Post-operative vocal cord rehabilitation assessment method and related apparatus

Abstract

The application is suitable for the field of audio processing technology, and provides a vocal cord postoperative rehabilitation evaluation method and related equipment. The method pre-processes preoperative audio signals and postoperative audio signals to obtain preoperative pre-processed voice signals and postoperative pre-processed voice signals; extracts stable voice segments from the preoperative pre-processed voice signals and the postoperative pre-processed voice signals to obtain preoperative stable voice segments and postoperative stable voice segments; extracts acoustic parameters from the preoperative stable voice segments and the postoperative stable voice segments to obtain preoperative acoustic parameters and postoperative acoustic parameters; compares the preoperative acoustic parameters and the postoperative acoustic parameters to obtain a vocal cord postoperative rehabilitation evaluation result. The application can directly obtain the vocal cord postoperative rehabilitation evaluation result according to the voice signals by uniformly pre-processing the preoperative and postoperative voice signals, extracting stable voice segments, and extracting acoustic parameters, and comparing the extracted acoustic parameters.

Description

Technical Field

[0001] This invention relates to the field of audio processing technology, and in particular to a method and related equipment for postoperative rehabilitation assessment of vocal cords. Background Technology

[0002] After vocal cord surgery (such as polyp removal or nodule removal), the recovery of a patient's voice function needs to be clinically assessed to determine the progress of rehabilitation and decide on subsequent treatment plans. Currently, the assessment of postoperative vocal cord rehabilitation mainly relies on laryngoscopy to observe the morphology of the vocal cords, subjective judgment by otolaryngologists based on auditory perception, or indirect assessment of recovery status by patients completing a self-assessment scale for voice disorders.

[0003] The above assessment methods all require the use of professional examination equipment, rely on the professional auditory judgment of physicians, or rely on the subjective description of the patient's own feelings. There is a lack of technical means that can directly and objectively quantify and compare the patient's pre- and post-operative voice signals and automatically output the assessment results. As a result, the assessment results of the post-operative recovery status of the vocal cords are not easy to quantify, record, track and compare. Summary of the Invention

[0004] This invention provides a method and related equipment for postoperative rehabilitation assessment of vocal cords, in order to solve the technical problems in the prior art where postoperative rehabilitation assessment of vocal cords relies on laryngoscopy or subjective human judgment and lacks objective quantitative comparison based on speech signals.

[0005] In a first aspect, embodiments of this application provide a method for postoperative rehabilitation assessment of the vocal cords, including: Pre-processing was performed on the pre-operative and post-operative audio signals to obtain pre-processed pre-operative and post-operative pre-processed audio signals. Stable speech segments are extracted from the preoperative preprocessed speech signal and the postoperative preprocessed speech signal to obtain preoperative stable speech segments and postoperative stable speech segments. Acoustic parameters were extracted from the preoperative stable speech segment and the postoperative stable speech segment to obtain the preoperative acoustic parameters and postoperative acoustic parameters. The results of the vocal cord postoperative rehabilitation assessment were obtained by comparing the preoperative acoustic parameters and the postoperative acoustic parameters.

[0006] Secondly, embodiments of this application provide a postoperative vocal cord rehabilitation assessment device, comprising: The preprocessing module is used to preprocess the preoperative audio signal and the postoperative audio signal respectively to obtain the preoperative preprocessed speech signal and the postoperative preprocessed speech signal. The stable extraction module is used to extract stable speech segments from the pre-operative pre-processed speech signal and the post-operative pre-processed speech signal respectively, to obtain pre-operative stable speech segments and post-operative stable speech segments. The parameter extraction module is used to extract acoustic parameters from the preoperative stable speech segment and the postoperative stable speech segment respectively, so as to obtain the preoperative acoustic parameters and the postoperative acoustic parameters. The comparative evaluation module is used to compare the preoperative acoustic parameters and the postoperative acoustic parameters to obtain the postoperative rehabilitation evaluation results of the vocal cords.

[0007] Thirdly, embodiments of this application provide a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the steps of the above-described postoperative vocal cord rehabilitation assessment method.

[0008] Fourthly, embodiments of this application provide a readable storage medium storing a computer program that, when executed by a processor, implements the steps of the above-described postoperative vocal cord rehabilitation assessment method.

[0009] Fifthly, embodiments of this application provide a computer program product, which includes a computer program that, when executed by a processor, enables the implementation of the steps of the above-described postoperative vocal cord rehabilitation assessment method.

[0010] In one solution provided by the aforementioned vocal cord postoperative rehabilitation assessment methods, devices, equipment, media, and procedures, the pre-processing of pre-operative and post-operative audio signals using the same procedure ensures consistency in basic conditions such as sampling rate, number of channels, and amplitude range, avoiding incomparability caused by differences in recording equipment or environment. Furthermore, by extracting stable speech segments from the pre-processed pre-operative and post-operative speech signals, the speech segments used for subsequent analysis are free from interference from non-target content such as silence and breath noise. This ensures that the input signals used for acoustic parameter extraction correspond to the patient's actual vocal state, rather than segments affected by occasional noise, thus allowing the extracted acoustic parameters to accurately reflect the vocal cord state at the corresponding time point. Since the pre-operative and post-operative acoustic parameters are obtained based on the same extraction procedure and the same data benchmark, they are comparable. The resulting assessment directly reflects the changes in the vocal cord state after surgery relative to the pre-operative state, providing an assessment basis for the post-operative vocal cord rehabilitation status without the need for laryngoscopy or manual auditory judgment. Attached Figure Description

[0011] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the description of the embodiments of the present invention will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0012] Figure 1 This is a schematic diagram of the structure of a postoperative vocal cord rehabilitation assessment system according to an embodiment of the present invention; Figure 2 This is a flowchart illustrating a method for assessing postoperative rehabilitation of the vocal cords in one embodiment of the present invention; Figure 3 yes Figure 2 A schematic diagram of the implementation process of step S10; Figure 4 yes Figure 2 A schematic diagram of the implementation process of step S20; Figure 5 yes Figure 2 A schematic diagram of the implementation process of step S30; Figure 6 This is a schematic diagram of the structure of a postoperative vocal cord rehabilitation assessment device according to an embodiment of the present invention; Figure 7 This is a schematic diagram of the structure of a computer device according to an embodiment of the present invention. Detailed Implementation

[0013] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0014] It should be understood that, when used in this specification and the appended claims, the term "comprising" indicates the presence of the described features, integrals, steps, operations, elements, and / or components, but does not exclude the presence or addition of one or more other features, integrals, steps, operations, elements, components, and / or collections thereof. It should also be understood that, as used in this specification and the appended claims, the term "and / or" refers to any combination of one or more of the associated listed items and all possible combinations, and includes such combinations.

[0015] Furthermore, in the description of this invention and the appended claims, the terms "first," "second," "third," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0016] References to "one embodiment" or "some embodiments" as described in this specification mean that one or more embodiments of the invention include a specific feature, structure, or characteristic described in connection with that embodiment. Therefore, the phrases "in one embodiment," "in some embodiments," "in other embodiments," "in still other embodiments," etc., appearing in different parts of this specification do not necessarily refer to the same embodiment, but rather mean "one or more, but not all, embodiments," unless otherwise specifically emphasized. The terms "comprising," "including," "having," and variations thereof mean "including but not limited to," unless otherwise specifically emphasized.

[0017] It should be understood that the sequence number of each step in the following embodiments does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0018] To illustrate the technical solution of the present invention, specific embodiments are described below.

[0019] To address the problems mentioned above in the background art, embodiments of this application propose a method, device, equipment, medium, and program product for postoperative vocal cord rehabilitation assessment. The postoperative vocal cord rehabilitation assessment method provided by the embodiments of this invention can be applied to, for example... Figure 1 The vocal cord postoperative rehabilitation assessment system shown includes a client and a server.

[0020] In one embodiment, such as Figure 2 As shown, a method for assessing postoperative rehabilitation of the vocal cords is provided, which is then applied to... Figure 1 Taking the postoperative vocal cord rehabilitation assessment system as an example, the following steps are included: S10: Preprocess the preoperative audio signal and the postoperative audio signal to obtain the preoperative preprocessed speech signal and the postoperative preprocessed speech signal. S20: Extract stable speech segments from the pre-processed speech signal and the post-processed speech signal to obtain pre-operative stable speech segments and post-operative stable speech segments; S30: Extract acoustic parameters from the preoperative stable speech segment and the postoperative stable speech segment to obtain the preoperative acoustic parameters and the postoperative acoustic parameters; S40: The preoperative acoustic parameters and postoperative acoustic parameters are compared to obtain the postoperative rehabilitation assessment results of the vocal cords.

[0021] In this embodiment, there is a certain correspondence between the patient's voice state before and after vocal cord surgery. The preoperative audio is usually recorded when the patient's voice state is relatively stable and can serve as a baseline reference for the patient. The postoperative audio is used to reflect the actual recovery of vocal cord function after surgery. Since the two audio streams come from different recording environments, their recording equipment, sampling rate, number of channels, and other parameters may not be completely consistent. If the two original audio streams are directly analyzed and compared, errors can easily be introduced due to the differences in recording conditions, affecting the comparability of subsequent analysis results.

[0022] Based on the above description, in this embodiment, after acquiring the preoperative audio signal and the postoperative audio signal, the system first preprocesses the two audio signals separately to ensure that the two audio signals are under a unified benchmark in terms of sampling rate, number of channels, amplitude range, etc., so as to obtain the preoperative preprocessed speech signal and the postoperative preprocessed speech signal, which provides a consistent data basis for the comparative analysis of preoperative and postoperative data in subsequent steps.

[0023] For example, if the preoperative audio signal is recorded by a certain type of recording device in dual-channel mode with a sampling rate of 44.1kHz, while the postoperative audio signal is recorded by the patient's mobile phone microphone in mono mode with a sampling rate of 16kHz, the system will convert both to the same number of channels and sampling rate during the preprocessing stage to avoid analysis deviations caused by differences in equipment.

[0024] In this embodiment, the speech signal emitted by the patient during audio recording typically includes not only the vowel vocalization segments used for analysis, but also initial silence, breathing sounds, and noise segments generated due to unstable breathing during vocalization. If these segments are included in subsequent acoustic parameter analysis, it can easily lead to unstable analysis results, or even make it impossible to extract effective formants or voice quality parameters.

[0025] Based on the above description, in this embodiment, after obtaining the pre-processed speech signal and the post-processed speech signal, the system further extracts stable speech segments from both signals respectively. Stable vocal segments suitable for analysis are located from the pre-processed speech signal, resulting in pre-operative and post-operative stable speech segments. It should be noted that the extraction processes for stable speech segments from the pre-operative and post-operative pre-processed speech signals are performed independently; that is, each stable speech segment is determined based on its respective signal content, and the two segments do not need to be identical in time position or duration.

[0026] For example, when a patient records postoperative audio, there may be an unstable vocalization at the beginning due to throat clearing or deep breathing. When extracting stable speech segments, the system will exclude the initial segment and only select segments with stable speech features for subsequent analysis.

[0027] In this embodiment, the preoperative stable speech segment and the postoperative stable speech segment correspond to the actual state of the patient's vocal tract structure and voice function before and after the operation, respectively. In order to make a quantitative comparison of the vocal cord state before and after the operation, it is necessary to extract acoustic parameters that can characterize the above state.

[0028] Based on the above description, in this embodiment, the system extracts acoustic parameters from the preoperative and postoperative stable speech segments respectively, obtaining preoperative and postoperative acoustic parameters. It should be noted that the specific parameter types and extraction methods used in this embodiment are consistent for both preoperative and postoperative data; that is, the preoperative and postoperative data independently obtain their corresponding acoustic parameters according to the same extraction process, thereby ensuring that the preoperative and postoperative data are within a comparable range during subsequent comparisons.

[0029] In this embodiment, the preoperative acoustic parameters and postoperative acoustic parameters reflect the patient's voice status at a certain point in time before and after the operation. However, the preoperative or postoperative acoustic parameters alone cannot directly indicate the patient's recovery status. It is necessary to compare the two to determine what changes have occurred after the operation relative to the preoperative state.

[0030] Based on the above description, in this embodiment, after obtaining the preoperative and postoperative acoustic parameters, the system compares the two to obtain the postoperative vocal cord rehabilitation assessment results. It should be noted that in this embodiment, the preoperative acoustic parameters are used as a reference standard in the comparison process. The changes in postoperative acoustic parameters relative to preoperative acoustic parameters are the focus of the comparison process. The postoperative vocal cord rehabilitation assessment results reflect the specific changes in postoperative acoustic parameters relative to preoperative acoustic parameters.

[0031] In one embodiment, such as Figure 3 As shown, step S10 specifically includes the following steps: S11: Perform channel merging processing on the preoperative audio signal and the postoperative audio signal to obtain a preoperative mono signal and a postoperative mono signal; S12: Normalize the preoperative mono signal and the postoperative mono signal according to the preset reference amplitude to obtain the preoperative normalized signal and the postoperative normalized signal. S13: Resample the preoperative normalized signal and the postoperative normalized signal according to the preset target sampling rate to obtain the preoperative preprocessed speech signal and the postoperative preprocessed speech signal.

[0032] In this embodiment, the recording devices used by the patient for the pre-operative and post-operative audio recordings are often different. For example, the pre-operative audio may be collected using professional recording equipment in a hospital examination room, while the post-operative follow-up audio may be recorded by the patient at home using a mobile phone. The two types of devices differ in their channel configuration; some devices output dual-channel audio, while others output only mono audio. If the number of channels is not standardized, the subtle differences between the left and right channel signals in the dual-channel audio will introduce additional uncertainty in subsequent analysis.

[0033] Based on the above description, after acquiring the preoperative and postoperative audio signals, the system first determines the number of channels in each audio path. If an audio path is a multi-channel signal, the average of the sampled values ​​at the corresponding positions of each channel of that audio path is taken to obtain the mono sampled value at that position. All sampled points of that audio path are processed in this way to obtain the corresponding mono signal. If the audio path itself is mono, it is directly used as the corresponding mono signal, thereby obtaining the preoperative mono signal and the postoperative mono signal respectively.

[0034] In this embodiment, the preoperative mono signal and the postoperative mono signal may also differ in amplitude range. This difference mainly stems from factors such as the gain setting of the recording equipment and the distance between the patient and the microphone, rather than differences in the strength of the patient's voice itself. If the two signals are directly analyzed and compared, the difference in equipment gain may be easily misinterpreted as a change in voice state.

[0035] Based on the above description, after obtaining the preoperative mono signal and the postoperative mono signal, the system acquires the maximum absolute value of the sampled values ​​in each signal path. Then, according to a preset reference amplitude, it scales all sampled values ​​of that signal path proportionally, so that the maximum absolute value of the scaled signal equals the preset reference amplitude. This yields the preoperative normalized signal and the postoperative normalized signal, respectively. It should be noted that in this embodiment, the preoperative and postoperative signals are normalized independently; that is, each signal determines its scaling ratio based on its own maximum absolute value, rather than using a uniform scaling factor, thereby eliminating the influence of differences in device gain.

[0036] For example, if the maximum absolute value of the preoperative mono signal sampling value is 0.3 and the maximum absolute value of the postoperative mono signal sampling value is 0.8, under the same preset reference amplitude, the two signals will be normalized according to different scaling ratios so that the peak amplitude of the two signals after processing is at the same level, avoiding deviations in the calculation results of subsequent energy parameters due to excessive recording volume after surgery.

[0037] In this embodiment, the original sampling rates of the preoperative mono signal and the postoperative mono signal may also be inconsistent. For example, the preoperative audio was recorded at a sampling rate of 44.1 kHz, while the postoperative audio was recorded at a sampling rate of 16 kHz. If the sampling rates are not unified, the relevant calculations involving frequency domain analysis in subsequent steps will be based on different frequency resolutions, resulting in the incomparability of preoperative and postoperative acoustic parameters.

[0038] Based on the above description, after obtaining the preoperative and postoperative normalized signals, the system determines whether the original sampling rate of each signal is equal to the preset target sampling rate. If not, the signal is resampled to convert its sampling rate to the preset target sampling rate; if it is already equal to the preset target sampling rate, no conversion is required. After the above processing, preoperative and postoperative preprocessed speech signals at a uniform sampling rate are obtained, respectively.

[0039] It should be noted that the resampling process in this embodiment can be implemented using polyphase filtering, that is, while changing the sampling rate, the filter suppresses the spectral aliasing that may be introduced by the change in sampling rate. For example, for preoperative audio with an original sampling rate of 44.1kHz and a preset target sampling rate of 16kHz, the system performs low-pass filtering on the signal while reducing the sampling rate, filtering out high-frequency components in the original signal that exceed the Nyquist frequency range of the target sampling rate, so as to prevent these high-frequency components from folding back into the analyzable frequency band after downsampling and interfering with the subsequent formant extraction.

[0040] In one embodiment, such as Figure 4 As shown, step S20 specifically includes the following steps: S21: The pre-processed speech signal and the post-processed speech signal are processed by framing to obtain multiple speech frames. The speech frames are then filtered according to their energy values ​​and zero-crossing rates to obtain pre-operative candidate speech segments and post-operative candidate speech segments. S22: Extract formant parameters from the preoperative candidate speech segments and the postoperative candidate speech segments respectively to obtain the preoperative candidate formant trajectory and the postoperative candidate formant trajectory. S23: Based on the fluctuation degree of the preoperative candidate formant trajectory and the postoperative candidate formant trajectory, determine the preoperative stable speech segment and the postoperative stable speech segment respectively from the preoperative candidate speech segment and the postoperative candidate speech segment.

[0041] In this embodiment, in addition to the vowel segments produced by the patient, the pre-processed speech signal and the post-processed speech signal usually contain unstable vocal components such as initial silence, breathing sounds, voiceless consonants, and breath sounds. The energy distribution and spectral characteristics of these components are significantly different from those of stable vowels. If they are not removed, they will interfere with subsequent formant extraction and voice quality analysis.

[0042] Based on the above description, the system first performs frame processing on the pre-processed speech signal and the post-processed speech signal according to preset frame length and preset frame shift, respectively, to obtain multiple speech frames. For each speech frame, its energy value and zero-crossing rate are calculated. The energy value reflects the overall amplitude level of the signal in that frame, while the zero-crossing rate reflects the frequency with which the signal waveform crosses the zero axis. Voiced vowels typically exhibit higher energy values ​​and lower zero-crossing rates, while voiceless consonants, breath sounds, or silent segments exhibit lower energy values, higher zero-crossing rates, or weak amplitudes. Based on preset energy and zero-crossing rate thresholds, the system retains speech frames with energy values ​​not lower than the preset energy threshold and zero-crossing rates not higher than the preset zero-crossing rate threshold as candidate speech frames. Candidate speech frames that are consecutive in time or have small intervals are merged into candidate speech segments, and candidate speech segments with a duration less than a preset duration threshold are removed. This yields pre-operative candidate speech segments and post-operative candidate speech segments, respectively.

[0043] For example, if a patient clears their throat at the beginning of recording postoperative audio, this segment has a high zero-crossing rate and does not meet the screening criteria for candidate speech frames. The system can exclude this segment from the candidate speech segments during the frame screening stage.

[0044] In this embodiment, although the preoperative and postoperative candidate speech segments have been freed from silence and noise interference, they may still contain transitional segments where the sound quality is not yet stable at the beginning of vocalization. For example, when the patient first begins to speak, the vocal tract has not yet reached the target vocalization position, and the formant position will shift significantly over time. If acoustic parameters are extracted from such segments, the results will not accurately represent the true vocal tract structure under stable vocalization conditions. Based on the above description, the system extracts formant parameters frame-by-frame from the preoperative and postoperative candidate speech segments, obtaining the trajectories of the formants over time for each segment—the preoperative and postoperative candidate formant trajectories.

[0045] In this embodiment, the system further determines stable vocal segments from candidate speech segments based on the fluctuation of preoperative and postoperative candidate formant trajectories over time. It should be noted that smaller fluctuations in the formant trajectories indicate a more stable vocal tract shape during that time period, suggesting the patient is in a state of continuous stable vocalization; larger fluctuations indicate ongoing changes in vocal tract shape during that time period, potentially corresponding to the onset or transition phase of vocalization. The system determines segments that meet the set fluctuation criteria from both preoperative and postoperative candidate speech segments, respectively, as the preoperative and postoperative stable vocal segments. In one embodiment, determining the preoperative stable speech segment and the postoperative stable speech segment from the preoperative candidate speech segment and the postoperative candidate speech segment respectively, based on the fluctuation degree of the preoperative candidate formant trajectory and the postoperative candidate formant trajectory, includes: Sliding window calculations were performed on the preoperative candidate resonance peak trajectories and the postoperative candidate resonance peak trajectories to obtain the preoperative fluctuation parameter sequence and the postoperative fluctuation parameter sequence. Based on the preoperative fluctuation parameter sequence and the postoperative fluctuation parameter sequence, a continuous stable window subset is determined from the preoperative candidate speech segment and the postoperative candidate speech segment respectively to obtain the candidate stable speech segment; If the duration of the candidate stable speech segment does not meet the preset conditions, the window position with the smallest fluctuation degree is determined according to the preoperative fluctuation parameter sequence or the postoperative fluctuation parameter sequence, or the center position of the preoperative candidate speech segment or the postoperative candidate speech segment is determined, so as to obtain the preoperative stable speech segment or the postoperative stable speech segment. If the duration of the candidate stable speech segment meets the preset condition, the candidate stable speech segment is determined as the preoperative stable speech segment or the postoperative stable speech segment.

[0046] In this embodiment, the preoperative candidate formant trajectory and the postoperative candidate formant trajectory are extracted frame by frame. The formant value of a single frame is easily affected by factors such as the position of the analysis window and local instantaneous noise, resulting in occasional jitter. If the stability of the sound is judged based solely on the value of a single frame, it is easy to misjudge the occasional fluctuations caused by local noise as actual changes in the shape of the vocal tract. It is also possible to misjudge a moment when the vocal tract is indeed stable as unstable due to abnormal values ​​in individual frames.

[0047] Based on the above description, the system calculates the preoperative and postoperative candidate formant trajectories by sliding along the time axis with a preset number of frames as the window length. Each time, it takes the formant values ​​of several consecutive frames within the window range, calculates the dispersion of these values, and uses this as the fluctuation parameter at the window's location. This process is repeated until the end of the candidate speech segment, yielding the preoperative and postoperative fluctuation parameter sequences. Thus, the local jitter in the frame-by-frame formant trajectory is smoothed into a measure reflecting the overall dispersion of the formants over a continuous time period, providing a basis for subsequent judgments on stability.

[0048] In this embodiment, the system performs continuity judgment on the time axis for windows whose fluctuation parameters meet preset fluctuation conditions based on the obtained fluctuation parameter sequence. Multiple windows that are consecutively arranged in time and whose fluctuation parameters all meet the preset fluctuation conditions are merged into a continuous stable window subset. It should be noted that in this embodiment, when performing continuity judgment, a small number of windows that do not meet the fluctuation conditions are allowed within the continuous stable window subset. That is, when the fluctuation parameter of a certain window does not meet the preset fluctuation conditions, but the windows before and after it all meet the conditions, and the number of windows that do not meet the conditions does not exceed a preset number, the windows on both sides of that window that meet the conditions are still merged into the same continuous stable window subset, without interrupting the continuity of the subset due to that window. The reason for this processing is that even if the vocal tract shape is in a generally stable state during actual sound production, it may still experience temporary deviations in individual frames due to subtle adjustments in the vocal muscles. If such intervals are not allowed, content that should be considered as the same stable segment is easily split into multiple shorter, independent segments, which is not conducive to obtaining sufficiently long analysis samples later. After the above merging, if there are multiple consecutive stable window subsets, the one with the longest duration is selected to obtain the candidate stable speech segment.

[0049] For example, in the fluctuation parameter sequence corresponding to a candidate speech segment, the fluctuation parameters of windows 5 to 20 all meet the preset fluctuation conditions. However, window 13 does not meet the conditions due to a small local disturbance, but the windows before and after it meet the conditions and the interval is only 1 window. When merging, the system still regards windows 5 to 20 as the same continuous stable window subset, rather than splitting them into two segments with window 13 as the boundary.

[0050] In this embodiment, after obtaining the candidate stable speech segment, the system further determines whether its duration meets the preset duration condition. If the duration of the candidate stable speech segment meets the condition, it means that the segment has sufficient length for subsequent acoustic parameter extraction, and the system directly identifies the candidate stable speech segment as a preoperative stable speech segment or a postoperative stable speech segment.

[0051] In this embodiment, if the fluctuation parameter sequence within a candidate speech segment consistently fails to form a continuous stable window subset whose duration meets the preset conditions—for example, if the patient's postoperative vocalization itself exhibits a certain degree of sound quality fluctuation, causing the formant trajectory to fail to reach the continuous duration required by the preset fluctuation conditions across the entire candidate speech segment—strictly meeting these duration conditions would prevent the audio path from entering the subsequent acoustic parameter extraction process, thus failing to provide an evaluation result. To avoid this situation, the system determines stable speech segments according to preset backtracking rules: one approach is to determine the position corresponding to the window with the smallest fluctuation parameter value (i.e., the most stable vocal tract shape) from the preoperative or postoperative fluctuation parameter sequence, and use this position as the basis for determining stable speech segments; another approach is to directly use the center position of the preoperative or postoperative candidate speech segment as a reference position to determine stable speech segments.

[0052] In one embodiment, such as Figure 5 As shown, step S30 specifically includes the following steps: S31: Extract structural dimension parameters from the preoperative stable speech segment and the postoperative stable speech segment to obtain preoperative structural acoustic parameters and postoperative structural acoustic parameters. S32: Extract functional dimension parameters from the preoperative stable speech segment and the postoperative stable speech segment to obtain preoperative functional acoustic parameters and postoperative functional acoustic parameters; S33: Obtain the preoperative acoustic parameters based on the preoperative structural acoustic parameters and the preoperative functional acoustic parameters, and obtain the postoperative acoustic parameters based on the postoperative structural acoustic parameters and the postoperative functional acoustic parameters.

[0053] In this embodiment, the changes in a patient's voice before and after vocal cord surgery are actually reflected on two different levels: first, changes at the level of vocal tract structure, such as adjustments to the shape of the vocal tract or the opening and closing of the oral cavity and pharynx during phonation, which are usually reflected in the shift of the formant frequency; second, changes in the functional state of the vocal cord vibration itself, such as incomplete vocal cord closure or irregular vibration period after surgery, which are usually reflected in the change in the proportion of noise components in the voice signal. If only one level is analyzed, it is easy to miss abnormalities that may exist on the other level, resulting in an incomplete evaluation result.

[0054] Based on the above description, in this embodiment, the system extracts acoustic parameters from the preoperative and postoperative stable speech segments from the two levels mentioned above: on the one hand, it extracts structural dimension parameters reflecting the state of the vocal tract structure, and on the other hand, it extracts functional dimension parameters reflecting the state of vocal cord vibration function, thus obtaining preoperative structural acoustic parameters, postoperative structural acoustic parameters, preoperative functional acoustic parameters, and postoperative functional acoustic parameters, respectively. It should be noted that the extraction of structural dimension parameters and functional dimension parameters are performed independently. The original signals used for both are the same preoperative or postoperative stable speech segment, but the extraction processes are independent of and do not affect each other.

[0055] In this embodiment, after obtaining the structural acoustic parameters and functional acoustic parameters before and after the operation, the system merges the structural acoustic parameters and functional acoustic parameters corresponding to the same time point (before or after the operation) to obtain the preoperative acoustic parameters and postoperative acoustic parameters that can simultaneously reflect the state of the vocal tract structure and the state of vocal function. These parameters serve as the complete set of acoustic parameters for the preoperative and postoperative comparative analysis in subsequent steps.

[0056] For example, a patient's preoperative structural acoustic parameters show a formant frequency distribution that is close to that of normal vocalization, but the preoperative functional acoustic parameters show a slightly higher proportion of noise components. If judged solely based on structural acoustic parameters, it might be concluded that the patient's vocal tract structure is basically normal. However, by combining functional acoustic parameters, the system can further discover that the patient already has a certain degree of abnormality in the level of vocal vibration function. The preoperative acoustic parameters obtained by combining the two can more completely reflect the patient's preoperative vocal condition compared to parameters of a single dimension.

[0057] In one embodiment, the step of extracting structural dimension parameters from the preoperative stable speech segment and the postoperative stable speech segment to obtain preoperative structural acoustic parameters and postoperative structural acoustic parameters includes: Based on the initial extraction parameters, formant extraction was performed on the preoperative stable speech segment and the postoperative stable speech segment to obtain the preoperative initial structural acoustic parameters and the postoperative initial structural acoustic parameters. If the number of formants in the preoperative initial structural acoustic parameters or the postoperative initial structural acoustic parameters does not meet the preset number condition, the initial extraction parameters are adjusted according to the preset parameter set, and the formants are re-extracted from the preoperative stable speech segment or the postoperative stable speech segment according to the adjusted extraction parameters to obtain multiple sets of candidate structural acoustic parameters. Based on the number of resonance peaks and the frequency distribution of resonance peaks in the multiple sets of candidate structural acoustic parameters, the preoperative structural acoustic parameters or the postoperative structural acoustic parameters are determined from the multiple sets of candidate structural acoustic parameters. If the number of resonance peaks in the preoperative initial structural acoustic parameters or the postoperative initial structural acoustic parameters meets the preset quantity condition, the preoperative initial structural acoustic parameters or the postoperative initial structural acoustic parameters are determined as the preoperative structural acoustic parameters or the postoperative structural acoustic parameters.

[0058] In this embodiment, formant extraction is typically based on linear predictive analysis. This involves pre-emphasizing and windowing the speech frame, calculating its autocorrelation sequence, and solving for the linear prediction coefficients to obtain the corresponding full-pole spectral envelope. The location of the local maximum amplitude within this envelope is then detected, yielding the formant frequencies. When extracting formants from preoperative and postoperative stable speech segments, the system first executes the above process using pre-set initial extraction parameters, including the linear prediction order and peak detection threshold. This yields preoperative and postoperative initial structural acoustic parameters, each containing the detected formant frequencies and their corresponding numbers.

[0059] In this embodiment, the voice signal of patients in the early postoperative period often has the characteristics of low energy, heavy breath sound component, and flat spectral envelope. When using fixed initial extraction parameters to extract formants on such signals, the position of the peak in the spectral envelope may not be prominent enough, resulting in fewer detected formants than the actual number of vocal tract formants, i.e., formant missed detection.

[0060] Based on the above explanation, after obtaining the initial preoperative or postoperative structural acoustic parameters, the system determines whether the number of detected formants meets the preset threshold. If the number of formants does not meet the threshold, it indicates that the extraction results under the current initial extraction parameters are insufficient to support subsequent analysis. In this case, the system adjusts the linear prediction order and peak detection threshold sequentially according to the preset parameter set, and uses each adjusted set of parameters to re-extract formants from the corresponding stable speech segment. Thus, multiple sets of candidate structural acoustic parameters based on different extraction parameters are obtained for the same speech segment.

[0061] It should be noted that in this embodiment, when the system tries each set of parameters in the preset parameter set in turn, once the number of resonance peaks detected under a certain set of parameters has reached the preset number condition, it can stop trying the remaining parameter combinations in the parameter set and use the obtained multiple sets of candidate structural acoustic parameters as the basis for subsequent selection, thereby avoiding unnecessary repeated calculations after obtaining results that meet the conditions.

[0062] In this embodiment, after obtaining multiple sets of candidate structural acoustic parameters, the system determines the final set of results as preoperative or postoperative structural acoustic parameters based on the number of resonance peaks and their distribution along the frequency axis in each set of candidate results. Specifically, the system prioritizes selecting a set of candidate results with a relatively large number of resonance peaks and at least a preset number of effective resonance peaks in a relatively low frequency range. If multiple sets of candidate results meet the above conditions, the set that first meets the conditions can be selected according to the order in which the parameters in the parameter set were tried.

[0063] For example, when extracting a stable audio segment recorded in the early postoperative period, the system only detected two formants when extracting according to the initial extraction parameters, which did not meet the preset number condition. Subsequently, the system tried multiple combinations of linear prediction order and peak detection threshold in the parameter set. Under one set of parameters, three formants were detected, and the frequencies of two of the formants were within the preset low-frequency range. The system then used this set of candidate results as the final structural acoustic parameters and did not continue to try the other combinations in the parameter set.

[0064] In this embodiment, if the number of formants detected in the preoperative initial structural acoustic parameters or postoperative initial structural acoustic parameters has reached the preset number condition, it means that the current initial extraction parameters can extract a sufficient number of formants from the corresponding stable speech segment. At this time, there is no need to trigger the parameter adjustment and re-extraction process, and the system directly determines the initial structural acoustic parameters as the corresponding preoperative structural acoustic parameters or postoperative structural acoustic parameters.

[0065] In one embodiment, the step of comparing the preoperative acoustic parameters and the postoperative acoustic parameters to obtain the postoperative vocal cord rehabilitation assessment result includes: The structural dimension comparison results are obtained by calculating the difference between the structural acoustic parameters in the preoperative acoustic parameters and the structural acoustic parameters in the postoperative acoustic parameters. The difference between the functional acoustic parameters in the preoperative acoustic parameters and the functional acoustic parameters in the postoperative acoustic parameters is calculated, and it is determined whether the calculated difference meets the preset conditions to obtain the functional dimension judgment result. Based on the structural dimension comparison results and the functional dimension judgment results, the postoperative rehabilitation assessment results of the vocal cords are obtained.

[0066] In this embodiment, the preoperative acoustic parameters are extracted from the audio recorded before the patient's surgery and can serve as reference data reflecting the patient's own vocal baseline state; the postoperative acoustic parameters reflect the patient's vocal state at the corresponding time point after surgery. After obtaining the preoperative and postoperative acoustic parameters, the system uses the preoperative acoustic parameters as a reference benchmark to determine the changes in the postoperative acoustic parameters relative to the preoperative acoustic parameters from both structural and functional dimensions.

[0067] Based on the above explanation, the system first calculates the difference between the preoperative and postoperative structural acoustic parameters. Specifically, it subtracts the preoperative structural acoustic parameters from the postoperative structural acoustic parameters to obtain the difference in their formant frequencies, which is then used as the structural dimension comparison result. It should be noted that the structural dimension comparison result reflects the direction and magnitude of the postoperative vocal tract structure shift relative to the preoperative state, rather than simply the magnitude of the numerical difference. The system can further determine potential changes in vocal tract structure, such as opening and closing, and tongue position, based on the sign and magnitude of this difference.

[0068] In this embodiment, the system simultaneously calculates the difference between the functional acoustic parameters in the preoperative acoustic parameters and the functional acoustic parameters in the postoperative acoustic parameters to obtain the difference in voice quality. It then determines whether the difference meets preset conditions to obtain the functional dimension judgment result. It should be noted that, unlike the structural dimension comparison result, the functional dimension judgment result is not a simple numerical difference, but a categorical result obtained after conditional judgment, used to characterize whether there has been a clinically significant decline in postoperative voice quality relative to preoperative quality.

[0069] In this embodiment, after obtaining the structural dimension comparison results and the functional dimension judgment results respectively, the system uses both as components of the postoperative vocal cord rehabilitation assessment results. It should be noted that the structural dimension comparison results reflect changes in the vocal tract's geometric structure, while the functional dimension judgment results reflect whether there are abnormalities in the vocal cord vibration function. Both, from different perspectives, constitute the basis for assessing the patient's postoperative rehabilitation status. The postoperative vocal cord rehabilitation assessment results are the combined output of the structural dimension comparison results and the functional dimension judgment results.

[0070] In one embodiment, the step of calculating the difference between the functional acoustic parameters in the preoperative acoustic parameters and the functional acoustic parameters in the postoperative acoustic parameters, and determining whether the calculated difference meets a preset condition to obtain a functional dimension judgment result, includes: The change in functional acoustic parameters is obtained based on the difference between the functional acoustic parameters in the postoperative acoustic parameters and the functional acoustic parameters in the preoperative acoustic parameters. Determine whether the change in the functional acoustic parameter meets the preset descent condition to obtain the functional dimension judgment result, which includes positive judgment result and negative judgment result.

[0071] In this embodiment, functional acoustic parameters are used to characterize the regularity of vocal cord vibration during a patient's phonation. When vocal cord vibration is relatively regular and periodic, the corresponding functional acoustic parameter values ​​are higher; when vocal cord vibration exhibits abnormalities such as irregular closure or air leakage, the glottal noise component increases, and the corresponding functional acoustic parameter values ​​decrease accordingly. The system uses preoperative functional acoustic parameters as a reference to calculate the changes in postoperative functional acoustic parameters relative to preoperative values. Specifically, the changes in functional acoustic parameters are obtained by subtracting preoperative functional acoustic parameters from postoperative functional acoustic parameters. The sign and magnitude of these changes reflect the direction and amplitude of the change in the regularity of vocal cord vibration after surgery relative to preoperative values.

[0072] Based on the above description, after obtaining the change in functional acoustic parameters, the system determines whether the change meets the preset descent condition. It should be noted that in this embodiment, the preset descent condition is only applied to cases where the change in functional acoustic parameters is in a downward direction. That is, when the postoperative functional acoustic parameters are lower than the preoperative functional acoustic parameters, and the decrease reaches the level required by the preset descent condition, the change is determined to meet the preset descent condition; when the change in functional acoustic parameters does not show a downward trend, or the decrease does not reach the preset descent condition, the change is determined not to meet the preset descent condition.

[0073] In this embodiment, if the change in functional acoustic parameters meets the preset decrease condition, the system obtains a positive judgment result. The positive judgment result indicates that the regularity of the patient's vocal cord vibration after surgery has decreased clinically significant compared to before surgery, suggesting that there may be an abnormality in the vocal cord vibration function. If the change in functional acoustic parameters does not meet the preset decrease condition, the system obtains a negative judgment result. The negative judgment result indicates that the regularity of the patient's vocal cord vibration after surgery has not decreased clinically significant compared to before surgery.

[0074] In one embodiment, after adjusting the initial extraction parameters according to a preset parameter set, and re-extracting formants from the preoperative stable speech segment or the postoperative stable speech segment according to the adjusted extraction parameters to obtain multiple sets of candidate structural acoustic parameters, the method further includes: If the number of formants in the multiple sets of candidate structural acoustic parameters does not meet the preset number condition, the preoperative stable speech segment or the postoperative stable speech segment is pre-emphasized and processed in a periodic manner. Based on the predicted residual signal of the speech signal after periodic processing, the sampling point with the largest predicted residual amplitude in each pitch period is determined to obtain the glottal closure time information. Based on the glottal closure time information and preset position parameters, duration parameters and transition parameters, a non-attenuation segment, an attenuation segment and a transition segment connecting the non-attenuation segment and the attenuation segment are generated in each pitch cycle to obtain the main excitation weighting function. The non-attenuation segment corresponds to the sampling point near the glottal closure time, and the attenuation segment corresponds to the sampling point during the glottal opening phase. The prediction error is weighted according to the main excitation weighting function, and the linear prediction coefficients are solved according to the past and future directions of the sampling points based on the weighted prediction error to obtain the preoperative or postoperative vocal tract filter parameters. The preoperative structural acoustic parameters or the postoperative structural acoustic parameters are determined based on the vocal tract filter parameters.

[0075] In this embodiment, if, after traversing all sets of linear prediction orders and peak detection thresholds in the preset parameter set, the number of formants in the multiple sets of candidate structural acoustic parameters for a stable speech segment still does not meet the preset quantity condition, it indicates that the difficulty in extracting formants from this speech segment is not solely due to improper settings of the linear prediction order or peak detection threshold. Rather, it may be because the standard linear prediction analysis assigns the same weight to all sampling points within the speech frame during the calculation process, failing to distinguish the contribution of different stages within the glottal vibration period to the vocal tract model estimation. This causes the energy pulses generated by the glottal vibration itself to interfere with the solution of the vocal tract filter coefficients, masking the spectral envelope peaks that should be dominated by the vocal tract resonance characteristics. In this case, the system no longer adjusts the linear prediction order or peak detection threshold, but instead uses a weighted linear prediction method that suppresses glottal vibration interference to re-obtain the formant parameters.

[0076] In this embodiment, the system first pre-emphasizes the preoperative or postoperative stable speech segments that trigger the processing branch to increase the proportion of high-frequency energy, which facilitates the subsequent resolution of formants. Based on the pre-emphasis processing, the system further performs periodic processing on the speech signal, that is, according to the fundamental frequency period in the signal, the signal is divided into multiple sequentially arranged fundamental frequency period segments, each of which corresponds to a complete cycle of vocal cord vibration.

[0077] In this embodiment, the system calculates the prediction residual signal of the speech signal after pre-emphasis and periodic processing. The prediction residual signal refers to the difference sequence between the original signal and the estimated signal obtained by linear prediction after linear prediction analysis of the speech signal. This difference sequence usually shows a significant increase in amplitude at the moment of glottal closure. This is because the vocal cords generate a concentrated excitation pulse at the moment of closure, which is difficult to accurately estimate by the linear prediction model, thus manifesting as a sudden increase in amplitude in the prediction residual signal at the corresponding moment. Based on this phenomenon, the system finds the sampling point position corresponding to the maximum amplitude of the prediction residual signal in each pitch period segment, and determines the sampling point position as the glottal closure moment in that pitch period. The system processes each pitch period segment in the speech signal in this way to obtain glottal closure moment information covering the entire speech signal.

[0078] For example, if the predicted residual signal corresponding to a certain pitch period segment is relatively stable at the beginning of the period, but the value at a certain sampling point within the period is significantly higher than that of the adjacent sampling points, the system will determine the position corresponding to that sampling point as the glottal closure time of that period. This process is repeated to determine the glottal closure time of each pitch period in the speech signal, thus obtaining the glottal closure time information.

[0079] In this embodiment, after obtaining the glottal closure time information, the system further constructs a main excitation weighting function for subsequent weighted linear prediction. Specifically, a complete cycle of vocal cord vibration can be divided into a glottal closure phase and a glottal opening phase in time. During the glottal closure phase, the vocal tract is approximately closed, and the acquired speech signal is mainly determined by the resonance characteristics of the vocal tract itself, with little direct influence from the glottal excitation source. However, during the glottal opening phase, the glottis is open, and factors such as turbulence noise of the glottal airflow will have additional effects on the signal acquired during this phase. If the sampling points of the opening phase and the sampling points of the closing phase are treated equally during the solution of the vocal tract filter coefficients, the estimation results of the vocal tract model are easily interfered with by the signal of the opening phase.

[0080] Based on the above description, within each pitch period, the system uses the glottal closure moment corresponding to that period as a reference position and determines the starting position of a non-attenuation segment according to preset position parameters. This non-attenuation segment is located near the glottal closure moment and is used to retain the original weights of the sampling points in this segment during subsequent prediction error weighting. The duration of this non-attenuation segment is determined according to preset duration parameters. In the region outside the non-attenuation segment, a transition segment connecting the non-attenuation segment and the attenuation segment is determined according to preset transition parameters. The weights in the transition segment gradually decrease from the non-attenuation segment side to the attenuation segment side. The remaining region outside the non-attenuation segment and the transition segment is determined as the attenuation segment. The attenuation segment corresponds to the sampling points during the glottal opening phase, and these sampling points are assigned lower weights during subsequent prediction error weighting. The system generates non-attenuation segments, attenuation segments, and transition segments in each pitch period in the above manner, and sequentially connects the weight values ​​corresponding to each pitch period to obtain the main excitation weighting function covering the entire speech signal.

[0081] It should be noted that the preset position parameter, duration parameter, and transition parameter control the starting position of the non-attenuation segment within the period, the relative length of the non-attenuation segment, and the relative duration of the transition between the attenuation segment and the non-attenuation segment, respectively. These three parameters collectively determine the specific shape of the main excitation weighting function within each pitch period. For example, setting the position parameter so that the starting position of the non-attenuation segment is slightly earlier than the glottal closure time allows the non-attenuation segment to cover sampling points before and after the glottal closure time. Setting the duration parameter to a smaller value allows the non-attenuation segment to cover only a narrow time range near the glottal closure time, thus enabling the weighting function to attenuate sampling points in the open-phase phase over a wider range.

[0082] In this embodiment, after obtaining the main excitation weighting function, the system uses this function to weight the prediction error. That is, in the error criterion on which the linear prediction coefficient is calculated, the squared value of the prediction error at each sampling point is multiplied by the value of the main excitation weighting function corresponding to that position and then accumulated. This makes the prediction error near the glottal closure moment account for a relatively higher proportion of the total error, while the contribution of the prediction error during the glottal opening phase to the total error is reduced accordingly.

[0083] In this embodiment, when solving for the linear prediction coefficients of the weighted prediction error, the system predicts each sampling point in the speech signal according to both its past and future directions. That is, it predicts using several sampling points before and after the current sampling point, incorporating the weighted prediction errors calculated in both directions into the error criterion for coefficient solving. The resulting vocal tract filter parameters utilize a larger number of effective sampling points within the analysis window compared to results obtained using only a single direction of prediction. The system performs the above weighted linear prediction process on preoperative and postoperative stable speech segments respectively to obtain preoperative or postoperative vocal tract filter parameters.

[0084] In this embodiment, the system calculates the corresponding spectral envelope based on the obtained duct filter parameters, detects the location of the local maximum amplitude within the spectral envelope, and determines the frequencies corresponding to each detected local maximum amplitude as the formant frequencies, thereby determining the preoperative or postoperative structural acoustic parameters. It should be noted that, since the master excitation weighting function has already reduced the influence of the sampling points during the glottal opening phase on the results during the solution of the duct filter parameters, compared to the spectral envelope obtained by standard linear predictive analysis which directly processes all sampling points with equal weights, the formant peaks in the spectral envelope obtained in this embodiment, which were originally difficult to detect due to glottal excitation interference, are more easily detected at the preset peak detection threshold after the degree of interference is reduced.

[0085] It should be understood that the sequence number of each step in the above embodiments does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0086] In one embodiment, a postoperative vocal cord rehabilitation assessment device is provided, which corresponds one-to-one with the postoperative vocal cord rehabilitation assessment methods described in the above embodiments. For example... Figure 6 As shown, the vocal cord postoperative rehabilitation assessment device includes: The preprocessing module 601 is used to preprocess the preoperative audio signal and the postoperative audio signal respectively to obtain the preoperative preprocessed speech signal and the postoperative preprocessed speech signal. The stable extraction module 602 is used to extract stable speech segments from the pre-processed speech signal and the post-processed speech signal respectively to obtain pre-operative stable speech segments and post-operative stable speech segments. The parameter extraction module 603 is used to extract acoustic parameters from the preoperative stable speech segment and the postoperative stable speech segment respectively to obtain preoperative acoustic parameters and postoperative acoustic parameters. The comparison and evaluation module 604 is used to compare the preoperative acoustic parameters and the postoperative acoustic parameters to obtain the postoperative rehabilitation evaluation results of the vocal cords.

[0087] It should be noted that the information interaction and execution process between the above-mentioned devices / units are based on the same concept as the method embodiments of this application. For details on their specific functions and technical effects, please refer to the method embodiments section, and they will not be repeated here.

[0088] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the above-described division of functional units and modules is merely an example. In practical applications, the above functions can be assigned to different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above. The functional units and modules in the embodiments can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit. Furthermore, the specific names of the functional units and modules are only for easy differentiation and are not intended to limit the scope of protection of this application. The specific working process of the units and modules in the above system can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0089] This application also provides a computer device, such as... Figure 7 As shown, the computer device includes: at least one processor, a memory, and a computer program stored in the memory and executable on the at least one processor. When the processor executes the computer program, it implements the steps in any of the above method embodiments, or when the processor executes the computer program, it implements the functions of each module / unit in the above device embodiments.

[0090] For example, the computer program may be divided into one or more modules / units, which are stored in the memory and executed by the processor to complete this application. The one or more modules / units may be a series of computer program instruction segments capable of performing a specific function, which describe the execution process of the computer program in the computer device.

[0091] Those skilled in the art will understand that Figure 7 The computer device described is merely an example and does not constitute a limitation on the computer device. It may include more or fewer components than shown, or combine certain components, or different components. For example, the computer device may also include input / output devices, network access devices, buses, etc.

[0092] The aforementioned processor can be a Central Processing Unit (CPU), or other general-purpose processors, Digital Signal Processors (DSPs), Application Specific Integrated Circuits (ASICs), or Field Programmable Gate Arrays (FPGAs). Programmable Gate Array (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor can be a microprocessor or any conventional processor.

[0093] The memory can be an internal storage unit of the computer device, such as a hard drive or RAM. The memory can also be an external storage device of the computer device, such as a plug-in hard drive, Smart Media Card (SMC), Secure Digital (SD) card, or Flash Card. Furthermore, the memory can include both internal and external storage units of the computer device.

[0094] This application also provides a readable storage medium storing a computer program that, when executed by a processor, implements the steps described in the various method embodiments above.

[0095] This application provides a computer program product that, when run on an electronic device, enables the electronic device to perform the steps described in the various method embodiments above.

[0096] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, all or part of the processes in the methods of the above embodiments of this application can be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. The computer-readable medium can include at least: any entity or device capable of carrying computer program code to a photographing device / terminal device, a recording medium, a computer memory, a read-only memory (ROM), a random access memory (RAM), an electrical carrier signal, a telecommunication signal, and a software distribution medium. Examples include USB flash drives, portable hard drives, magnetic disks, or optical disks. In some jurisdictions, according to legislation and patent practice, computer-readable media cannot be electrical carrier signals or telecommunication signals.

[0097] In the above embodiments, the descriptions of each embodiment have different focuses. For parts that are not described in detail or recorded in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0098] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0099] In the embodiments provided in this application, it should be understood that the disclosed apparatus / devices and methods can be implemented in other ways. For example, the apparatus / device embodiments described above are merely illustrative. For instance, the division of modules or units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or units may be electrical, mechanical, or other forms.

[0100] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0101] The above-described embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application, and should all be included within the protection scope of this application.

Claims

1. A method for assessing postoperative rehabilitation of the vocal cords, characterized in that, The method for assessing postoperative rehabilitation of the vocal cords: Pre-processing was performed on the pre-operative and post-operative audio signals to obtain pre-processed pre-operative and post-operative pre-processed audio signals. Stable speech segments are extracted from the preoperative preprocessed speech signal and the postoperative preprocessed speech signal to obtain preoperative stable speech segments and postoperative stable speech segments. Acoustic parameters were extracted from the preoperative stable speech segment and the postoperative stable speech segment to obtain the preoperative acoustic parameters and postoperative acoustic parameters. The results of the vocal cord postoperative rehabilitation assessment were obtained by comparing the preoperative acoustic parameters and the postoperative acoustic parameters.

2. The method for postoperative rehabilitation assessment of vocal cords according to claim 1, characterized in that, The preprocessing of preoperative and postoperative audio signals to obtain preoperative preprocessed speech signals and postoperative preprocessed speech signals includes: The preoperative audio signal and the postoperative audio signal are respectively processed by channel merging to obtain the preoperative mono signal and the postoperative mono signal. The preoperative mono signal and the postoperative mono signal are normalized according to a preset reference amplitude to obtain the preoperative normalized signal and the postoperative normalized signal. The preoperative normalized signal and the postoperative normalized signal are resampled according to a preset target sampling rate to obtain the preoperative preprocessed speech signal and the postoperative preprocessed speech signal.

3. The method for postoperative rehabilitation assessment of vocal cords according to claim 1, characterized in that, The step of extracting stable speech segments from the preoperative preprocessed speech signal and the postoperative preprocessed speech signal to obtain preoperative stable speech segments and postoperative stable speech segments includes: The pre-processed speech signal and the post-processed speech signal are processed by framing to obtain multiple speech frames. The speech frames are then filtered according to the energy value and zero-crossing rate of each speech frame to obtain pre-operative candidate speech segments and post-operative candidate speech segments. Formant parameters were extracted from the preoperative candidate speech segments and the postoperative candidate speech segments to obtain the preoperative candidate formant trajectories and the postoperative candidate formant trajectories. Based on the fluctuation degree of the preoperative candidate formant trajectory and the postoperative candidate formant trajectory, preoperative stable speech segments and postoperative stable speech segments are determined from the preoperative candidate speech segments and the postoperative candidate speech segments, respectively.

4. The method for postoperative rehabilitation assessment of vocal cords according to claim 3, characterized in that, The step of determining the preoperative stable speech segment and the postoperative stable speech segment from the preoperative candidate speech segment and the postoperative candidate speech segment respectively, based on the fluctuation degree of the preoperative candidate formant trajectory and the postoperative candidate formant trajectory, includes: Sliding window calculations were performed on the preoperative candidate resonance peak trajectories and the postoperative candidate resonance peak trajectories to obtain the preoperative fluctuation parameter sequence and the postoperative fluctuation parameter sequence. Based on the preoperative fluctuation parameter sequence and the postoperative fluctuation parameter sequence, a continuous stable window subset is determined from the preoperative candidate speech segment and the postoperative candidate speech segment respectively to obtain the candidate stable speech segment; If the duration of the candidate stable speech segment does not meet the preset conditions, the window position with the smallest fluctuation degree is determined according to the preoperative fluctuation parameter sequence or the postoperative fluctuation parameter sequence, or the center position of the preoperative candidate speech segment or the postoperative candidate speech segment is determined, so as to obtain the preoperative stable speech segment or the postoperative stable speech segment. If the duration of the candidate stable speech segment meets the preset condition, the candidate stable speech segment is determined as the preoperative stable speech segment or the postoperative stable speech segment.

5. The method for postoperative rehabilitation assessment of vocal cords according to claim 1, characterized in that, The acoustic parameters extracted from the preoperative stable speech segment and the postoperative stable speech segment are respectively obtained as follows: Based on the initial extraction parameters, formant extraction was performed on the preoperative stable speech segment and the postoperative stable speech segment to obtain the preoperative initial structural acoustic parameters and the postoperative initial structural acoustic parameters. If the number of formants in the preoperative initial structural acoustic parameters or the postoperative initial structural acoustic parameters does not meet the preset number condition, the initial extraction parameters are adjusted according to the preset parameter set, and the formants are re-extracted from the preoperative stable speech segment or the postoperative stable speech segment according to the adjusted extraction parameters to obtain multiple sets of candidate structural acoustic parameters. Based on the number of resonance peaks and the frequency distribution of resonance peaks in the multiple sets of candidate structural acoustic parameters, preoperative or postoperative structural acoustic parameters are determined from the multiple sets of candidate structural acoustic parameters. If the number of resonant peaks in the preoperative initial structural acoustic parameters or the postoperative initial structural acoustic parameters meets the preset number condition, the preoperative initial structural acoustic parameters or the postoperative initial structural acoustic parameters are determined as the preoperative structural acoustic parameters or the postoperative structural acoustic parameters. Functional dimension parameters were extracted from the preoperative stable speech segment and the postoperative stable speech segment to obtain preoperative functional acoustic parameters and postoperative functional acoustic parameters. The preoperative acoustic parameters are obtained based on the preoperative structural acoustic parameters and the preoperative functional acoustic parameters, and the postoperative acoustic parameters are obtained based on the postoperative structural acoustic parameters and the postoperative functional acoustic parameters.

6. The method for postoperative rehabilitation assessment of vocal cords according to claim 1, characterized in that, The comparison of the preoperative acoustic parameters and the postoperative acoustic parameters to obtain the postoperative vocal cord rehabilitation assessment results includes: The structural dimension comparison results are obtained by calculating the difference between the structural acoustic parameters in the preoperative acoustic parameters and the structural acoustic parameters in the postoperative acoustic parameters. The change in functional acoustic parameters is obtained based on the difference between the functional acoustic parameters in the postoperative acoustic parameters and the functional acoustic parameters in the preoperative acoustic parameters. Determine whether the change in the functional acoustic parameter meets the preset descent condition to obtain the functional dimension judgment result, which includes a positive judgment result and a negative judgment result. Based on the structural dimension comparison results and the functional dimension judgment results, the postoperative rehabilitation assessment results of the vocal cords are obtained.

7. A vocal cord surgery rehabilitation assessment device, characterized in that, include: The preprocessing module is used to preprocess the preoperative audio signal and the postoperative audio signal respectively to obtain the preoperative preprocessed speech signal and the postoperative preprocessed speech signal. The stable extraction module is used to extract stable speech segments from the pre-operative pre-processed speech signal and the post-operative pre-processed speech signal respectively, to obtain pre-operative stable speech segments and post-operative stable speech segments. The parameter extraction module is used to extract acoustic parameters from the preoperative stable speech segment and the postoperative stable speech segment respectively, so as to obtain the preoperative acoustic parameters and the postoperative acoustic parameters. The comparative evaluation module is used to compare the preoperative acoustic parameters and the postoperative acoustic parameters to obtain the postoperative rehabilitation evaluation results of the vocal cords.

8. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the postoperative vocal cord rehabilitation assessment method as described in any one of claims 1 to 6.

9. A readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the steps of the postoperative vocal cord rehabilitation assessment method as described in any one of claims 1 to 6.

10. A computer program product, characterized in that, The computer program product includes a computer program that, when executed by a processor, enables the implementation of the steps of the postoperative vocal cord rehabilitation assessment method as described in any one of claims 1 to 6.

Images