Intelligent speech recognition method and device, electronic equipment and storage medium

Abstract

The present application relates to speech processing technology, disclose a kind of intelligent speech recognition method, comprising: obtaining speech signal, the frame processing of speech signal is carried out, and the frame speech signal of obtaining is obtained;The characteristic parameter of the frame speech signal is analyzed, and the phoneme information parameter and the tone information parameter are obtained;According to the phoneme information parameter and the tone information parameter frame synchronization word recognition is carried out to the frame speech signal, and the frame word sequence of obtaining is obtained;Inquiry the starting end and the end of each sentence in the frame word sequence, the frame word sequence is split into the word sequence segment under each sentence;Word reorganization is carried out to the word sequence segment in turn using pre-constructed finite state machine, and the recognition sentence is generated.The present application also proposes a kind of intelligent speech recognition device, electronic equipment and storage medium.The present application can solve the problem that there is sequence disorder in speech recognition result.

Description

Technical Field

[0001] This invention relates to the field of speech processing technology, and in particular to an intelligent speech recognition method, device, electronic device, and storage medium. Background Technology

[0002] In most speech recognition applications, a speech acoustic model is typically used to match the input signal to obtain a set of candidate words. Then, a speech recognition model is used to find the optimal word sequence from these candidate words that meets syntactic constraints. While speech acoustic constraints and language constraints are relatively easy to implement in this method, when recognizing some continuous Chinese sentences, the recognized sentences may exhibit disordered word order, leading to incorrect speech recognition results. Summary of the Invention

[0003] This invention provides an intelligent speech recognition method, device, electronic device, and storage medium, the main purpose of which is to solve the problem of disordered word order in speech recognition results.

[0004] To achieve the above objectives, the present invention provides an intelligent speech recognition method, comprising:

[0005] Acquire the audio signal, perform frame segmentation processing on the audio signal to obtain framed audio signals;

[0006] Analyze the feature parameters of the segmented speech signal to obtain phoneme information parameters and tone information parameters;

[0007] Based on the phoneme information parameters and the tone information parameters, the frame-synchronized word recognition is performed on the frame-segmented speech signal to obtain the frame-segmented word sequence;

[0008] Query the start and end points of each statement in the framed word sequence, and split the framed word sequence into word sequence segments under each statement;

[0009] The word sequence segments are recombined sequentially using a pre-constructed finite state machine to generate a recognition sentence.

[0010] Optionally, performing framing processing on the speech signal to obtain framed speech signals includes:

[0011] Obtain the speech sampling points of the speech signal;

[0012] Based on the speech sampling points, the speech signal is subjected to overlapping segmentation processing to obtain framed speech signals.

[0013] Optionally, before performing framing processing on the speech signal, the method further includes:

[0014] The speech signal is amplified to obtain an amplified speech signal;

[0015] The amplified speech signal is filtered using a pre-constructed signal filter to obtain a filtered speech signal.

[0016] Speech endpoint detection is performed on the filtered speech signal, and the speech signals within the start endpoint and end endpoint of the endpoint detection are extracted to obtain the standard speech signal.

[0017] Optionally, the analysis of the feature parameters of the framed speech signal includes:

[0018] Obtain the fundamental frequency of the segmented speech signal and calculate the normal distribution density of the fundamental frequency to obtain the tone information parameters;

[0019] The framed speech signal is pre-emphasized, and the frequency analysis of the pre-emphasized framed speech signal is performed using a pre-constructed linear predictive coding model to obtain the frequency analysis results of the framed signal.

[0020] The frequency analysis results of the framed signal are represented spectrally using Mel scale to obtain phoneme information parameters.

[0021] Optionally, the step of performing frame-synchronized word recognition on the framed speech signal based on the phoneme information parameters and the tone information parameters to obtain a framed word sequence includes:

[0022] The OPV (One Pass Viterbi) algorithm is used to simultaneously perform word recognition on the framed speech signal using the phoneme information parameters and the tone information parameters, resulting in tone-based word recognition results.

[0023] The tonal word recognition results are concatenated using a cluster search mechanism to obtain the framed word sequence.

[0024] Optionally, the step of using a pre-constructed finite state machine to sequentially recombine the word sequence segments to generate a recognition sentence includes:

[0025] The predicted ranking score of each word in the word sequence segment is calculated sequentially using the pre-constructed finite state machine.

[0026] Based on the predicted ranking score, the words in the word sequence segment are sorted and combined to obtain the recognized statement.

[0027] Optionally, the step of using the pre-constructed finite state machine to sequentially calculate the predicted ranking score for each word in the word sequence segment includes:

[0028] The predicted ranking score for each word in the word sequence segment is calculated using the following formula:

[0029]

[0030] Where Pmax(i|j) is the predicted ranking score of the i-th word in the word sequence segment at the j-th ranking position, n is the i-th word, q is the word sequence segment, and j is the ranking position of the word in the word sequence segment. Predict the maximum sorting score for the i-th word at the j-th sorting position in the word sequence segment.

[0031] To address the above problems, the present invention also provides an intelligent voice recognition device, the device comprising:

[0032] The voice framing processing module is used to acquire voice signals, perform framing processing on the voice signals, and obtain framed voice signals.

[0033] The word sequence generation module is used to analyze the feature parameters of the framed speech signal to obtain phoneme information parameters and tone information parameters; and to perform frame-synchronized word recognition on the framed speech signal based on the phoneme information parameters and tone information parameters to obtain a framed word sequence.

[0034] The sentence word sequence segment generation module is used to query the start and end points of each sentence in the framed word sequence and split the framed word sequence into word sequence segments under each sentence.

[0035] The sentence word sequence segment recombination module is used to sequentially recombine the word sequence segments using a pre-constructed finite state machine to generate a recognized sentence.

[0036] To address the above problems, the present invention also provides an electronic device, the electronic device comprising:

[0037] At least one processor; and,

[0038] A memory communicatively connected to the at least one processor; wherein,

[0039] The memory stores a computer program that can be executed by the at least one processor, which enables the at least one processor to perform the intelligent speech recognition method described above.

[0040] To address the aforementioned problems, the present invention also provides a computer-readable storage medium storing at least one computer program, which is executed by a processor in an electronic device to implement the intelligent speech recognition method described above.

[0041] This invention performs frame-synchronized word recognition on frame-segmented speech signals based on extracted phoneme and tone information parameters to obtain frame-segmented word sequences. This ensures the continuity of information in the speech signal during speech recognition. By querying the start and end points of each sentence in the frame-segmented word sequence, the sequence can be divided into word segments based on sentences, defining the range of the segmented word sequences and ensuring the accuracy of sentence-based word order in speech recognition. Furthermore, a pre-constructed finite state machine is used to sequentially reassemble the word segments, ensuring the accuracy of word order within the corresponding sentences and thus solving the problem of disordered word order in speech recognition results. Attached Figure Description

[0042] Figure 1 This is a flowchart illustrating an intelligent speech recognition method provided in an embodiment of the present invention.

[0043] Figure 2 This is a functional block diagram of an intelligent voice recognition device provided in an embodiment of the present invention;

[0044] Figure 3 This is a schematic diagram of the structure of an electronic device that implements the intelligent speech recognition method according to an embodiment of the present invention.

[0045] The realization of the objective, functional features and advantages of the present invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0046] It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

[0047] This application provides an intelligent speech recognition method. The executing entity of the intelligent speech recognition method includes, but is not limited to, at least one of the following electronic devices that can be configured to execute the method provided in this application: a server, a terminal, etc. In other words, the intelligent speech recognition method can be executed by software or hardware installed on a terminal device or a server device, and the software can be a blockchain platform. The server includes, but is not limited to, a single server, a server cluster, a cloud server, or a cloud server cluster. The server can be an independent server or a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, content delivery networks (CDNs), and big data and artificial intelligence platforms.

[0048] Reference Figure 1The diagram shown is a flowchart of an intelligent speech recognition method provided in an embodiment of the present invention. In this embodiment, the intelligent speech recognition method includes the following steps S1-S5:

[0049] S1. Acquire the audio signal and perform frame segmentation processing on the audio signal to obtain a framed audio signal.

[0050] In this embodiment of the invention, the voice signal refers to the dialogue voice information received using a preset voice interaction system.

[0051] In this embodiment of the invention, before performing framing processing on the speech signal, it is necessary to preprocess the speech signal to enhance the speech features of the speech signal and remove noise from the speech signal.

[0052] As an embodiment of the present invention, before performing framing processing on the speech signal, the method further includes: performing signal amplification processing on the speech signal to obtain an amplified speech signal; performing signal filtering processing on the amplified speech signal using a pre-constructed signal filter to obtain a filtered speech signal; and performing speech endpoint detection on the filtered speech signal to extract the speech signals within the start endpoint and end endpoint of the endpoint detection to obtain a standard speech signal.

[0053] This invention provides a method for obtaining framed speech signals by performing frame-segmentation on the speech signal. This method can divide the speech signal into multiple stable and continuous speech signal segments, and refine the speech signal to extract speech information in more detail.

[0054] As an embodiment of the present invention, the step of performing framing processing on the speech signal to obtain a framed speech signal includes: acquiring the speech sampling points of the speech signal; and performing overlapping segmentation processing on the speech signal according to the speech sampling points to obtain a framed speech signal.

[0055] In this embodiment of the invention, the speech sampling point refers to the starting endpoint of the signal in the speech signal. Typically, the starting endpoint is separated by empty speech syllables.

[0056] By performing overlapping segmentation processing on the speech signal, this embodiment of the invention can achieve a smooth transition of the speech signal between frames after framing, thus ensuring the continuity of the speech signal after framing.

[0057] Furthermore, the process of performing overlapping segmentation on the speech signal based on the speech sampling points to obtain framed speech signals also requires the use of a pre-constructed windowing function to perform weighted calculations on the framed speech signals in order to ensure the stability of the framed speech signals.

[0058] In this embodiment of the invention, the windowing function refers to a finite window function that can perform weighted calculations on the speech signal. For example, the pre-constructed windowing function can use a rectangular window or a Hamming window.

[0059] In this embodiment of the invention, the following windowing function may be used:

[0060]

[0061] Where ω(n) is the windowing function, N is the frame length of the segmented speech signal, and n is the nth segmented speech signal.

[0062] S2. Analyze the feature parameters of the segmented speech signal to obtain phoneme information parameters and tone information parameters.

[0063] In this embodiment of the invention, the phoneme information parameter refers to the smallest phonetic unit that produces the speech of different languages. For example, in Chinese there are 22 initials (including zero initials) and 38 finals. The tone information parameter refers to the changes in tone in a language. For example, in Chinese, multiple candidate characters can be identified by combining the phoneme information parameter and the tone information parameter.

[0064] By analyzing the characteristic parameters of the framed speech signal, this embodiment of the invention can accurately obtain the constituent units of the speech signal, so as to accurately recognize the speech signal as text information.

[0065] As an embodiment of the present invention, the analysis of the characteristic parameters of the framed speech signal includes: obtaining the fundamental frequency of the framed speech signal and calculating the normal distribution density of the fundamental frequency to obtain tone information parameters; performing pre-emphasis processing on the framed speech signal and using a pre-constructed linear predictive coding model to perform frequency analysis on the pre-emphasis processed framed speech signal to obtain the frequency analysis result of the framed signal; and using Mel scale to represent the frequency analysis result of the framed signal to obtain phoneme information parameters.

[0066] S3. Perform frame-synchronized word recognition on the framed speech signal according to the phoneme information parameters and the tone information parameters to obtain the framed word sequence.

[0067] This invention enables frame-synchronized word recognition of the segmented speech signal based on the phoneme information parameters and the tone information parameters, thereby accurately identifying the target character using the basic unit of the speech signal.

[0068] In detail, the step of performing frame-synchronized word recognition on the framed speech signal based on the phoneme information parameters and the tone information parameters to obtain a framed word sequence includes: using the OPV (One Pass Viterbi) algorithm to synchronously perform word recognition on the framed speech signal using the phoneme information parameters and the tone information parameters to obtain tone-based word recognition results; and using a beam search mechanism to connect the tone-based word recognition results into words to obtain the framed word sequence.

[0069] In this embodiment of the invention, the OPV (One Pass Viterbi) algorithm refers to an algorithm that uses the phoneme information parameters and tone information parameters in the speech signal one after another to recognize words.

[0070] In this embodiment of the invention, the cluster search mechanism refers to an algorithm that establishes multiple search time steps, calculates the conditional probability of the toned word recognition results, and outputs the result with the highest conditional probability of the toned word recognition results in each search time step.

[0071] S4. Query the start and end points of each statement in the framed word sequence, and split the framed word sequence into word sequence segments of each statement.

[0072] This invention, through querying the start and end points of each sentence in a segmented word sequence, divides the segmented word sequence into word sequence segments for each sentence. This allows the successfully recognized words under each sentence to be integrated into a word set based on the sentence, thus defining the range of the segmented word sequence and ensuring the accuracy of the sentence-based word order in speech recognition.

[0073] S5. Using a pre-constructed finite state machine, the word sequence segments are sequentially recombined to generate a recognition sentence.

[0074] In this embodiment of the invention, the pre-constructed finite state machine refers to a mathematical model with discrete inputs and outputs, which is often used to process language-related tasks.

[0075] This invention utilizes a pre-constructed finite state machine to sequentially reorganize the word sequence segments, thereby ensuring the accuracy of the word order in the sentences corresponding to the speech signals.

[0076] As an embodiment of the present invention, the step of using a pre-constructed finite state machine to sequentially recombine words in the word sequence segment to generate a recognition statement includes: using the pre-constructed finite state machine to sequentially calculate the predicted ranking score of each word in the word sequence segment; and sorting and combining the words in the word sequence segment according to the predicted ranking score to obtain the recognition statement.

[0077] Further, the step of using the pre-constructed finite state machine to sequentially calculate the predicted ranking score for each word in the word sequence segment includes:

[0078] The predicted ranking score for each word in the word sequence segment is calculated using the following formula:

[0079]

[0080] Where Pmax(i|j) is the predicted ranking score of the i-th word in the word sequence segment at the j-th ranking position, n is the i-th word, q is the word sequence segment, and j is the ranking position of the word in the word sequence segment. Predict the maximum sorting score for the i-th word at the j-th sorting position in the word sequence segment.

[0081] This invention performs frame-synchronized word recognition on frame-segmented speech signals based on extracted phoneme and tone information parameters to obtain frame-segmented word sequences. This ensures the continuity of information in the speech signal during speech recognition. By querying the start and end points of each sentence in the frame-segmented word sequence, the sequence can be divided into word segments based on sentences, defining the range of the segmented word sequences and ensuring the accuracy of sentence-based word order in speech recognition. Furthermore, a pre-constructed finite state machine is used to sequentially reassemble the word segments, ensuring the accuracy of word order within the corresponding sentences and thus solving the problem of disordered word order in speech recognition results.

[0082] like Figure 2 The diagram shown is a functional block diagram of an intelligent voice recognition device provided in an embodiment of the present invention.

[0083] The intelligent speech recognition device 100 of this invention can be installed in an electronic device. Depending on the functions implemented, the intelligent speech recognition device 100 may include a speech frame processing module 101, a word sequence generation module 102, a sentence word sequence generation module 103, and a sentence word sequence segment reassembly module 104. The module described in this invention can also be called a unit, which refers to a series of computer program segments that can be executed by the processor of an electronic device and can perform a fixed function, and are stored in the memory of the electronic device.

[0084] In this embodiment, the functions of each module / unit are as follows:

[0085] The voice framing processing module 101 is used to acquire a voice signal, perform framing processing on the voice signal, and obtain a framed voice signal.

[0086] In this embodiment of the invention, the voice signal refers to the dialogue voice information received using a preset voice interaction system.

[0087] In this embodiment of the invention, before performing framing processing on the speech signal, it is necessary to preprocess the speech signal to enhance the speech features of the speech signal and remove noise from the speech signal.

[0088] As an embodiment of the present invention, before performing framing processing on the speech signal, the method further includes: performing signal amplification processing on the speech signal to obtain an amplified speech signal; performing signal filtering processing on the amplified speech signal using a pre-constructed signal filter to obtain a filtered speech signal; and performing speech endpoint detection on the filtered speech signal to extract the speech signals within the start endpoint and end endpoint of the endpoint detection to obtain a standard speech signal.

[0089] This invention provides a method for obtaining framed speech signals by performing frame-segmentation on the speech signal. This method can divide the speech signal into multiple stable and continuous speech signal segments, and refine the speech signal to extract speech information in more detail.

[0090] As an embodiment of the present invention, the step of performing framing processing on the speech signal to obtain a framed speech signal includes: acquiring the speech sampling points of the speech signal; and performing overlapping segmentation processing on the speech signal according to the speech sampling points to obtain a framed speech signal.

[0091] In this embodiment of the invention, the speech sampling point refers to the starting endpoint of the signal in the speech signal. Typically, the starting endpoint is separated by empty speech syllables.

[0092] By performing overlapping segmentation processing on the speech signal, this embodiment of the invention can achieve a smooth transition of the speech signal between frames after framing, thus ensuring the continuity of the speech signal after framing.

[0093] Furthermore, the process of performing overlapping segmentation on the speech signal based on the speech sampling points to obtain framed speech signals also requires the use of a pre-constructed windowing function to perform weighted calculations on the framed speech signals in order to ensure the stability of the framed speech signals.

[0094] In this embodiment of the invention, the windowing function refers to a finite window function that can perform weighted calculations on the speech signal. For example, the pre-constructed windowing function can use a rectangular window or a Hamming window.

[0095] In this embodiment of the invention, the following windowing function may be used:

[0096]

[0097] Where ω(n) is the windowing function, N is the frame length of the segmented speech signal, and n is the nth segmented speech signal.

[0098] The word sequence generation module 102 is used to analyze the feature parameters of the framed speech signal to obtain phoneme information parameters and tone information parameters; and to perform frame-synchronized word recognition on the framed speech signal based on the phoneme information parameters and tone information parameters to obtain a framed word sequence.

[0099] In this embodiment of the invention, the phoneme information parameter refers to the smallest phonetic unit that produces the speech of different languages. For example, in Chinese there are 22 initials (including zero initials) and 38 finals. The tone information parameter refers to the changes in tone in a language. For example, in Chinese, multiple candidate characters can be identified by combining the phoneme information parameter and the tone information parameter.

[0100] By analyzing the characteristic parameters of the framed speech signal, this embodiment of the invention can accurately obtain the constituent units of the speech signal, so as to accurately recognize the speech signal as text information.

[0101] As an embodiment of the present invention, the analysis of the characteristic parameters of the framed speech signal includes: obtaining the fundamental frequency of the framed speech signal and calculating the normal distribution density of the fundamental frequency to obtain tone information parameters; performing pre-emphasis processing on the framed speech signal and using a pre-constructed linear predictive coding model to perform frequency analysis on the pre-emphasis processed framed speech signal to obtain the frequency analysis result of the framed signal; and using Mel scale to represent the frequency analysis result of the framed signal to obtain phoneme information parameters.

[0102] This invention enables frame-synchronized word recognition of the segmented speech signal based on the phoneme information parameters and the tone information parameters, thereby accurately identifying the target character using the basic unit of the speech signal.

[0103] In detail, the step of performing frame-synchronized word recognition on the framed speech signal based on the phoneme information parameters and the tone information parameters to obtain a framed word sequence includes: using the OPV (One Pass Viterbi) algorithm to synchronously perform word recognition on the framed speech signal using the phoneme information parameters and the tone information parameters to obtain tone-based word recognition results; and using a beam search mechanism to connect the tone-based word recognition results into words to obtain the framed word sequence.

[0104] In this embodiment of the invention, the OPV (One Pass Viterbi) algorithm refers to an algorithm that uses the phoneme information parameters and tone information parameters in the speech signal one after another to recognize words.

[0105] In this embodiment of the invention, the cluster search mechanism refers to an algorithm that establishes multiple search time steps, calculates the conditional probability of the toned word recognition results, and outputs the result with the highest conditional probability of the toned word recognition results in each search time step.

[0106] The sentence word sequence segment generation module 103 is used to query the start and end points of each sentence in the framed word sequence and split the framed word sequence into word sequence segments under each sentence.

[0107] This invention, through querying the start and end points of each sentence in a segmented word sequence, divides the segmented word sequence into word sequence segments for each sentence. This allows the successfully recognized words under each sentence to be integrated into a word set based on the sentence, thus defining the range of the segmented word sequence and ensuring the accuracy of the sentence-based word order in speech recognition.

[0108] The sentence word sequence segment recombination module 104 is used to reconstruct the word sequence segments sequentially using a pre-constructed finite state machine to generate a recognition sentence.

[0109] In this embodiment of the invention, the pre-constructed finite state machine refers to a mathematical model with discrete inputs and outputs, which is often used to process language-related tasks.

[0110] This invention utilizes a pre-constructed finite state machine to sequentially reorganize the word sequence segments, thereby ensuring the accuracy of the word order in the sentences corresponding to the speech signals.

[0111] As an embodiment of the present invention, the step of using a pre-constructed finite state machine to sequentially recombine words in the word sequence segment to generate a recognition statement includes: using the pre-constructed finite state machine to sequentially calculate the predicted ranking score of each word in the word sequence segment; and sorting and combining the words in the word sequence segment according to the predicted ranking score to obtain the recognition statement.

[0112] Further, the step of using the pre-constructed finite state machine to sequentially calculate the predicted ranking score for each word in the word sequence segment includes:

[0113] The predicted ranking score for each word in the word sequence segment is calculated using the following formula:

[0114]

[0115] Where Pmax(i|j) is the predicted ranking score of the i-th word in the word sequence segment at the j-th ranking position, n is the i-th word, q is the word sequence segment, and j is the ranking position of the word in the word sequence segment. Predict the maximum sorting score for the i-th word at the j-th sorting position in the word sequence segment.

[0116] like Figure 3 The diagram shown is a structural schematic of an electronic device that implements an intelligent speech recognition method according to an embodiment of the present invention.

[0117] The electronic device 1 may include a processor 10, a memory 11, a communication bus 12 and a communication interface 13, and may also include a computer program, such as an intelligent voice recognition program, stored in the memory 11 and capable of running on the processor 10.

[0118] In some embodiments, the processor 10 may be composed of integrated circuits, such as a single packaged integrated circuit or multiple integrated circuits with the same or different functions, including combinations of one or more central processing units (CPUs), microprocessors, digital processing chips, graphics processors, and various control chips. The processor 10 is the control unit of the electronic device, connecting various components of the entire electronic device through various interfaces and lines. It executes programs or modules stored in the memory 11 (e.g., executing intelligent voice recognition programs) and calls data stored in the memory 11 to perform various functions of the electronic device and process data.

[0119] The memory 11 includes at least one type of readable storage medium, including flash memory, portable hard drive, multimedia card, card-type memory (e.g., SD or DX memory), magnetic memory, magnetic disk, optical disk, etc. In some embodiments, the memory 11 can be an internal storage unit of an electronic device, such as a portable hard drive. In other embodiments, the memory 11 can be an external storage device of the electronic device, such as a plug-in portable hard drive, Smart Media Card (SMC), Secure Digital (SD) card, Flash Card, etc. Furthermore, the memory 11 can include both internal and external storage units of the electronic device. The memory 11 can be used not only to store application software and various types of data installed on the electronic device, such as the code of an intelligent voice recognition program, but also to temporarily store data that has been output or will be output.

[0120] The communication bus 12 can be a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus, etc. This bus can be divided into an address bus, a data bus, a control bus, etc. The bus is configured to enable communication between the memory 11 and at least one processor 10, etc.

[0121] The communication interface 13 is used for communication between the aforementioned electronic device and other devices, including a network interface and a user interface. Optionally, the network interface may include a wired interface and / or a wireless interface (such as a Wi-Fi interface, Bluetooth interface, etc.), typically used to establish communication connections between the electronic device and other electronic devices. The user interface may be a display, an input unit (such as a keyboard), or, optionally, a standard wired or wireless interface. Optionally, in some embodiments, the display may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, or an OLED (Organic Light-Emitting Diode) touchscreen, etc. The display may also be appropriately referred to as a screen or display unit, used to display information processed in the electronic device and to display a visual user interface.

[0122] Figure 3 Only electronic devices with components are shown; it will be understood by those skilled in the art that... Figure 3 The structure shown does not constitute a limitation on the electronic device 1, and may include fewer or more components than shown, or combine certain components, or have different component arrangements.

[0123] For example, although not shown, the electronic device may also include a power supply (such as a battery) to power the various components. Preferably, the power supply can be logically connected to the at least one processor 10 through a power management device, thereby enabling functions such as charging management, discharging management, and power consumption management. The power supply may also include one or more DC or AC power supplies, recharging devices, power fault detection circuits, power converters or inverters, power status indicators, and other arbitrary components. The electronic device may also include various sensors, Bluetooth modules, Wi-Fi modules, etc., which will not be described in detail here.

[0124] It should be understood that the embodiments described are for illustrative purposes only and are not limited to this structure in the scope of the patent application.

[0125] The intelligent voice recognition program stored in the memory 11 of the electronic device 1 is a combination of multiple instructions, which, when run in the processor 10, can achieve the following:

[0126] Acquire the audio signal, perform frame segmentation processing on the audio signal to obtain framed audio signals;

[0127] Analyze the feature parameters of the segmented speech signal to obtain phoneme information parameters and tone information parameters;

[0128] Based on the phoneme information parameters and the tone information parameters, the frame-synchronized word recognition is performed on the frame-segmented speech signal to obtain the frame-segmented word sequence;

[0129] Query the start and end points of each statement in the framed word sequence, and split the framed word sequence into word sequence segments under each statement;

[0130] The word sequence segments are recombined sequentially using a pre-constructed finite state machine to generate a recognition sentence.

[0131] Specifically, the specific implementation method of the processor 10 for the above instructions can be referred to the description of the relevant steps in the corresponding embodiment of the accompanying drawings, and will not be repeated here.

[0132] Furthermore, if the modules / units integrated in the electronic device 1 are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. The computer-readable storage medium can be volatile or non-volatile. For example, the computer-readable medium may include: any entity or device capable of carrying the computer program code, a recording medium, a USB flash drive, a portable hard drive, a magnetic disk, an optical disk, a computer memory, or a read-only memory (ROM).

[0133] The present invention also provides a computer-readable storage medium storing a computer program, which, when executed by a processor of an electronic device, can perform the following:

[0134] Acquire the audio signal, perform frame segmentation processing on the audio signal to obtain framed audio signals;

[0135] Analyze the feature parameters of the segmented speech signal to obtain phoneme information parameters and tone information parameters;

[0136] Based on the phoneme information parameters and the tone information parameters, the frame-synchronized word recognition is performed on the frame-segmented speech signal to obtain the frame-segmented word sequence;

[0137] Query the start and end points of each statement in the framed word sequence, and split the framed word sequence into word sequence segments under each statement;

[0138] The word sequence segments are recombined sequentially using a pre-constructed finite state machine to generate a recognition sentence.

[0139] In the several embodiments provided by this invention, it should be understood that the disclosed devices, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of modules is only a logical functional division, and other division methods may be used in actual implementation.

[0140] The modules described as separate components may or may not be physically separate. The components shown as modules may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0141] Furthermore, the functional modules in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or in the form of hardware plus software functional modules.

[0142] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0143] Therefore, the embodiments should be considered exemplary and non-limiting in all respects, and the scope of the invention is defined by the appended claims rather than the foregoing description. Thus, all variations falling within the meaning and scope of equivalents of the claims are intended to be embraced within the invention. No appended diagram markings in the claims should be construed as limiting the scope of the claims.

[0144] The blockchain referred to in this invention is a novel application model of computer technologies such as distributed data storage, peer-to-peer transmission, consensus mechanisms, and encryption algorithms. Essentially, a blockchain is a decentralized database, a chain of data blocks linked together using cryptographic methods. Each data block contains information about a batch of network transactions, used to verify the validity of the information (anti-counterfeiting) and generate the next block. A blockchain can include an underlying blockchain platform, a platform product service layer, and an application service layer.

[0145] The embodiments of this application can acquire and process relevant data based on artificial intelligence technology. Artificial intelligence (AI) refers to the theories, methods, technologies, and application systems that use digital computers or machines controlled by digital computers to simulate, extend, and expand human intelligence, perceive the environment, acquire knowledge, and use that knowledge to obtain optimal results.

[0146] Furthermore, it is clear that the word "comprising" does not exclude other units or steps, and the singular does not exclude the plural. Multiple units or devices recited in a system claim may also be implemented by a single unit or device through software or hardware. The terms "first," "second," etc., are used to indicate names and do not indicate any specific order.

[0147] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. An intelligent speech recognition method, characterized in that, The method includes: Acquire the audio signal, perform frame segmentation processing on the audio signal to obtain framed audio signals; Analyze the feature parameters of the segmented speech signal to obtain phoneme information parameters and tone information parameters; The OPV algorithm is used to synchronously perform word recognition on the framed speech signal using the phoneme information parameters and the tone information parameters to obtain the tone-based word recognition results. The beam search mechanism is then used to connect the tone-based word recognition results to obtain the framed word sequence. Query the start and end points of each statement in the framed word sequence, and split the framed word sequence into word sequence segments under each statement; The predicted ranking score of each word in the word sequence segment is calculated sequentially using a pre-constructed finite state machine. The words in the word sequence segment are then sorted and combined according to the predicted ranking scores to generate a recognition statement.

2. The intelligent speech recognition method as described in claim 1, characterized in that, The step of performing framing processing on the speech signal to obtain framed speech signals includes: Obtain the speech sampling points of the speech signal; Based on the speech sampling points, the speech signal is subjected to overlapping segmentation processing to obtain framed speech signals.

3. The intelligent speech recognition method as described in claim 1, characterized in that, Before performing framing processing on the speech signal, the method further includes: The speech signal is amplified to obtain an amplified speech signal; The amplified speech signal is filtered using a pre-constructed signal filter to obtain a filtered speech signal. Speech endpoint detection is performed on the filtered speech signal, and the speech signals within the start endpoint and end endpoint of the endpoint detection are extracted to obtain the standard speech signal.

4. The intelligent speech recognition method as described in claim 1, characterized in that, The analysis of the characteristic parameters of the framed speech signal includes: Obtain the fundamental frequency of the segmented speech signal and calculate the normal distribution density of the fundamental frequency to obtain the tone information parameters; The framed speech signal is pre-emphasized, and the frequency analysis of the pre-emphasized framed speech signal is performed using a pre-constructed linear predictive coding model to obtain the frequency analysis results of the framed signal. The frequency analysis results of the framed signal are represented spectrally using Mel scale to obtain phoneme information parameters.

5. The intelligent speech recognition method as described in claim 1, characterized in that, The step of using the pre-constructed finite state machine to sequentially calculate the predicted ranking score for each word in the word sequence segment includes: The predicted ranking score for each word in the word sequence segment is calculated using the following formula: ； in, For the word sequence segment, the first The word in the first Predicted sort score at each sort position For the first One word, The word sequence segment, The word's sorting position within the word sequence segment. For the first The word is the th word in the word sequence segment. The value that predicts the largest sorting score at each sorting position.

6. An intelligent voice recognition device, characterized in that, The device includes: The voice framing processing module is used to acquire voice signals, perform framing processing on the voice signals, and obtain framed voice signals. The word sequence generation module is used to analyze the feature parameters of the framed speech signal to obtain phoneme information parameters and tone information parameters; the OPV algorithm is used to synchronously perform word recognition on the framed speech signal using the phoneme information parameters and tone information parameters to obtain tone-based word recognition results; and the beam search mechanism is used to connect the tone-based word recognition results to obtain the framed word sequence. The sentence word sequence segment generation module is used to query the start and end points of each sentence in the framed word sequence and split the framed word sequence into word sequence segments under each sentence. The sentence word sequence segment recombination module is used to calculate the predicted ranking score of each word in the word sequence segment sequentially using a pre-constructed finite state machine, and to sort and combine the words in the word sequence segment according to the predicted ranking score to generate a recognized sentence.

7. An electronic device, characterized in that, The electronic device includes: At least one processor; and, A memory communicatively connected to the at least one processor; wherein, The memory stores a computer program that can be executed by the at least one processor, the computer program being executed by the at least one processor to enable the at least one processor to perform the intelligent speech recognition method as described in any one of claims 1 to 5.

8. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the intelligent speech recognition method as described in any one of claims 1 to 5.

Images