device comprising:

[0024] The extraction module is used to extract features from the EEG signal to be identified, thereby obtaining the feature information of the EEG signal to be identified; the EEG signal to be identified is obtained based on the labeled audio signal.

[0025] The recognition module is used to input the feature information into the trained recognition model to obtain the encoded information corresponding to the EEG signal to be recognized;

[0026] The determining module is used to determine the audio signal corresponding to the EEG signal to be identified based on the encoded information.

[0027] Thirdly, this application also provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to perform the following steps:

[0028] Feature extraction is performed on the EEG signal to be identified to obtain the feature information of the EEG signal to be identified; the EEG signal to be identified is obtained based on the labeled audio signal.

[0029] The feature information is input into the trained recognition model to obtain the encoded information corresponding to the EEG signal to be recognized;

[0030] Based on the encoded information, the audio signal corresponding to the EEG signal to be identified is determined.

[0031] Fourthly, this application also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, performs the following steps:

[0032] Feature extraction is performed on the EEG signal to be identified to obtain the feature information of the EEG signal to be identified; the EEG signal to be identified is obtained based on the labeled audio signal.

[0033] The feature information is input into the trained recognition model to obtain the encoded information corresponding to the EEG signal to be recognized;

[0034] Based on the encoded information, the audio signal corresponding to the EEG signal to be identified is determined.

[0035] Fifthly, this application also provides a computer program product, including a computer program that, when executed by a processor, performs the following steps:

[0036] Feature extraction is performed on the EEG signal to be identified to obtain the feature information of the EEG signal to be identified; the EEG signal to be identified is obtained based on the labeled audio signal.

[0037] The feature information is input into the trained recognition model to obtain the encoded information corresponding to the EEG signal to be recognized;

[0038] Based on the encoded information, the audio signal corresponding to the EEG signal to be identified is determined.

[0039] The aforementioned speech recognition method, device, and computer equipment based on electroencephalogram (EEG) signals extract features from the EEG signals to be recognized, obtaining feature information of the EEG signals to be recognized; inputting the feature information into a trained recognition model to obtain the encoded information corresponding to the EEG signals to be recognized; and determining the corresponding audio signal based on the encoded information. Since determining the encoded information of the EEG signals to be recognized through the feature information of the EEG signals to be recognized has high accuracy, further conversion of the encoded information to obtain the audio signal improves the accuracy of the audio signal. Moreover, determining the EEG signals to be recognized based on the labeled audio signal improves the security of information transmission compared to directly acquiring the original speech audio. Attached Figure Description

[0040] To more clearly illustrate the technical solutions in the embodiments of this application or related technologies, the drawings used in the description of the embodiments of this application or related technologies will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0041] Figure 1 This is an application environment diagram of a speech recognition method based on electroencephalogram (EEG) signals in one embodiment;

[0042] Figure 2 This is a flowchart illustrating a language recognition method based on electroencephalogram (EEG) signals in one embodiment.

[0043] Figure 3 This is a flowchart illustrating a language recognition method based on electroencephalogram (EEG) signals in another embodiment;

[0044] Figure 4 This is a flowchart illustrating a language recognition method based on electroencephalogram (EEG) signals in another embodiment;

[0045] Figure 5 This is a flowchart illustrating a language recognition method based on electroencephalogram (EEG) signals in another embodiment;

[0046] Figure 6 This is a flowchart illustrating a language recognition method based on electroencephalogram (EEG) signals in another embodiment;

[0047] Figure 7 This is a schematic diagram of various labeled audio signals in one embodiment;

[0048] Figure 8 This is a flowchart illustrating a language recognition method based on electroencephalogram (EEG) signals in another embodiment;

[0049] Figure 9 This is a structural block diagram of a speech recognition device based on electroencephalogram (EEG) signals in one embodiment;

[0050] Figure 10 This is an internal structural diagram of a computer device in one embodiment. Detailed Implementation

[0051] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0052] The language recognition method based on electroencephalogram (EEG) signals provided in this application can be applied to, for example... Figure 1 In the application environment shown, electronic device 102 communicates with server 104 via a network. A data storage system can store the data that server 104 needs to process. The data storage system can be integrated onto server 104 or placed in the cloud or on other network servers. Electronic device 102 sends the EEG signal to be identified to server 104, so that server 104 can identify the EEG signal and determine the corresponding audio signal. Electronic device 102 can be a terminal or a server. Terminals can be, but are not limited to, various personal computers, laptops, smartphones, tablets, IoT devices, and portable wearable devices. IoT devices can be smart speakers, smart TVs, smart air conditioners, smart in-vehicle devices, projection devices, etc. Portable wearable devices can be smartwatches, smart bracelets, head-mounted devices, etc. Head-mounted devices can be virtual reality (VR) devices, augmented reality (AR) devices, smart glasses, etc. Server 104 can be a standalone physical server, a server cluster or distributed system composed of multiple physical servers, or a cloud server providing cloud computing services.

[0053] In one embodiment, such as Figure 2 As shown, a language recognition method based on electroencephalogram (EEG) signals is provided, which can be applied to... Figure 1 Taking the server in the example of this, the explanation includes:

[0054] S201, extract features from the EEG signal to be identified to obtain the feature information of the EEG signal to be identified.

[0055] In this embodiment, a labeled audio signal is played, and the target object's electroencephalogram (EEG) signal is acquired using a signal acquisition device. Feature extraction is then performed on the acquired EEG signal to obtain its feature information. Optionally, the feature information may include frequency domain feature information, time domain feature information, spatial domain feature information, etc.

[0056] Optionally, the signal acquisition device can be a wearable device or a medical-grade device. For example, a wearable device can be a head-mounted acquisition device, and a medical-grade device can be a multi-channel electroencephalogram (EEG) acquisition device.

[0057] S202, input the feature information into the trained recognition model to obtain the encoded information corresponding to the EEG signal to be recognized.

[0058] The encoding mode of the encoded information can be Morse code, binary code, etc.

[0059] In this embodiment, the recognition model can be a neural network model, which can annotate preset feature information to form sample information, thereby using the sample information to train the initial neural network model and obtain a trained recognition model. The feature information is then used as input to the recognition model, which analyzes and processes the feature information to obtain the encoded information corresponding to the EEG signal to be recognized.

[0060] Optionally, during the training phase, EEG signals under various labeled audio stimuli are collected multiple times, and event-related spectral perturbation (ERSP) features are extracted and used to train a support vector machine (SVM) binary classification model.

[0061] S203, Based on the encoding information, determine the audio signal corresponding to the EEG signal to be identified.

[0062] As an alternative implementation method, a conversion model can be generated based on the correspondence between each field in the encoded information and each signal in the audio signal. The encoded information can then be input into the conversion model to determine the audio signal corresponding to the EEG signal to be identified.

[0063] As another alternative implementation method, the text element corresponding to the EEG signal to be identified can be determined first based on the encoding information, and then the audio signal corresponding to the EEG signal to be identified can be generated based on the text element.

[0064] In the aforementioned speech recognition method based on electroencephalogram (EEG) signals, features are extracted from the EEG signals to be recognized, yielding feature information. This feature information is then input into a trained recognition model to obtain the encoded information corresponding to the EEG signals. Based on this encoded information, the corresponding audio signal is determined. Since determining the encoded information of the EEG signals from their feature information is highly accurate, further conversion of the encoded information to obtain the audio signal improves the accuracy of the audio signal. Moreover, determining the EEG signals based on the labeled audio signal enhances the security of information transmission compared to directly acquiring the original speech audio.

[0065] In one embodiment, an implementation of the above S203 is provided, such as... Figure 3 As shown, the above-mentioned "determining the audio signal corresponding to the EEG signal to be identified based on the encoding information" includes:

[0066] S301, determine the text information corresponding to the encoding information based on the encoding information and the encoding mode corresponding to the encoding information.

[0067] In this embodiment, a corresponding conversion table is determined based on the encoding mode corresponding to the encoding information, such as a Morse code table or an ASCII code table, so as to convert the encoding information into text information.

[0068] As an optional implementation, when the encoding mode of the encoded information is Morse code, the encoded information can be converted into text information according to the Morse code table.

[0069] As another alternative implementation method, when the encoding mode of the encoded information is binary code, the encoded information can be converted into text information according to the ASCII code correspondence table.

[0070] S302, based on the text information corresponding to the encoded information, generates the audio signal corresponding to the EEG signal to be recognized.

[0071] In this embodiment of the application, the text information corresponding to the encoded information is output as audio to obtain the audio signal corresponding to the EEG signal to be identified.

[0072] As an optional implementation, language conversion rules can be predefined, such as syllable pronunciation rules, linking rules, etc., and text information can be converted into audio signals according to the language conversion rules.

[0073] As another alternative implementation method, a large amount of speech and text data can be collected, and a statistical model can be used to learn the mapping relationship from text to speech. The text information corresponding to the encoded information can then be input into the statistical model to generate the audio signal corresponding to the EEG signal to be identified.

[0074] As another alternative implementation, deep neural network models, such as recurrent neural networks, long short-term memory networks, and transformers, can be pre-built. Deep learning models can automatically learn complex mapping relationships from text to speech, thereby converting text information into audio signals according to the deep neural network model.

[0075] In the above-mentioned application embodiments, the encoded information is first converted into text information according to the encoding mode corresponding to the encoded information, and then the text information is converted into audio signal. The audio signal corresponding to the EEG signal to be identified is determined through two conversions, thereby improving the accuracy of the audio signal.

[0076] In one embodiment, one implementation of the above-described S301 is provided, such as... Figure 4 As shown, the above-mentioned "determining the text information corresponding to the encoded information based on the encoded information and the encoding pattern corresponding to the encoded information" includes:

[0077] S401 divides the encoded information into multiple fields.

[0078] In this embodiment, the encoded information includes encoded symbols and interval positions, wherein the interval positions can be markers such as spaces, commas, semicolons, etc. The encoded information is then divided according to the interval positions to obtain multiple fields, each field including at least one encoded symbol. Optionally, when the encoding mode is binary code, the encoded symbols can be 0 and 1; when the encoding mode is Morse code, the encoded symbols can be dots and hyphens.

[0079] S402 converts multiple fields according to the encoding mode to obtain multiple text symbols.

[0080] The text symbols can include Chinese characters, English characters, etc.

[0081] In this embodiment, multiple fields are converted according to the conversion table corresponding to the encoding mode, and each field is converted into a text symbol to obtain multiple text symbols corresponding to the EEG signal to be identified.

[0082] For example, when the Morse code is “·-,·--·,·--·,·-··,·”, the encoded information is divided into 5 fields, namely ·-,·--·,·--·,·-··,·. The 5 alphanumeric symbols obtained by converting each field are A, P, P, L, and E.

[0083] S403 combines multiple character symbols to obtain the text information corresponding to the encoded information.

[0084] In this embodiment of the application, multiple text symbols are arranged sequentially according to the corresponding field order to obtain the text information corresponding to the encoded information.

[0085] For example, when multiple text symbols are A, P, P, L, E, the multiple text symbols are combined to obtain the text information corresponding to the encoded information, which is APPLE.

[0086] In the above-mentioned application embodiments, each field in the encoded information is converted sequentially, and the multiple text symbols obtained from the conversion are combined to obtain the text information corresponding to the encoded information, thus ensuring the reliability and accuracy of the text information.

[0087] In one embodiment, one implementation of S201 above is provided, such as... Figure 5 As shown, the above-mentioned "feature extraction of the EEG signal to be identified to obtain the feature information of the EEG signal to be identified" includes:

[0088] S501 performs spatial filtering on the EEG signal to be identified to obtain the filtered EEG signal.

[0089] In this embodiment of the application, if the collected EEG signal to be identified is a multi-lead signal, then the EEG signal to be identified is subjected to spatial filtering to reduce the interference signal in the EEG signal to be identified, and the filtered EEG signal is obtained.

[0090] S502, extract the Event Related Spectral Perturbation (ERSP) feature from the filtered EEG signal to obtain the feature information of the EEG signal to be identified.

[0091] In this embodiment, event-related spectral perturbation (ERSP) features are extracted from the filtered EEG signal, as shown in Equation 1:

[0092] (Equation 1)

[0093] in Represents the number of trials. Representative trial In frequency and time Spectral estimation at the location.

[0094] Optionally, after extracting ERSP features, The data is input into an SVM model for classification, and the system plays the corresponding audio based on the classification results.

[0095] It should be noted that in this embodiment of the application, if the collected EEG signal to be identified is not a multi-lead signal, then the above S501 step does not need to be performed.

[0096] In the above-mentioned embodiments, the EEG signal to be identified is preprocessed to reduce noise and other interference in the EEG signal, and then the preprocessed EEG signal is used for feature extraction, which improves the accuracy of feature information.

[0097] In one embodiment, such as Figure 6 As shown, the above-mentioned language recognition method based on EEG signals also includes:

[0098] S204, receives the labeled audio signal and obtains the EEG signal to be identified based on the labeled audio signal; the labeled audio signal is generated by converting the original language audio.

[0099] Optionally, the audio signal can be labeled as any one of the following: frequency modulated continuous wave (Chrip) signal, baseband signal, amplitude modulation (AM) signal, and frequency modulation (FM) signal.

[0100] In this embodiment, the electronic device acquires the original audio signal and converts it into encoded information. Based on the encoded information, the original audio signal is then converted into a labeled audio signal. The original audio signal includes original language elements, such as English letters, Chinese characters, and Cyrillic letters. The labeled audio signal can be a chirp signal, a frequency-modulated signal, or an amplitude-modulated signal. A chirp signal is a frequency-modulated continuous wave (FMCW) signal. The encoded information can be of the type of Morse code or binary code. Furthermore, the electronic device plays the labeled audio signal and acquires the electroencephalogram (EEG) signals of the target object during the playback using a data acquisition device.

[0101] Optionally, schematic diagrams of various labeled audio signals can be as follows: Figure 7 As shown, the electronic device can first generate a first correspondence between encoded information and language elements in the audio signal, and further, generate a second correspondence between the marked audio signal and the encoded information based on the amplitude and frequency of each marked audio signal. Thus, the original audio signal is converted into a marked audio signal based on the first and second correspondences.

[0102] For example, frequency increases and decreases in the chirp signal are categorized into two types of labeled audio. Frequency increase can be represented as "up," and frequency decrease as "down." "Up" is defined as a · in Morse code, and "down" as a - in Morse code, thus obtaining a coded sequence of 26 letters. For instance, the Morse code for the letter "A" is "·-," and the audio label for "A" is "up down," while the audio label for the letter "B" is "down up up up." During language communication, each letter is combined with an audio label. For example, the word "APPLE," after audio labeling, becomes "A up down, P up down down up, P up down down up, L up down up up, Eup." The electronic device sends the labeled audio signal to a server, which then plays it. The brain responds to the coded stimulus, recording the corresponding EEG signal. It should be noted that the selection of labeled audio can serve as an encryption method for information transmission. For each pair of communicating devices, a specific audio can be selected for that communication line, while other devices use different labeled audio.

[0103] In the above-mentioned embodiments, the marked audio signal is used as the communication content, and the marked audio signal is identified and converted into corpus information, thereby improving the security of information transmission.

[0104] In one embodiment, a complete language recognition method based on electroencephalogram (EEG) signals is provided, such as... Figure 8 As shown, the above method includes:

[0105] S1 receives the labeled audio signal and obtains the EEG signal to be identified based on the labeled audio signal; the labeled audio signal is generated by converting the original language audio; the labeled audio signal is any one of frequency-modulated continuous wave (Chrip) signal, baseband signal, amplitude modulation (AM) signal, and frequency modulation (FM) signal.

[0106] S2, perform spatial filtering on the EEG signal to be identified to obtain the filtered EEG signal.

[0107] S3 extracts the Event Related Spectral Perturbation (ERSP) feature from the filtered EEG signal to obtain the feature information of the EEG signal to be identified.

[0108] S4. Input the feature information into the trained recognition model to obtain the encoded information corresponding to the EEG signal to be recognized.

[0109] S5 divides the encoded information into multiple fields.

[0110] S6 converts multiple fields according to the encoding mode to obtain multiple text symbols.

[0111] S7 combines multiple character symbols to obtain the text information corresponding to the encoded information.

[0112] S8 generates the audio signal corresponding to the EEG signal to be identified based on the text information corresponding to the encoded information.

[0113] In the aforementioned speech recognition method based on electroencephalogram (EEG) signals, features are extracted from the EEG signals to be recognized, yielding feature information. This feature information is then input into a trained recognition model to obtain the encoded information corresponding to the EEG signals. Based on this encoded information, the corresponding audio signal is determined. Since determining the encoded information of the EEG signals from their feature information is highly accurate, further conversion of the encoded information to obtain the audio signal improves the accuracy of the audio signal. Moreover, determining the EEG signals based on the labeled audio signal enhances the security of information transmission compared to directly acquiring the original speech audio.

[0114] It should be understood that although the steps in the flowcharts of the embodiments described above are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the embodiments described above may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages of other steps.

[0115] Based on the same inventive concept, this application also provides an EEG-based language recognition device for implementing the above-described EEG-based language recognition method. The solution provided by this device is similar to the implementation described in the above method; therefore, the specific limitations of one or more EEG-based language recognition device embodiments provided below can be found in the limitations of the EEG-based language recognition method described above, and will not be repeated here.

[0116] In one embodiment, such as Figure 9 As shown, a speech recognition device based on electroencephalogram (EEG) signals is provided, comprising: an extraction module 10, a recognition module 11, and a determination module 12, wherein:

[0117] The extraction module 10 is used to extract features from the EEG signal to be identified, and obtain the feature information of the EEG signal to be identified; the EEG signal to be identified is obtained based on the labeled audio signal.

[0118] The recognition module 11 is used to input feature information into the trained recognition model to obtain the encoded information corresponding to the EEG signal to be recognized.

[0119] The determination module 12 is used to determine the audio signal corresponding to the EEG signal to be identified based on the encoded information.

[0120] In one embodiment, the determining module 12 includes: a determining unit and a generating unit, wherein:

[0121] The determining unit is used to determine the text information corresponding to the encoded information based on the encoded information and the encoding mode corresponding to the encoded information;

[0122] The generation unit is used to generate the audio signal corresponding to the EEG signal to be recognized based on the text information corresponding to the encoded information.

[0123] In one embodiment, the aforementioned determining unit is specifically used to divide the encoded information to obtain multiple fields; convert the multiple fields according to the encoding mode to obtain multiple text symbols; and combine the multiple text symbols to obtain the text information corresponding to the encoded information.

[0124] In one embodiment, the extraction module 10 includes: a processing unit and an extraction unit, wherein:

[0125] The processing unit is used to perform spatial filtering on the EEG signal to be identified, so as to obtain the filtered EEG signal.

[0126] The extraction unit is used to extract Event Related Spectral Perturbation (ERSP) features from the filtered EEG signal to obtain the feature information of the EEG signal to be identified.

[0127] In one embodiment, the above-mentioned language recognition device based on EEG signals further includes: a receiving module for receiving a labeled audio signal and obtaining the EEG signal to be recognized based on the labeled audio signal; the labeled audio signal is generated by converting the original language audio; the labeled audio signal is any one of frequency-modulated continuous wave (Chrip) signal, baseband signal, amplitude modulation (AM) signal, and frequency modulation (FM) signal.

[0128] The modules in the aforementioned speech recognition device based on electroencephalogram (EEG) signals can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in the processor of a computer device in hardware form or independent of it, or stored in the memory of the computer device in software form, so that the processor can call and execute the corresponding operations of each module.

[0129] In one embodiment, a computer device is provided, which may be a server, and its internal structure diagram may be as follows: Figure 10 As shown, this computer device includes a processor, memory, input / output (I / O) interfaces, and a communication interface. The processor, memory, and I / O interfaces are connected via a system bus, and the communication interface is also connected to the system bus via the I / O interfaces. The processor provides computational and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and a database. The internal memory provides the environment for the operating system and computer programs stored in the non-volatile storage media. The database stores speech recognition data based on electroencephalogram (EEG) signals. The I / O interfaces are used for information exchange between the processor and external devices. The communication interface is used for communication with external terminals via a network connection. When the computer program is executed by the processor, it implements a speech recognition method based on EEG signals.

[0130] Those skilled in the art will understand that Figure 10 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0131] In one embodiment, a computer device is provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to perform the following steps:

[0132] Feature extraction is performed on the EEG signal to be identified to obtain its feature information; the EEG signal to be identified is obtained based on the labeled audio signal.

[0133] The feature information is input into the trained recognition model to obtain the encoded information corresponding to the EEG signal to be recognized;

[0134] Based on the encoding information, the audio signal corresponding to the EEG signal to be identified is determined.

[0135] In one embodiment, the processor, when executing a computer program, also performs the following steps:

[0136] Based on the encoding information and the encoding pattern corresponding to the encoding information, determine the text information corresponding to the encoding information;

[0137] Based on the text information corresponding to the encoded information, an audio signal corresponding to the EEG signal to be identified is generated.

[0138] In one embodiment, the processor, when executing a computer program, also performs the following steps:

[0139] The encoded information is divided into multiple fields;

[0140] Multiple fields are converted according to the encoding pattern to obtain multiple text symbols;

[0141] Multiple text symbols are combined to obtain the text information corresponding to the encoded information.

[0142] In one embodiment, the processor, when executing a computer program, also performs the following steps:

[0143] Spatial filtering is performed on the EEG signal to be identified to obtain the filtered EEG signal.

[0144] Event-related spectral perturbation (ERSP) features are extracted from the filtered EEG signals to obtain the feature information of the EEG signals to be identified.

[0145] In one embodiment, the processor, when executing a computer program, also performs the following steps:

[0146] The system receives a labeled audio signal and obtains the EEG signal to be identified based on the labeled audio signal; the labeled audio signal is generated by converting the original language audio; the labeled audio signal is any one of frequency-modulated continuous wave (Chrip) signal, baseband signal, amplitude modulation (AM) signal, and frequency modulation (FM) signal.

[0147] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon, the computer program performing the following steps when executed by a processor:

[0148] Feature extraction is performed on the EEG signal to be identified to obtain its feature information; the EEG signal to be identified is obtained based on the labeled audio signal.

[0149] The feature information is input into the trained recognition model to obtain the encoded information corresponding to the EEG signal to be recognized;

[0150] Based on the encoding information, the audio signal corresponding to the EEG signal to be identified is determined.

[0151] In one embodiment, when the computer program is executed by a processor, it also performs the following steps:

[0152] Based on the encoding information and the encoding pattern corresponding to the encoding information, determine the text information corresponding to the encoding information;

[0153] Based on the text information corresponding to the encoded information, an audio signal corresponding to the EEG signal to be identified is generated.

[0154] In one embodiment, when the computer program is executed by a processor, it also performs the following steps:

[0155] The encoded information is divided into multiple fields;

[0156] Multiple fields are converted according to the encoding pattern to obtain multiple text symbols;

[0157] Multiple text symbols are combined to obtain the text information corresponding to the encoded information.

[0158] In one embodiment, when the computer program is executed by a processor, it also performs the following steps:

[0159] Spatial filtering is performed on the EEG signal to be identified to obtain the filtered EEG signal.

[0160] Event-related spectral perturbation (ERSP) features are extracted from the filtered EEG signals to obtain the feature information of the EEG signals to be identified.

[0161] In one embodiment, when the computer program is executed by a processor, it also performs the following steps:

[0162] The system receives a labeled audio signal and obtains the EEG signal to be identified based on the labeled audio signal; the labeled audio signal is generated by converting the original language audio; the labeled audio signal is any one of frequency-modulated continuous wave (Chrip) signal, baseband signal, amplitude modulation (AM) signal, and frequency modulation (FM) signal.

[0163] In one embodiment, a computer program product is provided, including a computer program that, when executed by a processor, performs the following steps:

[0164] Feature extraction is performed on the EEG signal to be identified to obtain its feature information; the EEG signal to be identified is obtained based on the labeled audio signal.

[0165] The feature information is input into the trained recognition model to obtain the encoded information corresponding to the EEG signal to be recognized;

[0166] Based on the encoding information, the audio signal corresponding to the EEG signal to be identified is determined.

[0167] In one embodiment, when the computer program is executed by a processor, it also performs the following steps:

[0168] Based on the encoding information and the encoding pattern corresponding to the encoding information, determine the text information corresponding to the encoding information;

[0169] Based on the text information corresponding to the encoded information, an audio signal corresponding to the EEG signal to be identified is generated.

[0170] In one embodiment, when the computer program is executed by a processor, it also performs the following steps:

[0171] The encoded information is divided into multiple fields;

[0172] Multiple fields are converted according to the encoding pattern to obtain multiple text symbols;

[0173] Multiple text symbols are combined to obtain the text information corresponding to the encoded information.

[0174] In one embodiment, when the computer program is executed by a processor, it also performs the following steps:

[0175] Spatial filtering is performed on the EEG signal to be identified to obtain the filtered EEG signal.

[0176] Event-related spectral perturbation (ERSP) features are extracted from the filtered EEG signals to obtain the feature information of the EEG signals to be identified.

[0177] In one embodiment, when the computer program is executed by a processor, it also performs the following steps:

[0178] The system receives a labeled audio signal and obtains the EEG signal to be identified based on the labeled audio signal; the labeled audio signal is generated by converting the original language audio; the labeled audio signal is any one of frequency-modulated continuous wave (Chrip) signal, baseband signal, amplitude modulation (AM) signal, and frequency modulation (FM) signal.

[0179] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile memory and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM). The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, artificial intelligence (AI) processors, etc., and are not limited to these.

[0180] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this application.

[0181] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by the appended claims.

Claims

1. A language recognition method based on electroencephalogram signals, characterized in that, The method comprises: feature extraction is performed on the brain electrical signal to be identified to obtain feature information of the brain electrical signal to be identified; the brain electrical signal to be identified is obtained according to a marked audio signal; the feature information is input into a trained identification model to obtain coding information corresponding to the brain electrical signal to be identified; the coding information is divided to obtain a plurality of fields; the plurality of fields are converted according to a coding mode corresponding to the coding information to obtain a plurality of character symbols; the plurality of character symbols are combined to obtain character information corresponding to the coding information; an audio signal corresponding to the brain electrical signal to be identified is generated according to the character information corresponding to the coding information.

2. The method of claim 1, wherein, The feature extraction performed on the brain electrical signal to be identified to obtain the feature information of the brain electrical signal to be identified comprises: spatial domain filtering is performed on the brain electrical signal to be identified to obtain a filtered brain electrical signal; event-related spectral perturbation (ERSP) features are extracted from the filtered brain electrical signal to obtain the feature information of the brain electrical signal to be identified.

3. The method of claim 1, wherein, The method further comprises: The marked audio signal is received, and the brain electrical signal to be identified is obtained according to the marked audio signal; the marked audio signal is generated by converting an original language audio.

4. The method of claim 3, wherein, The marked audio signal is any one of a chrip signal, a baseband signal, an amplitude modulation signal, or a frequency modulation signal.

5. The method of claim 1, wherein, The coding information includes coding symbols and interval positions, and the coding information is divided to obtain a plurality of fields, comprising: The coding information is divided according to the interval positions to obtain the plurality of fields; each field includes at least one coding symbol.

6. The method of claim 1, wherein, The coding mode includes binary code or Morse code.

7. An electroencephalogram-based language recognition apparatus, characterized by comprising: The device comprises: an extraction module configured to perform feature extraction on a brain electrical signal to be identified to obtain feature information of the brain electrical signal to be identified; the brain electrical signal to be identified is obtained according to a marked audio signal; an identification module configured to input the feature information into a trained identification model to obtain coding information corresponding to the brain electrical signal to be identified; a determination module configured to divide the coding information to obtain a plurality of fields; convert the plurality of fields according to a coding mode corresponding to the coding information to obtain a plurality of character symbols; combine the plurality of character symbols to obtain character information corresponding to the coding information; and generate an audio signal corresponding to the brain electrical signal to be identified according to the character information corresponding to the coding information.

8. A computer device comprising a memory and a processor, the memory storing a computer program, characterized in that, The processor executes the computer program to implement the steps of the method of any one of claims 1 to 6.

9. A computer-readable storage medium having stored thereon a computer program, characterized in that, The computer program is executed by the processor to implement the steps of the method of any one of claims 1 to 6.

10. A computer program product comprising a computer program, characterized in that, The computer program is executed by the processor to implement the steps of the method of any one of claims 1 to 6.

Images