Throat lesion monitoring system based on voice recognition

Abstract

The invention relates to the technical field of sound analysis, in particular to a throat lesion monitoring system based on sound recognition, which comprises a sound quality optimization module, a sound feature analysis module, an abnormality evaluation module and a treatment effect evaluation module. According to the invention, by utilizing the historical sound data of the user and comparing and evaluating the deviation of the current sound features, the monitoring accuracy is improved, the detection of early lesions is more reliable, the deviation detection threshold is adjusted according to the age and gender differences of individuals, the sound anomaly detection is more accurate, the sound anomaly of the user can be found in time, and the user experience is improved. The system has the advantages that the system is simple in structure and convenient to use, false alarm and missing alarm are reduced, real-time and visual treatment feedback is provided by collecting sound data before and after treatment, analyzing the sound change degree, combining the treatment time and evaluating the treatment effect, data support is provided for doctors, measure adjustment in the treatment process is optimized, and the disease monitoring and management fineness and effect are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of voice analysis, and in particular to a laryngeal lesion monitoring system based on voice recognition. Background Art

[0002] The technical field of voice analysis covers all steps from the acquisition, processing of voice signals to feature extraction and recognition. The technical field uses various algorithms to analyze voice data, so as to identify speech content, emotional state, identity features and other relevant information. During the processing, the voice signal is preprocessed to eliminate noise, and then key features of the voice are extracted through feature extraction techniques such as frequency analysis and harmonic analysis to extract useful information. In the medical field, voice analysis technology is applied to diagnose related diseases, such as detecting respiratory diseases by analyzing cough sounds.

[0003] Among them, the laryngeal lesion monitoring system is a system that uses voice analysis technology to detect and monitor the health status of the vocal cords and larynx. The system judges the health status of the larynx by analyzing the changes in the voice characteristics of the user, such as pitch and volume, and identifies signs of laryngeal lesions, such as abnormal voice frequency, voice quality changes and difficulty in pronunciation. The main uses include early diagnosis of conditions such as vocal cord inflammation, vocal cord tumors or chronic laryngitis, and assisting doctors in evaluating the treatment effect and disease progression.

[0004] Traditional monitoring systems lack consideration of individual differences in voice feature analysis, resulting in poor generalization of diagnosis, inability to provide accurate health status assessment for specific populations, prone to false positives and false negatives, difficult to immediately detect voice abnormalities of users and give corresponding reminders. In terms of evaluating the treatment effect, there is a lack of systematic comparison of voice data before and after treatment, making it difficult for doctors to accurately judge the treatment effect, thus affecting the timely adjustment of treatment strategies. As a result, it is difficult for patients to obtain the most suitable treatment advice for their condition, affecting the control and recovery speed of the condition. Summary of the Invention

[0005] The purpose of the present invention is to solve the deficiencies existing in the prior art, and a laryngeal lesion monitoring system based on voice recognition is proposed.

[0006] In order to achieve the above purpose, the present invention adopts the following technical solution: A laryngeal lesion monitoring system based on voice recognition, the system includes:

[0007] The voice quality optimization module, based on a voice capture device, collects the voice data of the user, and adjusts the voice filtering parameters according to the voice acquisition environment, performs voice filtering, removes environmental interference sounds, and obtains optimized voice data;

[0008] Based on the optimized voice data, the voice feature analysis module extracts the average pitch and volume features of the voice data by analyzing the user's historical voice data, compares the current voice data with the historical average voice features, evaluates the voice feature deviation, and obtains the voice feature deviation data;

[0009] Based on the voice feature deviation data, the anomaly evaluation module adjusts the deviation detection threshold according to the user's age and gender, compares the current voice feature deviation with the adjusted deviation detection threshold, evaluates the voice anomaly level, gives a laryngeal anomaly reminder to the user, and obtains the anomaly evaluation result;

[0010] Based on the anomaly evaluation result, the treatment effect evaluation module collects the voice data of the user before and after receiving laryngeal treatment, analyzes the degree of voice change, and combines the treatment duration to evaluate the laryngeal treatment effect, and obtains the evaluation result of the laryngeal lesion treatment effect.

[0011] The improvement of the present invention is that the acquisition steps of the optimized voice data are as follows:

[0012] Based on the voice capture device, collect the user's voice data and record the environmental noise data to obtain the basic voice data;

[0013] Based on the basic voice data, through the formula:

[0014]

[0015] Calculate the high-pass filter threshold θ, where N avg is the average value of the environmental noise level, N var is the variance of the noise level, N max is the maximum value of the noise level, N min is the minimum value of the noise level, α θ 、β θ and γ θ are adjustment coefficients, k θ is a stability constant, θ is the high-pass filter threshold, and e is the base of the natural logarithm;

[0016] Based on the high-pass filter threshold θ and the basic voice data, remove the part of the user's voice data that is lower than the high-pass filter threshold θ to obtain the optimized voice data.

[0017] The improvement of the present invention is that the method for extracting the average pitch and volume features of the voice data is as follows:

[0018] Based on the user voice data record, extract the voice data within a certain period of time, including voice pitch and volume, to obtain the user's historical voice data;

[0019] Based on the user's historical voice data, through the formula:

[0020]

[0021] and

[0022]

[0023] Calculate the historical average pitch P avg and the historical average volume V avg , and obtain the historical sound feature extraction result, where n is the total number of sound samples, P avg is the historical average pitch, V avg is the historical average volume, p i is the pitch of the i-th sound sample, v i is the volume of the i-th sound sample.

[0024] The improvement of the present invention is that the step of obtaining the sound feature deviation data is as follows:

[0025] Based on the historical sound feature extraction result and the optimized sound data, extract the average pitch and volume of the current user's sound to obtain sound deviation correlation data;

[0026] Based on the sound deviation correlation data, through the formula:

[0027]

[0028] Calculate the sound feature deviation index D to obtain the sound feature deviation data, where λ D is the weight factor of the pitch difference, κ D is the influence coefficient of the volume difference, P current and P avg represent the current and historical average pitches respectively, V current and V avg represent the current and historical average volumes respectively, and D is the sound feature deviation index.

[0029] The improvement of the present invention is that the method for adjusting the deviation detection threshold is as follows:

[0030] Collect the basic information of the user, including age and gender, and extract the basic deviation detection threshold to obtain the threshold correlation data;

[0031] Based on the threshold correlation data, through the formula:

[0032]

[0033] Calculate the adjusted deviation detection threshold T to obtain the deviation detection threshold adjustment result, where T base is the basic threshold, A is the age of the user, A baseis the age baseline, G mod is the gender modulus, γ T is the linear influence coefficient, δ T and η T are the non - linear influence coefficients, β T is the weight coefficient, e is the base of the natural logarithm, and T is the adjusted deviation detection threshold.

[0034] The improvement of the present invention is that the step of obtaining the abnormal evaluation result is as follows:

[0035] Based on the voice feature deviation data and the deviation detection threshold adjustment result, through the formula:

[0036]

[0037] Calculate the voice abnormality level L, where D is the voice feature deviation index, T is the adjusted deviation detection threshold, S L is the scale factor, ρ L is the dynamic adjustment parameter;

[0038] Based on the voice abnormality level L, compare it with the preset abnormal level table, and send a corresponding laryngeal abnormality reminder to the user to obtain the abnormal evaluation result.

[0039] The improvement of the present invention is that the method for analyzing the degree of voice change is as follows:

[0040] Collect the voice samples of the user before and after laryngeal treatment, including the volume and pitch before and after treatment, to obtain the voice data before and after treatment;

[0041] Based on the voice data before and after treatment, through the formula:

[0042]

[0043] Calculate the voice change index Z to obtain the voice change degree evaluation result, where P pre and P post respectively represent the average pitch before and after treatment, V pre and V post respectively represent the average volume before and after treatment, ω P and ω V are the weight coefficients, and Z is the voice change index.

[0044] The improvement of the present invention is that the step of obtaining the laryngeal lesion treatment effect evaluation result is as follows:

[0045] Based on the voice change degree evaluation result, extract the treatment duration according to the treatment start time and treatment end time;

[0046] Based on the treatment duration, through the formula:

[0047]

[0048] Calculate the laryngeal treatment effect index Y, where Z is the voice change index, T Y is the total time spent on treatment, k Y is the influence coefficient of voice change, λ Y is the time influence coefficient, Y is the laryngeal treatment effect index, and e is the base of the natural logarithm;

[0049] Based on the laryngeal treatment effect index Y, compare it with the target treatment effect index to determine whether the treatment effect meets the expectation, and obtain the evaluation result of the laryngeal lesion treatment effect.

[0050] Compared with the prior art, the advantages and positive effects of the present invention are as follows:

[0051] In the present invention, by using the user's historical voice data, comparing and evaluating the deviation of the current voice characteristics, the monitoring accuracy is improved, the detection of early lesions is made more reliable, and according to the age and gender differences of individuals, the deviation detection threshold is adjusted to make the voice abnormality detection more accurate, the user's voice abnormality can be detected in time, reducing false alarms and missed reports. By collecting the voice data before and after treatment, analyzing the degree of voice change, and combining with the treatment time, the treatment effect is evaluated, providing real-time and intuitive treatment feedback, and providing data support for doctors, optimizing the measure adjustment in the treatment process, and improving the fineness and effect of disease monitoring and management. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] Figure 1 is the system flowchart of the present invention;

[0053] Figure 2 is the flowchart of obtaining optimized voice data of the present invention;

[0054] Figure 3 is the flowchart of extracting the average pitch and volume characteristics of voice data of the present invention;

[0055] Figure 4 is the flowchart of obtaining voice feature deviation data of the present invention;

[0056] Figure 5 is the flowchart of adjusting the deviation detection threshold of the present invention;

[0057] Figure 6 is the flowchart of obtaining the abnormal evaluation result of the present invention;

[0058] Figure 7 is the flowchart of analyzing the degree of voice change of the present invention;

[0059] Figure 8The flowchart for obtaining the evaluation result of the treatment effect of laryngeal lesions in the present invention. Detailed implementation manners

[0060] In order to make the objectives, technical solutions and advantages of the present invention more clear and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0061] In the description of the present invention, it should be understood that the orientation or positional relationship indicated by the terms "length", "width", "upper", "lower", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", etc. is based on the orientation or positional relationship shown in the accompanying drawings. It is only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and thus cannot be construed as a limitation of the present invention. In addition, in the description of the present invention, the meaning of "a plurality of" is two or more unless otherwise specifically defined.

[0062] Please refer to Figure 1 , the present invention provides a technical solution: a laryngeal lesion monitoring system based on voice recognition, and the system includes:

[0063] The voice quality optimization module collects the voice data of the user based on the voice capture device, and adjusts the voice filtering parameters according to the voice acquisition environment to perform voice filtering, remove the environmental interference sound, and obtain the optimized voice data;

[0064] The voice feature analysis module extracts the average pitch and volume features of the voice data by analyzing the historical voice data of the user based on the optimized voice data, compares the current voice data with the historical average voice features, evaluates the voice feature deviation, and obtains the voice feature deviation data;

[0065] The abnormality evaluation module adjusts the deviation detection threshold according to the age and gender of the user based on the voice feature deviation data, compares the current voice feature deviation with the adjusted deviation detection threshold, evaluates the voice abnormality level, gives a laryngeal abnormality reminder to the user, and obtains the abnormality evaluation result;

[0066] The treatment effect evaluation module collects the voice data of the user before and after receiving laryngeal treatment based on the abnormality evaluation result, analyzes the degree of voice change, and combines the treatment duration to evaluate the laryngeal treatment effect, and obtains the evaluation result of the treatment effect of laryngeal lesions;

[0067] The optimized voice data includes the filtered audio signal and the audio waveform. The voice feature deviation data includes the fundamental frequency change, the amplitude deviation, and the formant shift. The abnormal evaluation result includes the abnormal level, the potential lesion type, and the laryngeal health rating. The evaluation result of the laryngeal lesion treatment effect includes the comparison result of the voice features before and after treatment, the voice recovery index, and the treatment effect assessment.

[0068] Please refer to Figure 2 , the steps for obtaining the optimized voice data are as follows:

[0069] Based on the voice capture device, collect the user's voice data and record the environmental noise data to obtain the basic voice data;

[0070] Based on the basic voice data, through the formula:

[0071]

[0072] Calculate the high-pass filter threshold θ, where N avg is the average value of the environmental noise level, N var is the variance of the noise level, N max is the maximum value of the noise level, N min is the minimum value of the noise level, α θ , β θ and γ θ are adjustment coefficients, k θ is the stability constant, θ is the high-pass filter threshold, and e is the base of the natural logarithm;

[0073] Based on the high-pass filter threshold θ and the basic voice data, remove the part of the user's voice data that is lower than the high-pass filter threshold θ to obtain the optimized voice data.

[0074] Formula:

[0075]

[0076] The meaning and acquisition method of the parameters

[0077] N avg is the average value of the environmental noise level, usually obtained by performing a series of voice measurements in a specific environment and calculated as the arithmetic mean of the collected noise data.

[0078] N var is the variance of the noise level, indicating the degree of fluctuation of the voice data, and is obtained by calculating the average of the squares of the differences between all voice measurement values and their average value.

[0079] N max and N minThey are the maximum and minimum values of the noise level during measurement, determined by recording the highest and lowest points among all the measured noise data within the same time period.

[0080] α θ , β θ and γ θ are pre-set adjustment coefficients used to adjust the sensitivity and response of the filtering threshold according to the specific conditions of the environment. These coefficients are usually determined based on experimental or historical data analysis to optimize the system's adaptability to different environmental noises.

[0081] k θ is the stability constant used to ensure that the input of the logarithmic function is never zero, usually set to 1 to ensure the stability of mathematical operations.

[0082] Calculation example

[0083] Set the environmental noise data set: 50dB, 55dB, 60dB, 50dB, 65dB.

[0084] Calculate the average value N avg

[0085]

[0086] Calculate the variance N var

[0087]

[0088] N max = 65dB, N min = 50dB, set the adjustment coefficient α θ = 0.5, β θ = 0.2, γ θ = 0.005, k θ = 1.

[0089] Calculate the high-pass filter threshold θ

[0090]

[0091] The calculated θ = 55.63dB, representing the adjusted high-pass filter threshold according to the statistical characteristics of the current environmental noise. This threshold will be used in the sound filtering system to remove background noise below the normal frequency range of human voices, thus optimizing the clarity and quality of sound data.

[0092] Please refer to Figure 3 , the method for extracting the average pitch and volume characteristics of sound data is:

[0093] Based on the user's voice data recording, extract the voice data within a certain period of time, including the voice pitch and volume, to obtain the user's historical voice data;

[0094] Based on the user's historical voice data, through the formula:

[0095]

[0096] and

[0097]

[0098] Calculate the historical average pitch P avg and the historical average volume V avg , to obtain the historical voice feature extraction result, where n is the total number of voice samples, P avg is the historical average pitch, V avg is the historical average volume, p i is the pitch of the i-th voice sample, v i the volume of the i-th voice sample.

[0099] Formula:

[0100]

[0101] and

[0102]

[0103] Meaning and acquisition method of parameters:

[0104] p i The pitch of the i-th voice sample, usually obtained by performing spectral analysis on the voice sample. In practical applications, this can be achieved through Fourier transform, converting the voice sample from the time domain to the frequency domain and extracting the main frequency component as the pitch.

[0105] v i The volume of the i-th voice sample. The calculation of volume is usually based on the average value or maximum value of the sound wave amplitude, and can be determined by measuring the waveform amplitude of the voice sample.

[0106] n The total number of voice samples, determined by counting the number of all valid voice samples collected within the specified time period.

[0107] Calculation example:

[0108] Suppose there are four voice samples, with pitches of 440Hz, 450Hz, 430Hz, 440Hz respectively, and volumes of 70dB, 72dB, 68dB, 71dB respectively.

[0109] Calculate the average pitch P avg

[0110]

[0111] Calculate the average volume V avg

[0112] The calculation results show that the average pitch obtained from the collected samples is 440 Hz and the average volume is 70.25 dB, indicating that the volume level of these sound samples is moderate, and the data provides typical characteristics of the sound samples.

[0113] Please refer to Figure 4 , and the steps for obtaining the sound feature deviation data are as follows:

[0114] Based on the historical sound feature extraction results and the optimized sound data, extract the average pitch and volume of the current user's voice to obtain the sound deviation correlation data;

[0115] Based on the sound deviation correlation data, through the formula:

[0116]

[0117] Calculate the sound feature deviation index D to obtain the sound feature deviation data, where λ D is the weight factor of the pitch difference, κ D is the influence coefficient of the volume difference, P current and P avg represent the current and historical average pitches respectively, V current and V avg represent the current and historical average volumes respectively, and D is the sound feature deviation index.

[0118] Formula:

[0119]

[0120] Meaning and acquisition method of parameters:

[0121] P current and P avg represent the pitch of the currently collected sound and the average pitch in the historical data respectively. The value is calculated by the sound analysis software, where P current is extracted from the real-time data, and P avg is the average value obtained by statistical analysis of the historical data set.

[0122] V current and V avg refer to the volume of the current sound and the historical average volume respectively. The volume value can be measured by a sound level meter or automatically calculated from the recording using audio processing software.

[0123] λD is a weight factor used to adjust the contribution of pitch differences in the total deviation calculation. It can be adjusted according to different application scenarios based on empirical or voice analysis requirements.

[0124] κ D is an influence coefficient used to adjust the sensitivity of volume differences. By adjusting this parameter, the

[0125]

[0126] proportion in the logarithmic transformation can be controlled so that the impact of volume differences on the total deviation is not too drastic.

[0127] Calculation example:

[0128] Set the following data: current average pitch P current = 450Hz, historical average pitch P avg = 440Hz, current average volume V current = 75dB, historical average volume V avg = 70dB, weight factor λ D = 0.5, influence coefficient κ D = 0.1.

[0129] Pitch difference:

[0130] P current - P avg = 450 - 440 = 10Hz

[0131] Logarithmic transformation of volume difference:

[0132] log(1 + κ D ·|V current - V avg |)

[0133] = log(1 + 0.1·|75 - 70|)

[0134] = log(1.5)

[0135] ≈0.176

[0136] Calculation of deviation index D:

[0137]

[0138] The calculated deviation index D is approximately 10.00, indicating that there is a significant difference between the current pitch and the historical average. The volume difference processed in logarithmic form contributes relatively little to the total deviation, which can balance the influence of pitch and volume in the sound feature evaluation and ensure that sound changes are appropriately reflected without overamplifying a single feature.

[0139] Please refer to Figure 5 , the method for adjusting the deviation detection threshold is as follows:

[0140] Collect the basic information of users, including age and gender, and extract the basic deviation detection threshold to obtain threshold correlation data;

[0141] Based on the threshold correlation data, through the formula:

[0142]

[0143] Calculate the adjusted deviation detection threshold T to obtain the deviation detection threshold adjustment result, where T base is the basic threshold, A is the age of the user, A base is the age baseline, G mod is the gender modulus, γ T is the linear influence coefficient, δ T and η T are the non - linear influence coefficients, β T is the weight coefficient, e is the base of the natural logarithm, and T is the adjusted deviation detection threshold.

[0144] Formula:

[0145]

[0146] Meaning and acquisition method of parameters:

[0147] T base The basic deviation detection threshold, usually obtained from extensive analysis of sound health data, serves as the starting comparison point for all users.

[0148] A The actual age of the user, directly provided by the user during device or service setup.

[0149] A base The age baseline, usually determined based on sound health research data, identifies the key turning age points for sound characteristic changes.

[0150] G mod The gender modulus, determined by the gender information reported by the user, is used to adjust the threshold to adapt to gender differences.

[0151] γ T , δ T , η T and β T The coefficients are determined through data analysis and previous research. γ T Controls the direct linear influence of age differences on the threshold, δ T and η T Adjust the non - linear part to simulate the decreasing rate of the impact of age growth on sound characteristic changes, βT Adjust the direct impact of gender and set it according to the calculation requirements.

[0152] Calculation example:

[0153] Set the following parameters: T base = 10, A = 25, the current user's age, A base = 20, the age baseline, G mod = 1, indicating that the user is male, γ T = 0.5, δ T = 0.3, η T = 0.05, β T = 2.

[0154] Calculation process:

[0155] Calculate the linear and non-linear impacts of age difference:

[0156]

[0157] Calculate the adjusted deviation detection threshold:

[0158] T = 10 + 2.425 + 2·1

[0159] = 14.425

[0160] The calculated adjusted deviation detection threshold is approximately 14.4. The result shows that for a 25-year-old male user, considering the increase in his age relative to the baseline age and gender factors, the basic deviation detection threshold can be appropriately increased.

[0161] Please refer to Figure 6 , and the steps to obtain the abnormal evaluation result are:

[0162] Based on the voice feature deviation data and the adjusted result of the deviation detection threshold, through the formula:

[0163]

[0164] Calculate the voice abnormality level L, where D is the voice feature deviation index, T is the adjusted deviation detection threshold, S L is the scale factor, ρ L is the dynamic adjustment parameter;

[0165] Based on the voice abnormality level L, compare it with the preset abnormal level table, and send the corresponding laryngeal abnormality reminder to the user to obtain the abnormal evaluation result.

[0166] Formula:

[0167]

[0168] Meaning and acquisition method of parameters:

[0169] D: Sound feature deviation index, obtained by the previous method.

[0170] T: Adjusted deviation detection threshold, obtained by the previous method.

[0171] S L : Basic proportionality factor of the anomaly level, a preset constant, whose value is based on historical data analysis or expert advice and is used to determine the sensitivity of sound anomaly assessment.

[0172] ρ L : Dynamic adjustment parameter. This parameter is introduced to increase the flexibility of the formula when dealing with the deviation magnitude, especially when the deviation magnitude is close to the threshold, and it can more finely control the calculation of the anomaly level.

[0173] Calculation example:

[0174] Set the parameters as follows: D = 15, the currently measured deviation index, T = 10, the adjusted deviation detection threshold, S L = 5, the basic proportionality factor of the anomaly level, ρ L = 0.1, the dynamic adjustment parameter.

[0175] Calculate the anomaly level L:

[0176]

[0177] The calculated anomaly level L = 1, indicating that the deviation shown by the current sound data exceeds the adjusted threshold, but the deviation is relatively small. The result means that the user needs to pay attention to potential laryngeal health problems, but the current degree of anomaly is not very serious.

[0178] Please refer to Figure 7 , the method for analyzing the degree of sound change is:

[0179] Collect the sound samples of the user before and after laryngeal treatment, including the volume and pitch before and after treatment, to obtain the sound data before and after treatment;

[0180] Based on the sound data before and after treatment, through the formula:

[0181]

[0182] Calculate the sound change index Z to obtain the evaluation result of the degree of sound change, where, P pre and P post respectively represent the average pitch before and after treatment, V pre and V post respectively represent the average volume before and after treatment, ω P and ωV is the weight coefficient, and Z is the voice change index.

[0183] Formula:

[0184]

[0185] Meaning and acquisition method of parameters:

[0186] P pre and P post : The parameters respectively represent the average pitch before and after treatment. The pitch is usually measured from the recorded audio samples through voice analysis software. The specific measurement involves spectral analysis of the voice signal to extract the main frequency components.

[0187] V pre and V post : The parameters respectively represent the average volume before and after treatment. The volume is obtained by analyzing the amplitude of the voice signal. Usually, decibels are used as the measurement unit, and it is determined through the calculation of the sound pressure level.

[0188] ω P and ω V : The weight coefficients of the pitch and volume change ratios. It is set according to the influence degree of pitch and volume on the treatment effect of the lesion, and is usually determined based on clinical research data or the suggestions of acoustic experts. The coefficient helps to adjust the contribution of different voice characteristics to the evaluation of the overall voice change degree.

[0189] Calculation example:

[0190] Set the following parameter values: P pre = 220Hz, P post = 230Hz, V pre = 70dB, V post = 75dB, ω P = 0.6, ω V = 0.4.

[0191] Calculate the pitch change ratio:

[0192]

[0193] Calculate the volume change ratio:

[0194]

[0195] Calculate the overall voice change index Z:

[0196]

[0197] The calculated degree of voice change Z is approximately 0.0573. The result indicates the comprehensive change in voice characteristics before and after treatment, where both pitch and volume have increased, and the increase in volume contributes more to the overall degree of change, which is reflected in the setting of the weight coefficient. The assessment of the degree of change can help medical providers judge the effectiveness of laryngeal treatment, especially in terms of the performance of voice function recovery.

[0198] Please refer to Figure 8 , and the steps to obtain the evaluation result of the treatment effect of laryngeal lesions are as follows:

[0199] Based on the evaluation result of the degree of voice change, extract the treatment duration according to the start time and end time of the treatment;

[0200] Based on the treatment duration, through the formula:

[0201]

[0202] Calculate the laryngeal treatment effect index Y, where Z is the voice change index, T Y is the total time spent on treatment, k Y is the influence coefficient of voice change, λ Y is the time influence coefficient, Y is the laryngeal treatment effect index, and e is the base of the natural logarithm;

[0203] Based on the laryngeal treatment effect index Y, compare it with the target treatment effect index to judge whether the treatment effect meets the expectation and obtain the evaluation result of the treatment effect of laryngeal lesions.

[0204] Formula:

[0205]

[0206] Meaning and acquisition method of parameters:

[0207] Z: Voice change index, obtained through the previous steps.

[0208] T Y : The total time spent on treatment, and the parameter is directly obtained from the patient's treatment record. It records the entire duration from the start to the end of the treatment.

[0209] k Y : Influence coefficient of voice change, a coefficient set according to the expected influence of voice change on the treatment effect. Usually determined based on clinical data or analysis of previous treatment cases to adjust the weight of voice change in the treatment effect evaluation.

[0210] λ Y: The time impact coefficient is used to adjust the influence of the time length on the evaluation of the treatment effect. The coefficient is set based on different types of treatments and their time sensitivities, which helps to balance the different impacts of treatment periods of different lengths on the effect.

[0211] Calculation example:

[0212] Set the following parameter values: Z = 0.75, the voice change index, indicating a significant improvement in the voice characteristics after treatment, T Y = 2 days, the total duration of the treatment, k Y = 1.5, the influence coefficient of voice change, λ Y = 0.2, the time impact coefficient.

[0213] Calculation process:

[0214] Calculate

[0215]

[0216] Use a calculator or mathematical software to get e 1.125 ≈ 3.0802.

[0217] Calculate the denominator 1 + λ Y ·T Y :

[0218] 1 + λ Y ·T Y = 1 + 0.2×2 = 1 + 0.4 = 1.4

[0219] Finally, calculate the evaluation result Y of the laryngeal treatment effect:

[0220]

[0221] The calculated evaluation result of the treatment effect Y ≈ 2.2. This value indicates that considering the combined influence of the degree of voice change and the treatment time, the treatment has a considerable effect. A higher Y value indicates that the treatment has led to a significant improvement in the voice characteristics and achieved this effect in a relatively short time, indicating that the treatment method is efficient. The evaluation method provides quantitative treatment effect information for doctors and patients, which helps to evaluate and compare the effects of different treatment methods.

[0222] The above is only a preferred embodiment of the present invention, and it does not limit the present invention in other forms. Any person skilled in the relevant art may use the disclosed technical content to make changes or modifications into equivalent embodiments with equivalent changes and apply them to other fields. However, any simple modification, equivalent change, and modification made to the above embodiments based on the technical essence of the present invention without departing from the technical solution content of the present invention still fall within the protection scope of the technical solution of the present invention.

Claims

1. A laryngeal lesion monitoring system based on sound recognition, characterized in that: The system comprises: The sound quality optimization module collects the user's sound data based on the sound capture device, and adjusts the sound filtering parameters according to the sound collection environment, performs sound filtering, removes environmental interference sounds, and obtains optimized sound data; The sound feature analysis module extracts the average pitch and volume features of the sound data by analyzing the user's historical sound data based on the optimized sound data, compares the current sound data with the historical average sound features, evaluates the sound feature deviation, and obtains the sound feature deviation data; The abnormality assessment module adjusts the deviation detection threshold based on the voice feature deviation data and the user's age and gender, compares the current voice feature deviation with the adjusted deviation detection threshold, assesses the voice abnormality level, issues a throat abnormality reminder to the user, and obtains an abnormality assessment result; Based on the abnormal evaluation results, the treatment effect evaluation module collects the user's voice data before and after receiving laryngeal treatment, analyzes the degree of voice change, and evaluates the laryngeal treatment effect in combination with the treatment duration to obtain a laryngeal lesion treatment effect evaluation result.

2. The laryngeal lesion monitoring system based on sound recognition according to claim 1, characterized in that: The steps for obtaining the optimized sound data are: Based on the sound capture device, the user's sound data is collected and the environmental noise data is recorded to obtain basic sound data; Based on the basic sound data, by the formula: Calculate the high-pass filter threshold θ, where N avg is the average value of the ambient noise level, N var is the variance of the noise level, N max is the maximum value of the noise level, N min is the minimum value of the noise level, α θ , β θ and γ θ is the adjustment factor, k θ is a stability constant, θ is the high-pass filter threshold, and e is the base of the natural logarithm; Based on the high-pass filter threshold θ and the basic sound data, the portion of the user's sound data that is lower than the high-pass filter threshold θ is removed to obtain optimized sound data.

3. The laryngeal lesion monitoring system based on sound recognition according to claim 1, characterized in that: The method for extracting the average pitch and volume characteristics of the sound data is: Based on the user's voice data records, extract the voice data within a certain period of time, including the voice pitch and volume, to obtain the user's historical voice data; Based on the historical voice data of the user, by the formula: and Calculate the historical average pitch P avg and the historical average volume V avg , and obtain the historical sound feature extraction results, where n is the total number of sound samples, P avg is the historical average pitch, V avg is the historical average volume, p i is the pitch of the ith sound sample, v i The volume of the i-th sound sample.

4. The laryngeal lesion monitoring system based on sound recognition according to claim 3, characterized in that: The steps for obtaining the sound feature deviation data are as follows: Based on the historical sound feature extraction results and the optimized sound data, extract the average pitch and volume of the current user's voice to obtain sound deviation associated data; Based on the sound deviation correlation data, by the formula: Calculate the sound feature deviation index D to obtain the sound feature deviation data, where λ D is the weighting factor for pitch difference, κ D is the influence coefficient of volume difference, P current and P avg Represent the current and historical average pitch, V current and V avg Represent the current and historical average volume respectively, and D is the sound feature deviation index.

5. The laryngeal lesion monitoring system based on sound recognition according to claim 1, characterized in that: The method for adjusting the deviation detection threshold is: Collect basic information of users, including age and gender, and extract basic deviation detection thresholds to obtain threshold-related data; Based on the threshold correlation data, through the formula: Calculate the adjusted deviation detection threshold T to obtain the deviation detection threshold adjustment result, where T base is the basic threshold, A is the user's age, and A base is the age baseline, G mod is the gender modulus, γ T is the linear influence coefficient, δ T and η T is the nonlinear influence coefficient, β T is the weight coefficient, e is the base of the natural logarithm, and T is the adjusted deviation detection threshold.

6. The laryngeal lesion monitoring system based on sound recognition according to claim 5, characterized in that: The steps for obtaining the abnormal evaluation result are: Based on the sound feature deviation data and the deviation detection threshold adjustment result, the formula: Calculate the sound abnormality level L, where D is the sound feature deviation index, T is the adjusted deviation detection threshold, S L is the proportionality factor, ρ L To dynamically adjust parameters; Based on the sound abnormality level L, a comparison is made with a preset abnormality level table, and a corresponding throat abnormality reminder is issued to the user to obtain an abnormality assessment result.

7. The laryngeal lesion monitoring system based on sound recognition according to claim 1, characterized in that: The method for analyzing the degree of sound change is: Collect the user's voice samples before and after laryngeal treatment, including the volume and pitch before and after treatment, to obtain the voice data before and after treatment; Based on the sound data before and after the treatment, the formula: Calculate the sound change index Z to obtain the sound change degree assessment result, where P pre and P post represent the average pitch before and after treatment, V pre and V post Represent the average volume before and after treatment, ω P and ω V is the weight coefficient, and Z is the sound change index.

8. The laryngeal lesion monitoring system based on sound recognition according to claim 7, characterized in that: The steps for obtaining the laryngeal lesion treatment effect evaluation result are as follows: Based on the sound change degree evaluation result, extracting the treatment duration according to the treatment start time and treatment end time; Based on the treatment duration, by the formula: Calculate the laryngeal treatment effect index Y, where Z is the voice change index, T Y is the total time spent on treatment, k Y is the influence coefficient of sound change, λ Y is the time influence coefficient, Y is the laryngeal treatment effect index, and e is the base of the natural logarithm; Based on the laryngeal treatment effect index Y, it is compared with the target treatment effect index to determine whether the treatment effect has reached the expectation, and the laryngeal lesion treatment effect evaluation result is obtained.

Images