Model training and detection method, communication module, device, equipment and storage medium

Abstract

The present application relates to a model training and detection method, communication module, device, equipment and storage medium, wherein the method comprises: obtaining multiple sets of training data of dual-tone multi-frequency signals; wherein each set of the training data comprises: the number of tones of the dual-tone multi-frequency signals to be detected, the detection window size and the bridge length; wherein the bridge length is used to indicate the number of consecutive no tones at the end of the dual-tone multi-frequency signal detection; obtaining a multivariate function model; wherein the multivariate function model is used to fit the relationship between the number of tones, the detection window size, the bridge length and the detection success rate of the dual-tone multi-frequency signal; the multivariate function model is trained using multiple sets of training data of the dual-tone multi-frequency signals to obtain the detection model of the dual-tone multi-frequency signal. The present application is used to solve the technical problem of low efficiency in the method of manually adjusting parameters.

Description

Technical Field

[0001] The present application relates to the field of speech recognition technology, and in particular to a model training and detection method, a communication module, a device, an equipment and a storage medium. Background Art

[0002] DTMF (Double Tone Multi Frequency) is a technology that can achieve fast and reliable transmission of telephone numbers. It has strong anti-interference ability and high transmission speed. Therefore, it can be widely used in telephone communication systems. DTMF is a coding technology that uses two specific single-tone combination signals to represent digital signals to achieve its functions. The frequencies of the two single tones are different, and the numbers represented or the functions achieved are also different. This type of telephone usually has 16 keys for audio dialing, including: 10 numeric keys 0 to 9 and 6 function keys *, #, A, B, C, D. Most of them are used for audio dialing. In fact, it can also be widely used in data communication systems to achieve remote transmission of various data streams and voice information.

[0003] At present, mobile communication modules are increasingly used in computers, and the module communication call mechanism is also indispensable. In many application scenarios, the call mechanism is not only used as a voice call. For example: in banks, shopping malls and many other scenarios, mobile communication modules are more often used as emergency alarms, and the transmission of alarm information is often done by transmitting special DTMF tones after the call communication is established. For example: banks may have ordinary civil cases or more serious criminal cases, and these two types can be distinguished and communicated through * and #, but DTMF tones are transmitted by dual-frequency tones, and false alarms or missed alarms may occur during the identification and reporting process.

[0004] In the prior art, when a problem occurs after use by a customer, the problem is fed back to the manufacturer, and then the technical staff manually adjusts the parameters, and the technology manufacturer performs internal debugging. During internal debugging, technical staff are required to debug and adjust the parameters according to the customer's usage environment and demand environment, and verify the effect. After it meets the requirements, the tester assists in stress testing, checks the effect, and then makes repeated adjustments to obtain a more suitable product status at this stage.

[0005] In the existing technology, most of them rely on the experience of technical personnel to make repeated adjustments, and the cycle is long. In order to ensure functional stability, repeated stress testing and parameter fine-tuning are required to correct the effect, and there is no guarantee that the most suitable parameter combination is obtained, so that after the subsequent return to the customer, there is still a certain probability that the problem will recur, resulting in low adjustment efficiency. Summary of the invention

[0006] The present application provides a model training and detection method, a communication module, an apparatus, a device and a storage medium to solve the technical problem of low efficiency of manual parameter adjustment in the prior art.

[0007] In a first aspect, an embodiment of the present application provides a method for training a detection model of a dual-tone multi-frequency signal, comprising:

[0008] Acquire multiple sets of dual-tone multi-frequency signal training data; wherein each set of the training data includes: the number of tones of the dual-tone multi-frequency signal to be detected, the detection window size and the bridge length; wherein the bridge length is used to indicate the number of consecutive no tones when the dual-tone multi-frequency signal detection ends;

[0009] Acquire a multivariate function model; wherein the multivariate function model is used to fit the relationship between the number of tones, the detection window size, the bridge length and the detection success rate of the dual-tone multi-frequency signal;

[0010] The multivariate function model is trained using a plurality of sets of training data of the dual-tone multi-frequency signal to obtain a detection model of the dual-tone multi-frequency signal.

[0011] Optionally, the multivariate function model is trained using multiple groups of training data of the dual-tone multi-frequency signal to obtain a detection model of the dual-tone multi-frequency signal, including:

[0012] Dividing the training data of multiple groups of dual-tone multi-frequency signals into multiple training data sets;

[0013] The following training process is performed for the training data of each dual-tone multi-frequency signal in each training data set:

[0014] For each training data of the dual-tone multi-frequency signal, the number of tones, the detection window size and the bridge length are input into the multivariate function model to obtain a detection success rate; and the difference between the detection success rate and a preset expected success rate is calculated;

[0015] Calculate an evaluation value using a preset evaluation algorithm and each of the differences; wherein the evaluation value is used to evaluate the training effect at a certain stage;

[0016] Calculate the difference between the evaluation value and the preset evaluation value to obtain a deviation value;

[0017] Determining whether the deviation value is greater than or equal to a preset threshold;

[0018] If yes, adjusting the parameters of the multivariate function model according to the deviation value;

[0019] If not, a detection model of the dual-tone multi-frequency signal is obtained.

[0020] Optionally, adjusting the parameters of the multivariate function model according to the deviation value includes:

[0021] If it is determined that the difference between the deviation value and the preset threshold value is greater than or equal to the preset threshold value, adjusting the weight of the parameter corresponding to the number of tones in the multivariate function model to be greater than the weight of each of the detection window size and the bridge length;

[0022] If it is determined that the difference between the deviation value and the preset threshold is less than the preset threshold value, the proportion of the parameter corresponding to the number of tones in the multivariate function model is adjusted to be less than the respective proportions of the detection window size and the bridge length.

[0023] Optionally, the acquiring of multiple sets of dual-tone multi-frequency signal training data includes:

[0024] Get multiple audio clips;

[0025] The following processing is performed for each of the audio clips:

[0026] Detecting a dual-tone multi-frequency signal from the audio clip, and performing de-jittering and de-duplication processing to obtain a processed dual-tone multi-frequency signal;

[0027] The number of tones, the detection window and the bridge length are detected from the processed dual tone multi-frequency signal as a set of training data of the dual tone multi-frequency signal.

[0028] In a second aspect, an embodiment of the present application provides a method for detecting a dual-tone multi-frequency signal, comprising:

[0029] Obtain a trained dual-tone multi-frequency signal detection model; wherein the detection model is trained based on the training method of the dual-tone multi-frequency signal detection model described in the first aspect;

[0030] Writing the trained dual-tone multi-frequency signal detection model into the audio calibration parameter file to obtain a modified audio calibration parameter file;

[0031] Reloading the modified audio calibration parameter file to make the trained dual-tone multi-frequency signal detection model effective;

[0032] The dual-tone multi-frequency signal is detected by using the effective dual-tone multi-frequency signal detection model, and the detected dual-tone multi-frequency signal is reported.

[0033] In a third aspect, an embodiment of the present application provides a communication module, which is used to execute the dual-tone multi-frequency signal detection method described in the second aspect.

[0034] In a fourth aspect, an embodiment of the present application provides a training device for a dual-tone multi-frequency signal detection model, comprising:

[0035] The first acquisition module is used to acquire multiple sets of dual-tone multi-frequency signal training data; wherein each set of the training data includes: the number of tones of the dual-tone multi-frequency signal to be detected, the detection window size and the bridge length; wherein the bridge length is used to indicate the number of consecutive no tones when the dual-tone multi-frequency signal detection ends;

[0036] A second acquisition module is used to acquire a multivariate function model; wherein the multivariate function model is used to fit the relationship between the number of tones, the detection window size, the bridge length and the detection success rate of the dual-tone multi-frequency signal;

[0037] The training module is used to train the multivariate function model using a plurality of sets of training data of the dual-tone multi-frequency signal to obtain a detection model of the dual-tone multi-frequency signal.

[0038] In a fifth aspect, an embodiment of the present application provides a dual-tone multi-frequency signal detection device, comprising:

[0039] A third acquisition module is used to acquire a trained dual-tone multi-frequency signal detection model; wherein the detection model is trained based on the training method of the dual-tone multi-frequency signal detection model described in the first aspect;

[0040] A writing module is used to write the trained dual-tone multi-frequency signal detection model into an audio calibration parameter file to obtain a modified audio calibration parameter file;

[0041] A reload module, used for reloading the modified audio calibration parameter file to make the trained dual-tone multi-frequency signal detection model effective;

[0042] The detection module is used to detect the dual-tone multi-frequency signal by using the effective dual-tone multi-frequency signal detection model, and report the detected dual-tone multi-frequency signal.

[0043] In a sixth aspect, an embodiment of the present application provides an electronic device, comprising: a processor, a memory, and a communication bus, wherein the processor and the memory communicate with each other via the communication bus;

[0044] The memory is used to store computer programs;

[0045] The processor is used to execute the program stored in the memory to implement the dual-tone multi-frequency signal detection model training method described in the first aspect or the dual-tone multi-frequency signal detection method described in the second aspect.

[0046] In a seventh aspect, an embodiment of the present application provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the training method for the dual-tone multi-frequency signal detection model described in the first aspect or the dual-tone multi-frequency signal detection method described in the second aspect.

[0047] The above-mentioned technical scheme provided by the embodiment of the present application has the following advantages compared with the prior art: the method provided by the embodiment of the present application adopts the method of machine learning, utilizes the training data of multiple groups of dual-tone multi-frequency signals to train the multivariate function model, obtains the detection model of dual-tone multi-frequency signals, does not need to manually adjust parameters according to experience, greatly reduces the low efficiency of manual adjustment process and test process in research and development, and improves the parameter adjustment efficiency. Moreover, the multivariate function model is trained by a large amount of training data, and the detection model obtained has high accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the invention and, together with the description, serve to explain the principles of the invention.

[0049] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0050] Figure 1 A schematic flow chart of a method for training a dual-tone multi-frequency signal detection model provided in an embodiment of the present application;

[0051] Figure 2 An overall flow chart of a training method for a dual-tone multi-frequency signal detection model provided in an embodiment of the present application;

[0052] Figure 3 A schematic diagram of the process of modifying and taking effect the model parameters provided in the embodiment of the present application;

[0053] Figure 4 A schematic flow chart of a dual-tone multi-frequency signal detection method provided in an embodiment of the present application;

[0054] Figure 5 A schematic diagram of the structure of a dual-tone multi-frequency signal detection model training device provided in an embodiment of the present application;

[0055] Figure 6 A schematic diagram of the structure of a dual-tone multi-frequency signal detection device provided in an embodiment of the present application;

[0056] Figure 7A schematic diagram of the structure of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0057] In order to make the purpose, technical solution and advantages of the embodiments of the present application clearer, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of this application.

[0058] The method provided in the embodiment of the present application can be applied to an electronic device, which can be a module capable of realizing a communication function or a terminal device including the module, etc. The terminal device can be a mobile terminal or a smart terminal. The mobile terminal can be at least one of a mobile phone, a tablet computer, a laptop computer, etc.; the smart terminal can be a terminal including a wireless communication module such as a smart car, a smart watch, a shared bicycle, a smart cabinet, etc.; the module can be a wireless communication module, such as any one of a 2G communication module, a 3G communication module, a 4G communication module, a 5G communication module, and a NB-IOT communication module.

[0059] like Figure 1 As shown, the embodiment of the present application provides a training method for a detection model of a dual-tone multi-frequency signal, which specifically includes the following steps:

[0060] Step 101, obtaining multiple sets of dual tone multi-frequency signal training data; wherein each set of training data includes: the number of tones of the dual tone multi-frequency signal to be detected, the detection window size and the bridge length; wherein the bridge length is used to indicate the number of consecutive no tones when the dual tone multi-frequency signal detection ends;

[0061] In a specific implementation, multiple sets of dual-tone multi-frequency signal training data can be obtained in the following manner: first, multiple audio clips are obtained; each audio clip is processed as follows: the dual-tone multi-frequency signal is detected from the audio clip, and de-jitter and de-duplicate processing is performed to obtain the processed dual-tone multi-frequency signal; from the processed dual-tone multi-frequency signal, the number of tones, the detection window and the bridge length are detected as a set of dual-tone multi-frequency signal training data. Among them, the dual-tone multi-frequency signal can be detected using the Goertzel algorithm.

[0062] The number of tones refers to the number of tones required to trigger the detection of dual-tone multi-frequency signals, that is, how many tones are there in the detection window to consider that a DTMF tone is detected. There will be no tone segments or interference-free tone segments that do not meet the standards in the detection window, and the number of matching tones is identified in the detection window. The bridge length is used to determine the number of consecutive no-tone indications required to end the detection. After a tone is detected, after how many consecutive no-tones there are, the tone detection can be considered complete, and the detected tone and all the last no-tone segments are connected to form a complete DTMF tone segment. Duplicate tones detected during the detection process are deleted.

[0063] In addition, it should be noted that the training data can be obtained by putting it into a specific use environment and actually occurring, or it can be obtained based on the dual-tone multi-frequency signals that occurred in history. In other words, the training process can be training separately from the actual use environment, training first and then putting it into the actual use environment; or it can be applied in the actual use environment and continuously trained, that is, putting it into the use environment first, and then using the reported DTMF signals that occurred in actual use for training.

[0064] Step 102, obtaining a multivariate function model; wherein the multivariate function model is used to fit the relationship between the number of tones, the detection window size, the bridge length and the detection success rate of the dual-tone multi-frequency signal;

[0065] Among them, the multivariate function model can adopt a ternary linear regression model, with the number of tones, detection window size and bridge length as input parameters and the detection success rate as output.

[0066] Step 103: training the multivariate function model using multiple sets of dual tone multi-frequency signal training data to obtain a dual tone multi-frequency signal detection model.

[0067] During training, you can evaluate each preset number of training data sets, and then adjust the multivariate function model parameters to make the trained model more accurate. For example, you can evaluate once every 1,000 trainings, and adjust the model parameters based on the evaluation results. Specifically, you can train the model as follows:

[0068] First, according to a preset number, the training data of multiple groups of dual-tone multi-frequency signals are divided into multiple training data sets; the training data of each dual-tone multi-frequency signal in each training data set is subjected to the following training process:

[0069] For each dual-tone multi-frequency signal training data, the number of tones, the detection window size and the bridge length are input into a multivariate function model to obtain a detection success rate; the difference between the detection success rate and the preset expected success rate is calculated; the evaluation value is calculated using a preset evaluation algorithm and each difference; wherein the evaluation value is used to evaluate the staged training effect; the difference between the evaluation value and the preset evaluation value is calculated to obtain a deviation value; whether the deviation value is greater than or equal to a preset threshold is determined; if yes, the parameters of the multivariate function model are adjusted according to the deviation value, and the training is continued using the next training set after the adjustment; if no, the training is stopped to obtain a dual-tone multi-frequency signal detection model.

[0070] The preset expected success rate can be set according to the actual situation, or determined according to the acceptable false alarm omission ratio provided by the user. For example, the preset expected success rate can be set to 95%; for another example, if the user's acceptable false alarm omission ratio is 3%, the preset expected success rate can be set to 97%.

[0071] The preset evaluation algorithm can be selected according to actual needs, for example, the least square method can be used to calculate the evaluation value, and other evaluation algorithms can also be used, such as calculating the sum of square differences. The preset evaluation value can also be set according to actual needs.

[0072] If the difference between the evaluation value and the preset evaluation value, that is, the deviation value, is greater than the preset threshold, it means that the currently trained model parameters need to be further adjusted; otherwise, it is considered that no further adjustment is required, that is, a trained dual-tone multi-frequency signal detection model is obtained.

[0073] In a specific embodiment, adjusting the parameters of the multivariate function model according to the deviation value includes:

[0074] If it is determined that the difference between the deviation value and the preset threshold value is greater than or equal to the preset threshold value, the weight of the parameter corresponding to the number of tones in the multivariate function model is adjusted to be greater than the weight of the detection window size and the bridge length respectively;

[0075] If the difference between the deviation value and the preset threshold is determined to be smaller than the preset threshold value, the weight of the parameter corresponding to the number of tones in the multivariate function model is adjusted to be smaller than the weights of the detection window size and the bridge length.

[0076] like Figure 2As shown, during the training and learning process, if the deviation value is relatively large, a large strategy can be used to adjust the parameters. In practical applications, the parameters corresponding to the number of tones have a greater impact on the output results of the model. When adjusting the parameters according to the large strategy, the proportion of the parameters corresponding to the number of tones can be adjusted to be greater than the proportion of the detection window size and the bridge length respectively; if the deviation value is relatively small, a small strategy can be used to adjust the parameters. In practical applications, the parameters corresponding to the detection window size and the bridge length have a smaller impact on the output results of the model. When adjusting the parameters according to the small strategy, the proportion of the parameters corresponding to the number of tones in the multivariate function model can be adjusted to be less than the proportion of the detection window size and the bridge length respectively. If the deviation value is less than the preset threshold, the learning is stopped.

[0077] In the present application embodiment, adopt the method for machine learning, utilize the training data of multiple groups of dual-tone multi-frequency signals to train the multivariate function model, obtain the detection model of dual-tone multi-frequency signals, do not need to manually adjust parameters according to experience, greatly reduce the low efficiency of internal manual adjustment process and test process of research and development, improve parameter adjustment efficiency. And, by a large amount of training data, the multivariate function model is trained, and the detection model obtained has high accuracy.

[0078] In a specific embodiment, in order to better adapt to the usage environment of various industries, multiple groups of dual-tone multi-frequency signal training data are classified according to industry, and the training data of dual-tone multi-frequency signals corresponding to different industries are obtained, and training is performed, so that the obtained dual-tone multi-frequency signal detection model is more suitable for different usage scenarios, thereby improving the adaptability of the product to the market and improving the satisfaction of customers in different industries.

[0079] In order to facilitate understanding of the technical solutions provided in the embodiments of the present application, Figure 2 A training method for a dual-tone multi-frequency signal detection model provided in an embodiment of the present application is further described.

[0080] 1. Model creation: The three parameters of the number of tones, detection window size and bridge length mentioned above are used as input parameters to build a multivariate function model. The evaluation system is built according to the false alarm and missed alarm ratio. The system is briefly put into the customer use environment to obtain initial use data, build a relatively fitting preliminary function, load the periodic missed alarm ratio as the basis for evaluation, and formulate two parameter adjustment strategies to deal with deviation scenarios with large or small deviation values.

[0081] 2. Model learning: After the model is embedded in the system, it is handed over to the customer manufacturer to train the model so that the model can be fully tested in the usage scenario. The evaluated threshold and parameter adjustment strategy are loaded together with the model. During use, the deviation value is calculated through the evaluation algorithm (such as the least squares method, etc.), and the parameters are automatically adjusted according to the adjustment strategy to continue running.

[0082] Among them, parameter adaptation parameter modification can be mainly divided into two strategies, large deviation value strategy and small deviation value strategy. When the deviation value is large, it means that the customer's usage environment has changed greatly over a period of time, and a large adjustment is required. The large deviation strategy adjustment is performed on each parameter. When the deviation value is small, it means that the user's usage scenario has changed little from the previous one during this period, but because it exceeds the preset threshold, it also needs to be adjusted, and the small strategy model can be applied for fine-tuning.

[0083] like Figure 3 As shown, after the parameters are modified according to the strategy, the corresponding parameters in the ACDB audio calibration parameter file are modified through the Qualcomm interface. After the modification, the program responsible for loading the ACDB file is restarted to reload the parameter file so that the parameters take effect to detect the dual-tone multi-frequency signal.

[0084] In addition, if Figure 4 As shown, the embodiment of the present application provides a method for detecting a dual-tone multi-frequency signal, which specifically includes the following steps:

[0085] Step 401, obtaining a trained detection model of a dual-tone multi-frequency signal; wherein the detection model is trained based on the training method of the detection model of the dual-tone multi-frequency signal;

[0086] Step 402, writing the trained dual-tone multi-frequency signal detection model into the audio calibration parameter file to obtain a modified audio calibration parameter file;

[0087] Step 403, reloading the modified audio calibration parameter file to make the trained dual-tone multi-frequency signal detection model effective;

[0088] Step 404: Detect the dual-tone multi-frequency signal using the effective dual-tone multi-frequency signal detection model, and report the detected dual-tone multi-frequency signal.

[0089] In the specific implementation, the corresponding parameters in the ACDB audio calibration parameter file can be modified through the data interface according to the parameters of the trained detection model. After the modification, the program responsible for loading the ACDB audio calibration file is restarted to reload the ACDB audio calibration file so that the parameters take effect, so as to detect the dual-tone multi-frequency signal and report it.

[0090] In addition, an embodiment of the present application further provides a communication module, which is used to execute the above-mentioned dual-tone multi-frequency signal detection method.

[0091] Among them, in a specific implementation, the communication module can be used in scenes such as banks and shopping malls to transmit alarm information. Of course, it can also be used in other scenes to realize dual-tone multi-frequency signal detection. In a specific implementation, the communication module can be integrated in an electronic device or externally placed in an electronic device.

[0092] Based on the same concept, a training device for a detection model of a dual-tone multi-frequency signal is provided in the embodiment of the present application. The specific implementation of the device can be found in the description of the method embodiment part, and the repetitions will not be repeated. Figure 5 As shown, the device mainly includes:

[0093] The first acquisition module 501 is used to acquire multiple sets of dual-tone multi-frequency signal training data; wherein each set of the training data includes: the number of tones of the dual-tone multi-frequency signal to be detected, the detection window size and the bridge length; wherein the bridge length is used to indicate the number of consecutive no tones when the dual-tone multi-frequency signal detection ends;

[0094] The second acquisition module 502 is used to obtain a multivariate function model; wherein the multivariate function model is used to fit the relationship between the number of tones, the detection window size, the bridge length and the detection success rate of the dual-tone multi-frequency signal;

[0095] The training module 503 is used to train the multivariate function model using multiple sets of training data of the dual-tone multi-frequency signal to obtain a detection model of the dual-tone multi-frequency signal.

[0096] In a specific embodiment, the training module 503 is used to divide the training data of multiple groups of dual-tone multi-frequency signals into multiple training data sets; the training data of each dual-tone multi-frequency signal in each training data set is subjected to the following training process: for each training data of the dual-tone multi-frequency signal, the number of tones, the detection window size and the bridge length are input into the multivariate function model to obtain a detection success rate; the difference between the detection success rate and a preset expected success rate is calculated; an evaluation value is calculated using a preset evaluation algorithm and each of the differences; wherein the evaluation value is used to evaluate a staged training effect; the difference between the evaluation value and the preset evaluation value is calculated to obtain a deviation value; it is determined whether the deviation value is greater than or equal to a preset threshold value; if so, the parameters of the multivariate function model are adjusted according to the deviation value; if not, a detection model of the dual-tone multi-frequency signal is obtained.

[0097] In a specific embodiment, the training module 503 is used to adjust the proportion of the parameter corresponding to the number of tones in the multivariate function model to be greater than the respective proportions of the detection window size and the bridge length when it is determined that the difference between the deviation value and the preset threshold is greater than or equal to the preset threshold value; and to adjust the proportion of the parameter corresponding to the number of tones in the multivariate function model to be less than the respective proportions of the detection window size and the bridge length when it is determined that the difference between the deviation value and the preset threshold is less than the preset threshold value.

[0098] In a specific embodiment, the first acquisition module 501 is used to acquire multiple audio clips; perform the following processing on each of the audio clips: detect a dual-tone multi-frequency signal from the audio clip, and perform de-jitter and de-duplication processing to obtain a processed dual-tone multi-frequency signal; and detect the number of tones, the detection window and the bridge length from the processed dual-tone multi-frequency signal as a group of dual-tone multi-frequency signal training data.

[0099] Based on the same concept, a dual-tone multi-frequency signal detection device is provided in the embodiment of the present application. The specific implementation of the device can refer to the description of the method embodiment part, and the repetition will not be repeated. Figure 6 As shown, the device mainly includes:

[0100] The third acquisition module 601 is used to acquire a trained detection model of a dual-tone multi-frequency signal; wherein the detection model is trained based on the training method of the detection model of the dual-tone multi-frequency signal;

[0101] A writing module 602 is used to write the trained dual-tone multi-frequency signal detection model into an audio calibration parameter file to obtain a modified audio calibration parameter file;

[0102] The reload module 603 is used to reload the modified audio calibration parameter file to make the trained dual-tone multi-frequency signal detection model effective;

[0103] The detection module 604 is used to detect the dual-tone multi-frequency signal by using the effective dual-tone multi-frequency signal detection model, and report the detected dual-tone multi-frequency signal.

[0104] In addition, an embodiment of the present application further provides a communication module, which is used to execute the above-mentioned training method of the dual-tone multi-frequency signal detection model, or the above-mentioned dual-tone multi-frequency signal detection method.

[0105] Based on the same concept, an electronic device is also provided in the embodiment of the present application, such as Figure 7 As shown, the electronic device mainly includes: a processor 701, a memory 702 and a communication bus 703, wherein the processor 701 and the memory 702 communicate with each other through the communication bus 703. The memory 702 stores a program that can be executed by the processor 701, and the processor 701 executes the program stored in the memory 702 to implement the following steps:

[0106] Acquire multiple sets of dual-tone multi-frequency signal training data; wherein each set of the training data includes: the number of tones of the dual-tone multi-frequency signal to be detected, the detection window size and the bridge length; wherein the bridge length is used to indicate the number of consecutive no tones when the dual-tone multi-frequency signal detection ends;

[0107] Acquire a multivariate function model; wherein the multivariate function model is used to fit the relationship between the number of tones, the detection window size, the bridge length and the detection success rate of the dual-tone multi-frequency signal;

[0108] Utilize multiple groups of training data of dual tone multi-frequency signals to train the multivariate function model to obtain a detection model of the dual tone multi-frequency signals;

[0109] or,

[0110] Obtain a trained dual-tone multi-frequency signal detection model; wherein the detection model is trained based on the training method of the dual-tone multi-frequency signal detection model;

[0111] Writing the trained dual-tone multi-frequency signal detection model into the audio calibration parameter file to obtain a modified audio calibration parameter file;

[0112] Reloading the modified audio calibration parameter file to make the trained dual-tone multi-frequency signal detection model effective;

[0113] The dual-tone multi-frequency signal is detected by using the effective dual-tone multi-frequency signal detection model, and the detected dual-tone multi-frequency signal is reported.

[0114] The communication bus 703 mentioned in the above electronic device can be a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus. The communication bus 703 can be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 7 Only one thick line is used in the diagram, but this does not mean that there is only one bus or only one type of bus.

[0115] The memory 702 may include a random access memory (RAM) or a non-volatile memory, such as at least one disk memory. Alternatively, the memory may also be at least one storage device located away from the processor 701.

[0116] The above-mentioned processor 701 can be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc., and can also be a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field programmable gate array (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components.

[0117] In another embodiment of the present application, a computer-readable storage medium is provided, in which a computer program is stored. When the computer program is run on a computer, the computer executes a training method for a dual-tone multi-frequency signal detection model described in the above embodiment, or a dual-tone multi-frequency signal detection method.

[0118] In the above embodiments, it can be implemented in whole or in part by software, hardware, firmware or any combination thereof. When implemented using software, it can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer instruction is loaded and executed on a computer, the process or function described in the embodiment of the present application is generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network or other programmable device. The computer instruction can be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium, for example, the computer instruction is transmitted from a website site, a computer, a server or a data center by wired (e.g., coaxial cable, optical fiber, digital subscriber line (DSL)) or wireless (e.g., infrared, microwave, etc.) mode to another website site, computer, server or data center. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server, a data center, etc. that contains one or more available media integration. The available medium can be a magnetic medium (e.g., a floppy disk, a hard disk, a tape, etc.), an optical medium (e.g., a DVD) or a semiconductor medium (e.g., a solid-state hard disk), etc.

[0119] It should be noted that, in this article, relational terms such as "first" and "second" are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprise a ..." do not exclude the existence of other identical elements in the process, method, article or device including the elements.

[0120] The foregoing is merely a specific embodiment of the present invention, which enables those skilled in the art to understand or implement the present invention. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but rather to the widest scope consistent with the principles and novel features claimed herein.

Claims

1. A training method for a detection model of a dual-tone multi-frequency signal, characterized in that, include: Acquire multiple sets of dual-tone multi-frequency signal training data; wherein each set of the training data includes: the number of tones of the dual-tone multi-frequency signal to be detected, the detection window size and the bridge length; wherein the bridge length is used to indicate the number of consecutive no tones when the dual-tone multi-frequency signal detection ends; Acquire a multivariate function model; wherein the multivariate function model is used to fit the relationship between the number of tones, the detection window size, the bridge length and the detection success rate of the dual-tone multi-frequency signal; The multivariate function model is trained using a plurality of sets of training data of the dual-tone multi-frequency signal to obtain a detection model of the dual-tone multi-frequency signal.

2. The training method of the detection model of dual-tone multi-frequency signal according to claim 1, wherein The method of using a plurality of groups of training data of the dual-tone multi-frequency signal to train the multivariate function model to obtain the detection model of the dual-tone multi-frequency signal comprises: Dividing the training data of multiple groups of dual-tone multi-frequency signals into multiple training data sets; The following training process is performed for the training data of each dual-tone multi-frequency signal in each training data set: For each training data of the dual-tone multi-frequency signal, the number of tones, the detection window size and the bridge length are input into the multivariate function model to obtain a detection success rate; and the difference between the detection success rate and a preset expected success rate is calculated; Calculate an evaluation value using a preset evaluation algorithm and each of the differences; wherein the evaluation value is used to evaluate the training effect at a certain stage; Calculate the difference between the evaluation value and the preset evaluation value to obtain a deviation value; Determining whether the deviation value is greater than or equal to a preset threshold; If yes, adjusting the parameters of the multivariate function model according to the deviation value; If not, a detection model of the dual-tone multi-frequency signal is obtained.

3. the training method of the detection model of dual tone multi-frequency signal according to claim 2 is characterized in that, The adjusting the parameters of the multivariate function model according to the deviation value comprises: If it is determined that the difference between the deviation value and the preset threshold value is greater than or equal to the preset threshold value, adjusting the weight of the parameter corresponding to the number of tones in the multivariate function model to be greater than the weight of each of the detection window size and the bridge length; If it is determined that the difference between the deviation value and the preset threshold is less than the preset threshold value, the proportion of the parameter corresponding to the number of tones in the multivariate function model is adjusted to be less than the respective proportions of the detection window size and the bridge length.

4. according to the training method of the detection model of the dual-tone multi-frequency signal described in any one of claims 1 to 3, it is characterized in that, The step of obtaining multiple sets of dual-tone multi-frequency signal training data includes: Get multiple audio clips; The following processing is performed for each of the audio clips: Detecting a dual-tone multi-frequency signal from the audio clip, and performing de-jittering and de-duplication processing to obtain a processed dual-tone multi-frequency signal; The number of tones, the detection window and the bridge length are detected from the processed dual tone multi-frequency signal as a set of training data of the dual tone multi-frequency signal.

5. A dual-tone multi-frequency signal detection method, characterized in that: include: Obtain a trained dual-tone multi-frequency signal detection model; wherein the detection model is trained based on the training method of the dual-tone multi-frequency signal detection model described in any one of claims 1 to 4; Writing the trained dual-tone multi-frequency signal detection model into the audio calibration parameter file to obtain a modified audio calibration parameter file; Reloading the modified audio calibration parameter file to make the trained dual-tone multi-frequency signal detection model effective; The dual-tone multi-frequency signal is detected by using the effective dual-tone multi-frequency signal detection model, and the detected dual-tone multi-frequency signal is reported.

6. A communication module, characterized in that: include; An acquisition unit, for acquiring a trained dual-tone multi-frequency signal detection model; wherein the detection model is trained based on the training method of the dual-tone multi-frequency signal detection model according to any one of claims 1 to 4; A writing unit is used to write the trained dual-tone multi-frequency signal detection model into an audio calibration parameter file to obtain a modified audio calibration parameter file; A reloading unit, used for reloading the modified audio calibration parameter file to make the trained dual-tone multi-frequency signal detection model effective; The detection unit is used to detect the dual-tone multi-frequency signal by using the effective dual-tone multi-frequency signal detection model, and report the detected dual-tone multi-frequency signal.

7. A training device for a detection model of a dual-tone multi-frequency signal, characterized in that: include: The first acquisition module is used to acquire multiple sets of dual-tone multi-frequency signal training data; wherein each set of the training data includes: the number of tones of the dual-tone multi-frequency signal to be detected, the detection window size and the bridge length; wherein the bridge length is used to indicate the number of consecutive no tones when the dual-tone multi-frequency signal detection ends; A second acquisition module is used to acquire a multivariate function model; wherein the multivariate function model is used to fit the relationship between the number of tones, the detection window size, the bridge length and the detection success rate of the dual-tone multi-frequency signal; The training module is used to train the multivariate function model using a plurality of sets of training data of the dual-tone multi-frequency signal to obtain a detection model of the dual-tone multi-frequency signal.

8. A dual-tone multi-frequency signal detection device, characterized in that: include: A third acquisition module is used to acquire a trained dual-tone multi-frequency signal detection model; wherein the detection model is trained based on the training method of the dual-tone multi-frequency signal detection model described in any one of claims 1 to 4; A writing module is used to write the trained dual-tone multi-frequency signal detection model into an audio calibration parameter file to obtain a modified audio calibration parameter file; A reload module, used for reloading the modified audio calibration parameter file to make the trained dual-tone multi-frequency signal detection model effective; The detection module is used to detect the dual-tone multi-frequency signal by using the effective dual-tone multi-frequency signal detection model, and report the detected dual-tone multi-frequency signal.

9. An electronic device, characterized in that: include: A processor, a memory and a communication bus, wherein the processor and the memory communicate with each other via the communication bus; The memory is used to store computer programs; The processor is used to execute the program stored in the memory to implement the dual-tone multi-frequency signal detection model training method described in any one of claims 1 to 4 or the dual-tone multi-frequency signal detection method described in claim 5.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the training method for the dual-tone multi-frequency signal detection model described in any one of claims 1 to 4 or the dual-tone multi-frequency signal detection method described in claim 5 is implemented.

Images