Cochlear implant sound signal encoding method, processor, medium and cochlear implant

Abstract

The application provides an artificial cochlea sound signal coding method, a processor, a medium and an artificial cochlea. Different coding methods are applied to high-frequency characteristic parts and low-frequency characteristic parts in collected sound signals, so that the electrode channel of the artificial cochlea implant can generate a stimulating current covering the low-frequency domain and the high-frequency domain at the same time, the problems of low intelligibility and low clarity in the existing artificial cochlea technology are solved, the signal is not easily affected by the electric signal, and the performance of the artificial cochlea sound signal coding algorithm in a low-power and low-computing signal processing device is improved.

Description

Technical Field

[0001] This application relates to the field of cochlear implant signal processing, and in particular to a method for encoding cochlear implant sound signals, a processor, a medium, and a cochlear implant. Background Technology

[0002] Cochlear implant technology is currently recognized worldwide as the only effective solution for restoring hearing to patients with bilateral severe or profound sensorineural hearing loss. For most mainstream cochlear implant systems, external sound is first collected by a microphone and converted into an electrical signal. This signal then undergoes pre-emphasis, compression, noise reduction, and encoding processing before being transmitted to the body via a transmitting coil mounted behind the ear. When the receiving coil in the implant senses the signal, a decoding chip decodes the received electrical signal, causing the implant's stimulation electrodes to generate electrical stimulation, ultimately helping the patient's auditory nerve to produce hearing.

[0003] In the operation of cochlear implants, the sound coding algorithm is a crucial factor determining the effectiveness of artificial hearing reconstruction. However, in clinical trials, the process of hearing reconstruction using sound coding algorithms is easily limited by the performance and power consumption of signal processing equipment, resulting in significant losses in both time and frequency resolution of the encoded signal. Existing technologies commonly employ the CIS (Computed Analog Sounding) speech coding scheme, which improves upon the electrical interference problem caused by simultaneous electrode stimulation in traditional compressed analog (CA) schemes through a continuous interval sampling strategy and utilizes a higher stimulation rate to obtain better temporal domain representation information. The ACE (Acute Enzyme and Cochlear Imaging) speech coding scheme uses waveform coding. While this scheme improves upon the insufficient frequency resolution of the CIS scheme and selects the channels with the highest energy from sub-bands for stimulation, it still has shortcomings.

[0004] Current technologies suffer from low clarity and intelligibility in reconstructed speech signals, and the temporal resolution of the signal is easily affected by electrical signals. These problems are particularly pronounced in low-power, low-computing-power signal processing devices. Therefore, to achieve better auditory results, it is necessary to extract speech signal features as effectively as possible and improve speech coding strategies to provide patients with a better wearing experience. Summary of the Invention

[0005] In view of the shortcomings of the prior art described above, the purpose of this application is to provide a cochlear implant sound signal encoding method, processor, medium and cochlear implant, to solve the problems of low clarity and intelligibility and the susceptibility of the signal to electrical signals in the existing cochlear implant technology, and at the same time improve the performance of the cochlear implant sound signal encoding algorithm in low power consumption and low computing power signal processing devices.

[0006] To achieve the above and other related objectives, a first aspect of this application provides a method for encoding sound signals in a cochlear implant, applied to a cochlear implant signal processor; the method includes: acquiring a sound signal; extracting low-frequency features and high-frequency features from the acquired sound signal; performing interval mapping operations on the extracted low-frequency features and high-frequency features based on corresponding preset parameters to obtain a low-frequency mapping interval and a high-frequency mapping interval respectively; merging and outputting the low-frequency mapping interval and the high-frequency mapping interval so that the electrode channel of the cochlear implant can generate a stimulation current covering the low-frequency and high-frequency domains.

[0007] In some embodiments of the first aspect of this application, the process of extracting low-frequency features from the acquired sound signal includes: performing a full-wave rectification operation on the acquired sound signal; performing a filtering operation on the sound signal after the full-wave rectification operation to obtain low-frequency signals of multiple channels; and performing a downsampling operation on the low-frequency signals according to a preset stimulation rate to obtain the low-frequency features of the sound signal.

[0008] In some embodiments of the first aspect of this application, a low-pass filter is used to filter the sound signal that has undergone full-wave rectification.

[0009] In some embodiments of the first aspect of this application, the process of extracting high-frequency features from the acquired sound signal includes: filtering the acquired sound signal through a preset filter bank to obtain a high-frequency complex spectrum; calculating the energy value of the high-frequency complex spectrum to obtain a high-frequency energy spectrum; performing channel merging based on the high-frequency energy spectrum and outputting the feature energy spectra of multiple sub-bands; sorting the feature energy spectra of the multiple sub-bands from largest to smallest, retaining the feature energy spectrum values ​​of the first few sub-bands in the energy spectrum ranking, and setting the feature energy spectrum values ​​of the remaining sub-bands to empty, thereby generating high-frequency features.

[0010] In some embodiments of the first aspect of this application, the preset filter bank includes the WOLA filter bank.

[0011] In some embodiments of the first aspect of this application, the process of performing interval mapping operations on the extracted low-frequency features and high-frequency features includes: using multiple sets of threshold parameters to perform mapping window limiting operations on the low-frequency features and high-frequency features of the input mapping window respectively; the mapping window is a set of threshold parameters containing a maximum value parameter and a minimum value parameter; mapping the low-frequency features and high-frequency features after the mapping window limiting operation to a current energy level range within a preset range to obtain the timing information of the low-frequency features and the timing information of the high-frequency features.

[0012] In some embodiments of the first aspect of this application, the process of performing interval mapping operation on the high-frequency feature includes, after obtaining the timing information of the high-frequency feature and before outputting it, performing the following: upsampling the timing information of the high-frequency feature in the current energy level interval distributed and mapped to the preset range, and aligning the timing information of the high-frequency feature with the timing information of the low-frequency feature, so that the timing information of the low-frequency feature and the high-frequency feature can be merged and output.

[0013] To achieve the above and other related objectives, a second aspect of this application provides a cochlear implant signal processor that establishes a communication connection with a cochlear implant. The cochlear implant signal processor includes: a signal acquisition module for acquiring sound signals; a feature extraction module for extracting low-frequency and high-frequency features from the acquired sound signals; an interval mapping module for performing interval mapping operations on the extracted low-frequency and high-frequency features based on corresponding preset parameters to obtain low-frequency and high-frequency mapping intervals, respectively; and a signal output module for merging and outputting the low-frequency and high-frequency mapping intervals to generate a stimulation current covering the low-frequency and high-frequency domains in the electrode channel of the cochlear implant.

[0014] To achieve the above and other related objectives, a third aspect of this application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the method.

[0015] To achieve the above and other related objectives, a fourth aspect of this application provides a cochlear implant, comprising a cochlear signal processor and a cochlear implant as provided in the second aspect of this application; wherein the cochlear signal processor establishes a communication connection with the cochlear implant.

[0016] As described above, this application relates to the field of cochlear implant signal processing, and in particular to a cochlear implant sound signal encoding method, processor, medium, and cochlear implant. This application has the following beneficial effects: by applying different encoding methods to the high-frequency and low-frequency characteristic components of the acquired sound signal, it improves the clarity and intelligibility of existing cochlear implant devices, enhances the anti-interference capability of the cochlear implant, and improves the performance of existing cochlear implant sound signal encoding algorithms in low-power, low-computing-power signal processing devices. Attached Figure Description

[0017] Figure 1 The diagram shown is a flowchart of an embodiment of a cochlear implant sound signal encoding method of this application.

[0018] Figure 2A The diagram shown is a flowchart of an embodiment of a cochlear implant sound signal encoding method of this application.

[0019] Figure 2B The diagram shown is a flowchart of an embodiment of a cochlear implant sound signal encoding method of this application.

[0020] Figure 3A The diagram shown is a flowchart of an embodiment of a cochlear implant sound signal encoding method of this application.

[0021] Figure 3B The diagram shown is a flowchart of an embodiment of a cochlear implant sound signal encoding method of this application.

[0022] Figure 4A The diagram shown is a flowchart of an embodiment of a cochlear implant sound signal encoding method of this application.

[0023] Figure 4B The diagram shown is a flowchart of an embodiment of a cochlear implant sound signal encoding method of this application.

[0024] Figure 5 The diagram shown is a schematic representation of the interval mapping principle in an embodiment of a cochlear implant sound signal encoding method of this application.

[0025] Figure 6 The diagram shown is a schematic representation of the principle of interpolation of high-frequency features in an embodiment of the cochlear implant sound signal encoding method of this application.

[0026] Figure 7 The diagram shown is a structural schematic of an embodiment of a cochlear implant sound signal encoding method of this application.

[0027] Figure 8 The diagram shown is a flowchart of an embodiment of a cochlear implant sound signal encoding method of this application.

[0028] Figure 9 The diagram shown is a time-domain waveform of the input signal in an embodiment of a cochlear implant sound signal encoding method of this application.

[0029] Figure 10 The image shown is a waveform diagram of the sub-signal output after processing by a bandpass filter bank in an embodiment of a cochlear implant sound signal encoding method of this application.

[0030] Figure 11 The image shown is an envelope diagram of the sub-signals output by the bandpass filter bank in an embodiment of the cochlear implant sound signal encoding method of this application.

[0031] Figure 12 The diagram shows the electrode channels of the cochlear implant sound signal processor after the cochlear implant sound signal encoding operation is performed in an embodiment of the cochlear implant sound signal encoding method of this application. Detailed Implementation

[0032] The following specific examples illustrate the implementation of this application. Those skilled in the art can easily understand other advantages and effects of this application from the content disclosed in this specification. This application can also be implemented or applied through other different specific embodiments, and various details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of this application. It should be noted that, unless otherwise specified, the following embodiments and features in the embodiments can be combined with each other.

[0033] It should be noted that in the following description, reference is made to the accompanying drawings, which illustrate several embodiments of this application. It should be understood that other embodiments may also be used, and changes in mechanical composition, structure, electrical system, and operation may be made without departing from the spirit and scope of this application. The following detailed description should not be considered limiting, and the scope of the embodiments of this application is defined only by the claims of the published patent. The terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the application. Spatially related terms, such as “upper,” “lower,” “left,” “right,” “below,” “below,” “lower part,” “above,” “upper part,” etc., may be used herein to illustrate the relationship between one element or feature shown in the figures and another element or feature.

[0034] In this application, unless otherwise expressly specified and limited, the terms "installation," "connection," "linking," "fixing," and "holding" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection between two components. Those skilled in the art can understand the specific meaning of the above terms in this application according to the specific circumstances.

[0035] Furthermore, as used herein, the singular forms “a,” “an,” and “the” are intended to include the plural forms as well, unless the context indicates otherwise. It should be further understood that the terms “comprising,” “including,” indicate the presence of the stated feature, operation, element, component, item, kind, and / or group, but do not preclude the presence, occurrence, or addition of one or more other features, operations, elements, components, items, kinds, and / or groups. The terms “or” and “and / or” as used herein are interpreted as inclusive, or mean any one or any combination thereof. Thus, “A, B, or C” or “A, B, and / or C” means “any one of: A; B; C; A and B; A and C; B and C; A, B, and C.” Exceptions to this definition arise only when combinations of elements, functions, or operations are inherently mutually exclusive in some manner.

[0036] To address the problems mentioned above, this invention provides a solution to the issues of low clarity and intelligibility, and susceptibility to electrical signal interference, present in existing cochlear implant technology. Furthermore, to make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions in the embodiments of this invention are further described in detail below with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0037] Before providing a further detailed description of the present invention, the nouns and terms used in the embodiments of the present invention are explained, and the nouns and terms used in the embodiments of the present invention are subject to the following interpretations:

[0038] <1> Cochlear implant: A cochlear implant is an electronic device that uses an external speech processor to convert sound into electrical signals in a specific coded form. These signals are then transmitted through an implanted electrode system to directly stimulate the auditory nerve, thereby restoring or reconstructing the patient's hearing function.

[0039] <2> Full-wave rectification: Full-wave rectification is a circuit that rectifies alternating current. In this type of rectifier circuit, current flows through one rectifier device (such as a crystal diode) during one half-cycle, and through a second rectifier device during the other half-cycle. The connection of the two rectifier devices allows the current flowing through them to flow through the load in the same direction.

[0040] <3> Filtering: Filtering is the process of removing specific frequency bands from a signal. Filtering is an important measure for suppressing and preventing interference, and it is divided into classical filtering and modern filtering.

[0041] <4> Low-pass filter: A low-pass filter is an electronic filter that allows signals below the cutoff frequency to pass through, but blocks signals above the cutoff frequency.

[0042] <5> Bandpass filter: A bandpass filter is a filter that allows frequency components within a specific frequency range to pass through, but attenuates frequency components in other ranges to extremely low levels. It is the opposite of a bandstop filter. Bandpass filters can also be created by combining low-pass and high-pass filters.

[0043] <6> Downsampling: Downsampling is a multi-rate digital signal processing technique, which can also be understood as a process of reducing the signal sampling rate. It is usually used to reduce the data transmission rate or data size. During downsampling, filters are used to reduce distortion caused by aliasing. Because aliasing can occur during downsampling, the part of the system that has downsampling functionality is called a frequency downsampling unit.

[0044] <7> Upsampling: Upsampling is the process of acquiring samples of an analog signal. Sampling converts a signal that is continuous in both time and amplitude into a discrete signal in both time and amplitude under the action of a sampling pulse. Therefore, sampling is also known as the discretization process of waveform.

[0045] <8> Complex spectrum: A method of representing a signal or noise as a frequency function using a sequence of complex coefficients of a Fourier transform or Fourier series.

[0046] <9> Subband: In subband coding, the original signal is transformed from the time domain to the frequency domain, then divided into several subbands, and each subband is digitally encoded. The frequency band with specific characteristics generated by passing the original signal through a bandpass filter is called a subband.

[0047] <10> WOLA (Weighted Overlap and Add) filter bank: The multiphase structure and weighted superposition and addition structure of the filter bank is an efficient implementation of the DFT filter bank. Its characteristics are that it does not restrict the relationship between the data decimation rate and the number of channels multiplied by an integer, which has high flexibility and can efficiently and conveniently process data in multiple channels.

[0048] <11> DFT (Discrete Fourier Transform): The Discrete Fourier Transform can transform a signal from the time domain to the frequency domain, and both the time domain and the frequency domain are discrete.

[0049] <12> FFT (Fast Fourier transform): The Fast Fourier Transform is an efficient algorithm of the DFT, which can be divided into time decimation method and frequency decimation method.

[0050] This invention provides a cochlear implant sound signal encoding method, a cochlear implant signal processor, a storage medium storing an executable program for implementing the cochlear implant sound signal encoding method, and a cochlear implant for implementing the cochlear implant sound signal encoding method. Regarding the implementation of the cochlear implant sound signal encoding method, this invention will describe exemplary implementation scenarios.

[0051] like Figure 1 The diagram illustrates a flowchart of a cochlear implant sound signal encoding method according to an embodiment of the present invention. The cochlear implant sound signal encoding method in this embodiment mainly includes the following steps:

[0052] Step S11: Acquire sound signals.

[0053] In some embodiments of the present invention, the cochlear implant acquires sound signals through a built-in microphone. The process includes: sound waves propagate through the air to the diaphragm of the microphone, the diaphragm vibrates with the air and generates a corresponding electrical signal according to the amplitude of the vibration; then the acquired electrical signal is converted into a digital signal by an analog-to-digital converter in the microphone, so as to be input into the cochlear implant sound processor and processed based on the cochlear implant sound signal encoding method provided in this application.

[0054] Step S12: Extract low-frequency features and high-frequency features from the acquired sound signal.

[0055] To facilitate a better understanding by those skilled in the art of low-frequency feature extraction from acquired sound signals provided by this invention, the following will combine... Figure 2A and Figure 2B This will provide a more comprehensive explanation of the entire process of low-frequency feature extraction.

[0056] In this embodiment of the invention, the process of extracting low-frequency features in step S12 is as follows: Figure 2A As shown, it includes three processes: full-wave rectification, low-pass filtering, and downsampling. Specifically, as... Figure 2B The diagram includes the following sub-steps:

[0057] Step S21: Perform full-wave rectification on the acquired sound signal.

[0058] In a preferred embodiment of the invention, before performing the full-wave rectification operation on the acquired sound signal, the following is further performed: The acquired sound signal is decomposed into multiple sub-signals using a bandpass filter bank. It should be understood that the bandpass filter bank consists of multiple bandpass filters with preset ranges. Specifically, the bandpass filter utilizes a high-pass filter and a low-pass filter to work together to produce a filtering effect. The cutoff frequencies of the high-pass and low-pass filters can be used simultaneously as the lower and upper limits of the bandpass filter, thereby filtering out sub-signals of specific frequencies from the input signal. The bandpass filter bank limits the frequency range of the multiple sub-signals according to the number of input electrodes in the cochlear implant, thereby decomposing the acquired complete sound signal into multiple sub-signals.

[0059] It is worth noting that the reason for using a bandpass filter bank to decompose the acquired sound signal into multiple sub-signals is that the technical problem this invention aims to solve is to improve the intelligibility and clarity of existing cochlear implant coding algorithms. However, in reality, interference signals such as public AC power can easily cause aliasing if introduced into subsequent coding algorithms, thus affecting the intelligibility and clarity of the cochlear implant. This invention uses a bandpass filter bank to filter out multiple low-frequency sub-signals suitable for cochlear implant electrodes, making the sub-signal coverage more accurate and flexible. It also has a better effect on eliminating interference signals.

[0060] In some embodiments of this invention, full-wave rectification is achieved through a full-wave rectifier circuit. In this invention, the full-wave rectifier circuit can be a center-tapped full-wave rectifier circuit or a bridge rectifier circuit. Inputting the acquired digital signal into the full-wave rectifier circuit converts the current from a complete cycle of alternating current into pulsating direct current. Pulsating direct current refers to current whose direction does not change, but whose magnitude changes over time. Full-wave rectification utilizes the signal characteristics of the two half-waves of alternating current, improving the rectifier's efficiency and making the rectified current easier to smooth. This provides a smoother, more stable current signal with richer features for cochlear implant sound signal encoding methods, facilitating subsequent encoding processing and minimizing signal feature loss during encoding.

[0061] Step S22: Filter the sound signal after full-wave rectification to obtain low-frequency signals in multiple channels.

[0062] In some embodiments of the present invention, the process of filtering the sound signal after full-wave rectification includes the following steps: The full-wave rectified sound signal is input into a low-pass filter bank of a preset frequency. It should be understood that the full-wave rectified sound signal consists of sub-signals from multiple channels, and the low-pass filter bank contains multiple low-pass filters with different preset frequencies. Each low-pass filter allows only signals below the preset frequency to pass through without attenuation, while signals above the preset frequency cannot pass through. The low-pass filter bank can further remove high-frequency noise present in the sound signal, increase the smoothness of the sound signal, and help reduce errors in the subsequent low-frequency feature extraction process.

[0063] Furthermore, this invention employs cascaded second-order IIR filters to implement a low-pass filter operation, thereby obtaining low-frequency signals from multiple channels. It should be noted that the second-order IIR low-pass filtering algorithm is a circuit analog filter that uses a second-order difference equation to describe signal variations and eliminates noise by attenuating high-frequency signals. In this invention, the cascaded implementation of second-order IIR filters allows for better extraction of sound features containing richer low-frequency signal characteristics.

[0064] Step S23: Perform downsampling on the low-frequency signal according to a preset stimulation rate to obtain the low-frequency characteristics of the sound signal.

[0065] It should be understood that the preset stimulation rate refers to a stimulation rate that is preset according to the actual application requirements. The embodiments of the present invention do not limit the specific value of the preset stimulation rate.

[0066] To facilitate a better understanding by those skilled in the art of extracting high-frequency features from acquired sound signals, the following will combine... Figure 3A and Figure 3B This will provide a more comprehensive explanation of the entire process of high-frequency feature extraction.

[0067] In this embodiment of the invention, the process of extracting high-frequency features in step S12 is as follows: Figure 3A As shown, this includes two processes: calculating the energy spectrum and channel merging. Specifically, as... Figure 3B The diagram includes the following sub-steps:

[0068] Step S31: The acquired sound signal is filtered through a preset filter bank to obtain a high-frequency complex spectrum.

[0069] Preferably, the filter bank used for filtering operations during high-frequency feature extraction is a Weighted Overlap-Add (WOLA) filter bank. The WOLA filter bank can perform channelization processing on the acquired audio signal; the channelization process divides the broadband audio signal into several narrower sub-bands and processes the signals of multiple sub-bands using different operation types and parameters. The channelization process is as follows: the WOLA filter uses a window function to sequentially shift the input audio signal in the time domain according to the window length of the window function, and applies different operation types and parameters to different segments of the audio signal during the shifting operation. The operation types applied to the audio signal include: data weighting operation, superposition operation, FFT (Fast Fourier Transform) operation, and complex modulation operation.

[0070] It is worth noting that the WOLA filter bank overcomes many limitations of the DFT (Discrete Fourier Transform) multinomial filtering structures commonly used in existing signal processing. For example, the DFT requires the number of channels to be consistent with the decimation factor and cannot segment the signal. The WOLA filter bank, through its segmented and refined processing of the audio signal and flexible signal processing methods, preserves richer high-frequency features of the original audio signal, improving the intelligibility and resolution of high-frequency audio signals in cochlear implants.

[0071] Step S32: Calculate the energy value of the high-frequency complex spectrum to obtain the high-frequency energy spectrum.

[0072] For example, the energy spectrum of the high-frequency complex spectrum obtained by filtering through the WOLA filter bank is calculated; wherein, in order to avoid data overflow of the energy spectrum, the square root operation is performed on the energy spectrum value; the energy spectrum calculation process is to take the root mean square (RMS) of the high-frequency complex spectrum value as the energy value of each sub-band to obtain the energy spectrum of the signal, and the energy spectrum is used to characterize the energy distribution of the energy values ​​of each sub-band of the signal.

[0073] Step S33: Perform channel merging operation based on the high-frequency energy spectrum and output the characteristic energy spectra of multiple sub-bands.

[0074] For example, the process of channel merging based on the high-frequency energy spectrum includes: customizing a frequency band allocation table according to actual needs, and merging the energy spectrum into a characteristic energy spectrum output with multiple sub-frequency bands based on the frequency band allocation table. The advantage of customizing the frequency band allocation table is that it improves the flexibility of the high-frequency characteristic energy spectrum merging method. The merging method can be adjusted according to actual needs to retain high-frequency features to the greatest extent, thereby improving the resolution, intelligibility, and accuracy of the cochlear implant.

[0075] Step S34: Sort the feature energy spectra of the multiple sub-bands from largest to smallest, retain the feature energy spectrum values ​​of the multiple sub-bands with the highest energy spectra, and set the feature energy spectrum values ​​of the remaining sub-bands to empty, so as to generate high-frequency features.

[0076] In some embodiments of the present invention, the characteristic energy spectra of the multiple sub-bands are sorted from largest to smallest, and the characteristic energy spectrum values ​​of the top-ranked sub-bands are retained, while the characteristic energy spectrum values ​​of the remaining sub-bands are set to zero to generate high-frequency features. This process is called the eigenvalue maximization operation. The purpose of the eigenvalue maximization operation is to select the characteristic energy containing the main sound information, and by setting the energy values ​​of the remaining sub-bands to zero, the interference of high-frequency noise, interference, and invalid information noise with small energy values ​​is eliminated, while retaining effective high-frequency characteristic information. This improves the resolution of the high-frequency signal and ultimately enhances the clarity and intelligibility of the received high-frequency sound information for cochlear implant wearers.

[0077] Preferably, the bandpass filter bank and low-frequency feature extraction processes described in steps S21 to S23 can be calculated by a filter engine; the WOLA filter bank and high-frequency feature extraction processes described in steps S31 to S34 can be implemented by a configurable signal processing accelerator (HEAR Configurable Accelerator), and the remaining encoding functions can be implemented by a digital signal processor. The advantage of this approach is that the three processors can perform targeted calculations and management for different functions, fully utilize the advantages of each processor, and work together to reduce the computational load of a single processor, improve computational efficiency, reduce latency in the cochlear implant signal encoding process, and enhance the user experience.

[0078] Step S13: Perform interval mapping operations on the extracted low-frequency features and high-frequency features based on the corresponding preset parameters to obtain the low-frequency mapping interval and the high-frequency mapping interval, respectively.

[0079] To facilitate a better understanding by those skilled in the art of performing interval mapping operations on high-frequency and low-frequency features provided by this invention, the following will combine... Figure 4A and Figure 4B This will provide a more comprehensive explanation of the entire process of the interval mapping operation.

[0080] In this embodiment of the invention, the process of performing the interval mapping operation in step S13 is as follows: Figure 4A As shown, this includes window mapping operations, logarithm operations, and current mapping operations. Specifically, as... Figure 4B The diagram includes the following sub-steps:

[0081] Step S41: Use multiple sets of threshold parameters to perform mapping window limiting operations on the low-frequency and high-frequency features of the input mapping window respectively.

[0082] In some embodiments of the present invention, the specific process of using multiple sets of threshold parameters to perform mapping window limiting operations on the low-frequency and high-frequency features of the input mapping window is as follows: Based on the threshold parameters of the mapping window, the energy feature value of the input mapping interval is judged. If the energy feature value of the current sub-band of the input mapping interval is less than the minimum value of the mapping window, the energy feature value is set to zero, that is, the energy feature value will not generate the corresponding output. If the energy feature value of the current sub-band of the input mapping interval is greater than the maximum value of the mapping window, the energy feature value of the input mapping interval is set to the maximum value of the mapping window. Otherwise, the energy feature value of the input mapping interval is kept as is.

[0083] It is worth noting that the purpose of the mapping window limiting operation is to restrict the values ​​of energy features to a certain range, thereby eliminating the adverse effects that may be introduced by the energy feature values ​​of singular samples. The mapping window limiting operation improves the accuracy of audio signal processing, eliminates errors that may be introduced by singular samples, reduces computational costs, and further reduces errors generated during audio signal encoding.

[0084] Step S42: Map the low-frequency and high-frequency feature distributions that have undergone the mapping window limiting operation to a current energy level range within a preset range.

[0085] In some embodiments of the present invention, the process of obtaining the timing information of the low-frequency features and the timing information of the high-frequency features includes: mapping the energy feature values ​​after the mapping window limiting operation to an energy level in decibels (dB) using a logarithmic mapping method, to obtain the logarithmic mapping of the energy features after the mapping window limiting operation. The advantage of performing the logarithmic operation is that it can amplify the feature information present in the sound signal, enabling the linear mapping operation to provide the electrode channel corresponding to the cochlear implant electrode with a current containing more sound signal features, thereby improving the resolution and clarity of the cochlear implant.

[0086] Furthermore, the linear mapping operation maps the logarithmic value obtained by the logarithmic operation to a specified current energy level range and outputs it to the electrodes of the cochlear implant, so that the cochlear implant electrodes can generate a stimulating current at a specific rate and amplitude according to the magnitude of the output and its corresponding electrode channel.

[0087] Step S14: The low-frequency mapping interval and the high-frequency mapping interval are merged and output so that the electrode channel of the cochlear implant can generate a stimulation current covering the low-frequency domain and the high-frequency domain.

[0088] like Figure 5The diagram illustrates the process of merging the low-frequency and high-frequency mapping intervals. The low-frequency and high-frequency feature outputs are controlled by different mapping windows. This is because, depending on the feature extraction method—low-frequency features are extracted by a bandpass filter bank, and high-frequency features by a WOLA filter bank—even when the input low- and high-frequency acoustic signals have the same sound pressure level, their feature output amplitudes are usually not the same. Therefore, by applying different mapping window thresholds to the low- and high-frequency features, acoustic signals with the same sound pressure level can be normalized to the same current level.

[0089] More preferably, such as Figure 6 As shown, after obtaining the timing information of the high-frequency features and before outputting the merged timing features, the timing information of the high-frequency features in the current energy level range mapped to the preset range is upsampled. The timing information of the high-frequency features is then aligned with the timing information of the low-frequency features for merging and output. The upsampling process can employ linear interpolation or second-order Bezier spline interpolation.

[0090] It's worth noting that the upsampling process ensures that the timing information of high-frequency features is consistent with that of low-frequency features. This timing information refers to the sampling frequencies of the high-frequency and low-frequency features. The reason for the inconsistent sampling rates of high-frequency and low-frequency features is that when using a WOLA filter bank to transform the sound signal from the time domain to the frequency domain, the WOLA filter bank causes the sampling rate of high-frequency features to be much lower than that of low-frequency features. This inconsistency in sampling frequencies prevents the high-frequency and low-frequency features from being combined, leading to a problem where no output can be achieved. Upsampling the timing information of high-frequency features unifies the sampling frequencies of high-frequency and low-frequency features, allowing them to be combined for output. This benefits the cochlear implant electrode output, ensuring that the stimulation current contains both high-frequency and low-frequency features, thus improving the cochlear implant's resolution and intelligibility.

[0091] The above description provides an exemplary method for encoding cochlear implant sound signals in embodiments of the present invention. The following description will describe the cochlear implant signal processor provided in embodiments of the present invention.

[0092] like Figure 7 The diagram illustrates the structure of a cochlear implant signal processor according to an embodiment of the present invention. In this embodiment, the cochlear implant signal processor 700 includes:

[0093] Signal acquisition module 71: Used to acquire sound signals;

[0094] Feature extraction module 72: used to extract low-frequency and high-frequency features from the acquired sound signals;

[0095] Interval mapping module 73: used to perform interval mapping operations on the extracted low-frequency features and high-frequency features based on corresponding preset parameters, so as to obtain low-frequency mapping intervals and high-frequency mapping intervals respectively;

[0096] Signal output module 74: used to merge and output the low-frequency mapping range and the high-frequency mapping range so that the electrode channel of the cochlear implant can generate a stimulation current covering the low-frequency domain and the high-frequency domain.

[0097] It should be noted that the cochlear implant signal processor provided in the above embodiments is only illustrated by the division of the above-described program modules when encoding cochlear implant sound signals. In practical applications, the above processing can be assigned to different program modules as needed, that is, the internal structure of the device can be divided into different program modules to complete all or part of the processing described above. In addition, the cochlear implant signal processor and the cochlear implant sound signal encoding method embodiments provided in the above embodiments belong to the same concept, and their specific implementation process can be found in the method embodiments, which will not be repeated here.

[0098] To facilitate a better understanding of the cochlear implant sound signal encoding method provided by this invention by those skilled in the art, the following will be combined with... Figure 8 The present invention relates to an embodiment of the present invention, which will be used to further explain the entire process of the cochlear sound signal encoding method of the applicant.

[0099] In this embodiment of the invention, a bandpass filter bank composed of multiple fourth-order IIR sub-filters is used for filtering operations in the low-frequency feature extraction process, and a WOLA filter bank is used for filtering operations in the high-frequency feature extraction process.

[0100] Specifically, the cochlear implant sound signal encoding method in this embodiment mainly includes the following steps:

[0101] Step S801: Acquire the sound signal and transmit the sound signal to the input stage.

[0102] The input stage in this embodiment is a signal frame composed of a time-domain signal of a specific length. For example... Figure 9 As shown, the sound information collected in this embodiment is a sound signal with a sampling rate of Fs = 16kHz and a duration of 2.5s. The input level frame length is 64, that is, 4ms is one frame.

[0103] Step S802: Pass the input audio signal into the bandpass filter bank to obtain multiple sub-signals.

[0104] like Figure 10 As shown, the bandpass filter bank used in this embodiment consists of 10 (P=10) IIR sub-filters, where each sub-filter is a cascaded implementation of two second-order sections. For any low-frequency electrode channel p=1,2,…,10, its sub-filter structure can be represented by a 2×6 matrix S. (p) Indicates, that is

[0105]

[0106] Where a represents the denominator coefficient of the above filter coefficient matrix; b represents its numerator coefficient.

[0107] Its transfer function H (p) (z) can be represented as:

[0108]

[0109] Where q represents the second-order section number.

[0110] Furthermore, the input signal X is filtered sequentially to obtain 10 sub-signals Y. (p) Its frequency response range gradually increases, as shown in:

[0111] Y (p) (z)=H (p) (z)X(z), (Formula 3)

[0112] Where X(z) represents the Laplace transform of X; Y (p) (z) represents Y (p) Laplace transform.

[0113] Step S803: Extract low-frequency features from multiple sub-signals.

[0114] like Figure 11 As shown, for the above sub-signal Y (p) Perform full-wave rectification, i.e.

[0115] Y ′(p) =|Y (p) |, (Formula 4)

[0116] Furthermore, a second-order IIR filter is used for Y. ′(p) The signal is low-pass filtered to obtain its envelope Y. ″(p) .

[0117] Assuming the cochlear implant system used in this example has a single-channel stimulation rate of 1 kHz, meaning it outputs 4 energy feature values ​​per frame to generate the stimulation signal, and downsampling the envelope with a decimation factor D = 16, the low-frequency feature output for any frame in this example is:

[0118]

[0119] in This represents the i-th sampled value of the output envelope of the p-th channel.

[0120] Step S804: Input the input audio signal to the WOLA filter bank for WOLA analysis to obtain the high-frequency complex spectrum.

[0121] In one embodiment of the present invention, a single-channel Fast Fourier Transform (FFT) window length K is set to 256, and a WOLA analysis window length La is set to 512. The signal within the frequency band from 0 to Fs / 2 = 8kHz is transformed into a complex spectrum composed of 128 sub-bands. Each sub-band has a bandwidth of 62.5Hz. The formula for the high-frequency complex spectrum can be expressed as:

[0122]

[0123] Among them W K =e j2π / K d = 64 is the downsampling factor, k represents the output sub-band number, τ represents the input frame number, h(τ) is the WOLA analysis prototype filter, and each signal frame has a corresponding high-frequency complex spectrum U. (k) .

[0124] Step S805: Perform high-frequency feature extraction on the high-frequency complex spectrum to obtain the energy features of multiple sub-bands.

[0125] In this embodiment, for any frame, the energy E(k) of each sub-band can be calculated as follows:

[0126] E (k) =|U (k) | 2 , (Formula 7)

[0127] Furthermore, based on the frequency band allocation table, the 128 sub-band energies are merged into M new sub-bands. In this example, M = 12, represented as:

[0128] E ′(m) =g·(E (t) +E (t+1) +…+E (t′) ), m=1,2,…,12, (Formula 8)

[0129] Where E ′(m) This represents the energy of the m-th sub-band after merging, where t and t′ represent the start and end indices of this sub-band in the frequency band allocation table, respectively. g is an attenuation factor used to prevent data overflow. The square root of this value is used to obtain its energy characteristic A.(m) , is represented as:

[0130]

[0131] Step S806: Perform a feature value maximization operation on the multiple sub-band energy features output to obtain high-frequency features.

[0132] In this embodiment, the eigenvalue maximization operation includes: retaining N maxima from the 12 sub-band energy features generated after merging and recording their corresponding electrode numbers, and setting the remaining channel energy feature values ​​to zero. In this example, N=6, so the output high-frequency feature after eigenvalue maximization is... It can be represented as:

[0133]

[0134] in, This indicates the selection of N larger sub-band energy eigenvalues, denoted as B. (1) B (2) ,…,B (N) and the rest The energy characteristic value of the channel is set to zero.

[0135] Step S807: Input the obtained low-frequency features and high-frequency features into interval mapping I and interval mapping II, respectively.

[0136] In this embodiment, the interval mapping operation includes: using two different sets of mapping window parameters Xmin (1) Xmax (1) and Xmin (2) Xmax (2) Output the above low-frequency features respectively and high-frequency characteristic output B (1) B (2) ,…,B (N) Perform window limiting operation, i.e.

[0137]

[0138]

[0139] Furthermore, for those subjected to window limiting operations B (n) Logarithmic compression can be represented as:

[0140]

[0141]

[0142] Furthermore, mapping the above equation linearly to the interval [C,T], it can be expressed as:

[0143]

[0144]

[0145] in

[0146]

[0147]

[0148] Step S808: Interpolate the high-frequency feature values ​​of the interval mapping II operation to align the sampling rate of the high-frequency features with the sampling rate of the low-frequency features.

[0149] In this embodiment, linear interpolation is used to upsample the high-frequency mapping output of the system to align it with the timing of the low-frequency component. For the high-frequency mapping output of the τth frame... Calculate the first interpolation result Represented as:

[0150]

[0151] Furthermore, the second interpolation result is calculated. Represented as:

[0152]

[0153] Furthermore, the second interpolation result is calculated. Represented as:

[0154]

[0155] The high-frequency output of the current frame can then be expressed as:

[0156]

[0157] Step S809: Merge and output the low-frequency feature values ​​that have undergone interval mapping I and the high-frequency feature values ​​that have undergone interval mapping II and interpolation operations.

[0158] In this embodiment, the high-frequency output and the low-frequency output are combined. Under the current parameter configuration, the output of the τth frame of the input frame can take the following form:

[0159]

[0160] In the low-frequency section, channels 1 to 10 are fixed output channels. In the high-frequency section, channels 11 to 22 are selected, and six channels with higher energy characteristics are used to generate stimulus outputs, as shown in channels 12, 14, 16, 18, 19, and 20 above. Channels 11, 13, 15, 17, 21, and 22 are set to zero. This process is repeated until all signal frames of the input example signal have been traversed. Figure 12 As shown, the cochlear implant used in conjunction with the above algorithm has a total of 22 (N+P=22) electrode channels that generate stimulation output, and each channel has an independent signal envelope waveform to define the electrode output.

[0161] Those skilled in the art will understand that all or part of the steps of the above-described method embodiments can be implemented using computer program-related hardware. The aforementioned computer program can be stored in a computer-readable storage medium. When executed, the program performs the steps of the above-described method embodiments; and the aforementioned storage medium includes various media capable of storing program code, such as ROM, RAM, magnetic disks, or optical disks.

[0162] In the embodiments provided in this application, the computer-readable and writable storage medium may include read-only memory, random access memory, EEPROM, CD-ROM or other optical disc storage devices, disk storage devices or other magnetic storage devices, flash memory, USB flash drive, portable hard drive, or any other medium capable of storing desired program code in the form of instructions or data structures and accessible by a computer. Additionally, any connection may be appropriately referred to as a computer-readable medium. For example, if instructions are transmitted from a website, server, or other remote source using coaxial cable, fiber optic cable, twisted pair, digital subscriber line (DSL), or wireless technologies such as infrared, radio, and microwave, then the coaxial cable, fiber optic cable, twisted pair, DSL, or wireless technologies such as infrared, radio, and microwave are included in the definition of the medium. However, it should be understood that computer-readable and writable storage media and data storage media do not include connections, carrier waves, signals, or other transient media, but are intended for non-transient, tangible storage media. The disks and optical discs used in the application include compact discs (CDs), laser discs, optical discs, digital multifunction discs (DVDs), floppy disks, and Blu-ray discs, where disks typically copy data magnetically, while optical discs use lasers to copy data optically.

[0163] In summary, this application provides a cochlear implant sound signal encoding method, a processor, a medium, and a cochlear implant. This invention offers a cochlear implant sound signal encoding method that improves the clarity and intelligibility of the cochlear implant. Different encoding methods are applied to the high-frequency and low-frequency characteristic components of the acquired sound signal, thus improving the performance of existing cochlear implant sound signal encoding algorithms in low-power, low-computing-power signal processing devices. Therefore, this application effectively overcomes the various shortcomings of existing technologies and has high industrial applicability.

[0164] The above embodiments are merely illustrative of the principles and effects of this application and are not intended to limit this application. Any person skilled in the art can modify or alter the above embodiments without departing from the spirit and scope of this application. Therefore, all equivalent modifications or alterations made by those skilled in the art without departing from the spirit and technical concept disclosed in this application should still be covered by the claims of this application.

Claims

1. A method for encoding sound signals in a cochlear implant, characterized in that, Applied to a cochlear implant signal processor; the method includes: Collect sound signals; Low-frequency and high-frequency features are extracted from the acquired sound signals; The extracted low-frequency and high-frequency features are subjected to interval mapping operations based on corresponding preset parameters to obtain low-frequency mapping intervals and high-frequency mapping intervals, respectively. Multiple sets of threshold parameters are used to perform mapping window limiting operations on the low-frequency and high-frequency features in the input mapping window. The mapping window is a set of threshold parameters containing maximum and minimum value parameters. The low-frequency and high-frequency features, after the mapping window limiting operation, are distributed and mapped to a current energy level interval within a preset range to obtain the timing information of the low-frequency features and the timing information of the high-frequency features. The timing information of the high-frequency features distributed and mapped to the current energy level interval within the preset range is upsampled, and the timing information of the high-frequency features is aligned with the timing information of the low-frequency features for merging and outputting the timing information of the low-frequency and high-frequency features. The low-frequency mapping range and the high-frequency mapping range are merged and output to generate a stimulation current covering both the low-frequency and high-frequency domains in the electrode channel of the cochlear implant.

2. The cochlear implant sound signal encoding method according to claim 1, characterized in that, The process of extracting low-frequency features from the acquired sound signals includes: The acquired sound signal is subjected to full-wave rectification. The sound signal that has undergone full-wave rectification is filtered to obtain low-frequency signals in multiple channels; The low-frequency signal is downsampled according to a preset stimulation rate to obtain the low-frequency characteristics of the sound signal.

3. The cochlear implant sound signal encoding method according to claim 2, characterized in that, The audio signal, after full-wave rectification, is filtered using a low-pass filter.

4. The cochlear implant sound signal encoding method according to claim 1, characterized in that, The process of extracting high-frequency features from the acquired sound signal includes: The acquired sound signal is filtered through a preset filter bank to obtain a high-frequency complex spectrum; Calculate the energy value of the high-frequency complex spectrum to obtain the high-frequency energy spectrum; Channel merging is performed based on the high-frequency energy spectrum, and the characteristic energy spectra of multiple sub-bands are output. The characteristic energy spectra of the multiple sub-bands are sorted from largest to smallest, and the characteristic energy spectrum values ​​of the first few sub-bands in the energy spectrum are retained. The characteristic energy spectrum values ​​of the remaining sub-bands are set to empty to generate high-frequency features.

5. The cochlear implant sound signal encoding method according to claim 4, characterized in that, The preset filter bank includes the WOLA filter bank.

6. A cochlear implant signal processor, characterized in that, The cochlear implant signal processor establishes a communication connection with the cochlear implant, and includes: Signal acquisition module: used to acquire sound signals; Feature extraction module: used to extract low-frequency and high-frequency features from the acquired sound signals; The interval mapping module is used to perform interval mapping operations on the extracted low-frequency and high-frequency features based on corresponding preset parameters to obtain low-frequency mapping intervals and high-frequency mapping intervals, respectively. Multiple sets of threshold parameters are used to perform mapping window limiting operations on the low-frequency and high-frequency features in the input mapping window. The mapping window is a set of threshold parameters containing maximum and minimum value parameters. The low-frequency and high-frequency features, after the mapping window limiting operation, are distributed and mapped to a current energy level interval within a preset range to obtain the timing information of the low-frequency features and the timing information of the high-frequency features. The timing information of the high-frequency features distributed and mapped to the current energy level interval within the preset range is upsampled, and the timing information of the high-frequency features is aligned with the timing information of the low-frequency features for merging and outputting the timing information of the low-frequency and high-frequency features. Signal output module: used to merge and output the low-frequency mapping range and the high-frequency mapping range so that the electrode channel of the cochlear implant can generate a stimulation current covering the low-frequency domain and the high-frequency domain.

7. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the method of any one of claims 1 to 5.

8. A cochlear implant, characterized in that, It includes the cochlear implant signal processor and the cochlear implant as described in claim 6; wherein the cochlear implant signal processor establishes a communication connection with the cochlear implant.

Images