Fistulization blood flow prediction method and system, wearable device and monitoring platform

Abstract

The invention provides a fistulization blood flow prediction method and system, wearable equipment and a monitoring platform. The prediction method comprises the following steps: acquiring fistulization sample data of different sample users at a plurality of different moments, wherein each piece of fistulization sample data comprises a fistulization voiceprint signal and a blood flow volume at the same moment; processing each fistulization voiceprint signal to obtain a corresponding voiceprint feature; based on each group of voiceprint features and blood flow, constructing a voiceprint blood flow mapping relation; acquiring a target fistulization voiceprint signal of the target object at the acquisition moment; processing the target fistulization voiceprint signal to obtain a corresponding target voiceprint feature; based on the target voiceprint feature and the voiceprint blood flow mapping relation, the fistulization blood flow of the target object is obtained through prediction. According to the method, the change trend of the blood flow can be accurately and continuously displayed in real time, the health condition of the fistula can be more accurately evaluated, measures can be taken in time, remote monitoring and data analysis can be achieved, and the experience feeling of a user is improved.

Description

Technical Field

[0001] This disclosure relates to the field of stoma health monitoring technology, and in particular to a method, system, wearable device and monitoring platform for predicting stoma blood flow. Background Technology

[0002] Arteriovenous fistulas (AVFs) are the lifeline for patients undergoing maintenance hemodialysis. Maintaining good AVF function can improve the quality of hemodialysis treatment for uremia patients and prolong their survival time. However, traditional fistula monitoring requires invasive surgery and has low accuracy in predicting blood flow, making it impossible to provide timely warnings of potential risks such as arterial and venous thrombosis and stenosis. Summary of the Invention

[0003] The technical problem to be solved by this disclosure is to overcome the shortcomings of the prior art in that the prediction accuracy of fistula blood flow is not high and the potential risks cannot be warned in a timely manner, and to provide a method, system, wearable device and monitoring platform for predicting fistula blood flow.

[0004] This disclosure solves the above-mentioned technical problems through the following technical solution:

[0005] According to a first aspect of this disclosure, a method for predicting stoma blood flow is provided, the prediction method comprising:

[0006] Acquire stoma sample data from different sample users at several different times, each stoma sample data including stoma acoustic signature signal and blood flow at the same time;

[0007] Each of the fistula-forming voiceprint signals is processed to obtain the corresponding voiceprint features;

[0008] Based on the voiceprint features and blood flow in each group, a voiceprint-blood flow mapping relationship is constructed.

[0009] Collect the target fistula acoustic signal of the target object at the time of collection;

[0010] The target fistula voiceprint signal is processed to obtain the corresponding target voiceprint features;

[0011] Based on the target voiceprint features and the voiceprint blood flow mapping relationship, the stoma blood flow of the target object is predicted.

[0012] Preferably, the step of processing each of the stoma voiceprint signals to obtain the corresponding voiceprint features includes:

[0013] The voiceprint signal segments that meet the preset conditions are selected from each of the fistula voiceprint signals;

[0014] Extract the voiceprint features from the voiceprint signal segment.

[0015] Preferably, the step of filtering out the voiceprint signal segments that meet the preset conditions from each of the stoma voiceprint signals includes:

[0016] Determine whether the length of each of the stoma voiceprint signals meets the preset number of acquisition points. If so, the stoma voiceprint signal is taken as the voiceprint signal segment. Otherwise, the stoma voiceprint signal is filled with a preset filling method to obtain the voiceprint signal segment that meets the preset number of acquisition points.

[0017] The preset number of sampling points represents the number of sampling points that the sensor pre-sets for a single sampling.

[0018] Preferably, the step of filtering out the voiceprint signal segments that meet the preset conditions from each of the stoma voiceprint signals further includes:

[0019] Each of the fistula-forming voiceprint signals is filtered using a preset autocorrelation function to obtain the voiceprint signal segments that meet the preset conditions.

[0020] Preferably, the step of filtering each of the stoma voiceprint signals using a preset autocorrelation function to obtain the voiceprint signal segments that meet the preset conditions includes:

[0021] Obtain the ACF autocorrelation function (an autocorrelation function) of the stoma acoustic signature signal under each sliding window;

[0022] Iterate through each sliding window and obtain the maximum autocorrelation value and hysteresis value of the current sliding window based on the ACF autocorrelation function of the current sliding window;

[0023] Determine whether the maximum autocorrelation value and the hysteresis value meet the preset threshold. If so, determine that the voiceprint signal segment under the current sliding window meets the preset condition.

[0024] The voiceprint signal segments that meet the preset conditions are selected.

[0025] Preferably, the step of extracting the voiceprint features from the voiceprint signal segment includes:

[0026] The voiceprint signal segment is normalized and denoised using wavelet denoising to obtain the target voiceprint signal segment.

[0027] Extract the voiceprint features from the target voiceprint signal segment.

[0028] Preferably, the step of extracting the voiceprint features from the target voiceprint signal segment includes:

[0029] The target voiceprint signal segment is processed using the MFCC (Mel frequency cepstral coefficients) function to obtain the voiceprint feature matrix;

[0030] Based on the aforementioned voiceprint feature matrix, the MFCC coefficients at different times are obtained;

[0031] The MFCC coefficients at several different times are weighted to obtain the voiceprint features.

[0032] Preferably, the voiceprint blood flow mapping relationship includes a voiceprint blood flow mapping table and / or a voiceprint blood flow change curve;

[0033] And / or,

[0034] The step of collecting the target fistula acoustic signature signal of the target object at the time of collection includes:

[0035] Collect multiple target fistula voiceprint signals of the target object at different times;

[0036] When acquiring multiple target stoma voiceprint signals of the target object at multiple consecutive acquisition times, after the step of predicting the stoma blood flow of the target object based on the target voiceprint features and the voiceprint blood flow mapping relationship, the prediction method further includes:

[0037] Based on multiple stoma blood flow rates corresponding to multiple target stoma voiceprint signals, the actual blood flow change trend of the target object at multiple consecutive acquisition times is obtained; and / or, the actual blood flow change trend of the target object at multiple consecutive acquisition times is obtained, and the actual blood flow change trend is compared with the voiceprint blood flow change curve. If a target curve segment in the voiceprint blood flow change curve with a similarity greater than a set value to the actual blood flow change trend is obtained, the blood flow change trend of the target curve segment within a future set time period is taken as the predicted blood flow change trend of the target object within the future set time period; wherein, the voiceprint blood flow mapping relationship is the voiceprint blood flow change curve.

[0038] According to a second aspect of this disclosure, a system for predicting stoma blood flow is provided, the system comprising an acquisition module, a processing module, a construction module, a collection module, and a prediction module;

[0039] The acquisition module is used to acquire stoma sample data of different sample users at several different times. Each stoma sample data includes stoma acoustic signature signal and blood flow at the same time.

[0040] The processing module is used to process each of the fistula voiceprint signals to obtain the corresponding voiceprint features;

[0041] The construction module is used to construct a voiceprint-blood flow mapping relationship based on each group of voiceprint features and blood flow.

[0042] The acquisition module is used to acquire the target fistula acoustic signature signal of the target object at the acquisition time;

[0043] The processing module is also used to process the target fistula voiceprint signal to obtain the corresponding target voiceprint features;

[0044] The prediction module is used to predict the stoma blood flow of the target object based on the target voiceprint features and the voiceprint blood flow mapping relationship.

[0045] Preferably, the processing module includes a filtering unit and an extraction unit;

[0046] The filtering unit is used to filter out voiceprint signal segments that meet preset conditions from each of the fistula voiceprint signals.

[0047] The extraction unit is used to extract the voiceprint features from the voiceprint signal segment.

[0048] Preferably, the filtering unit is specifically used to determine whether the length of the voiceprint signal of each of the stoma voiceprint signals meets the preset number of acquisition points condition. If so, the stoma voiceprint signal is used as the voiceprint signal segment; otherwise, the stoma voiceprint signal is filled with a preset filling method to obtain the voiceprint signal segment that meets the preset number of acquisition points condition.

[0049] The preset number of sampling points represents the number of sampling points that the sensor pre-sets for a single sampling.

[0050] Preferably, the filtering unit is further configured to use a preset autocorrelation function to filter each of the stoma voiceprint signals to obtain the voiceprint signal segments that meet the preset conditions.

[0051] Preferably, the filtering unit is specifically used to obtain the autocorrelation function (ACF) of the stoma voiceprint signal under each sliding window; traverse each sliding window in a loop, and obtain the maximum autocorrelation value and hysteresis value under the current sliding window based on the ACF autocorrelation function under the current sliding window; determine whether the maximum autocorrelation value and the hysteresis value meet a preset threshold, and if so, determine that the voiceprint signal segment under the current sliding window meets the preset condition; and filter out the voiceprint signal segment that meets the preset condition.

[0052] Preferably, the extraction unit is further configured to perform normalization processing and wavelet denoising processing on the voiceprint signal segment to obtain a target voiceprint signal segment; and extract the voiceprint features in the target voiceprint signal segment.

[0053] Preferably, the extraction unit is specifically used to process the target voiceprint signal segment using the MFCC (Mel frequency cepstral coefficients) function to obtain a voiceprint feature matrix; to obtain MFCC coefficients at different times based on the voiceprint feature matrix; and to perform weighted processing on the MFCC coefficients at several different times to obtain the voiceprint features.

[0054] Preferably, the voiceprint blood flow mapping relationship includes a voiceprint blood flow mapping table and / or a voiceprint blood flow change curve;

[0055] And / or,

[0056] The acquisition module is also used to acquire multiple target fistula voiceprint signals of the target object at different times;

[0057] When multiple target stoma voiceprint signals of the target object are continuously acquired at multiple acquisition times of continuous time, the prediction module is further used to obtain the actual blood flow change trend of the target object at multiple acquisition times of continuous time based on the multiple stoma blood flow corresponding to the multiple target stoma voiceprint signals; and / or, to obtain the actual blood flow change trend of the target object at multiple acquisition times of continuous time, and to compare the actual blood flow change trend with the voiceprint blood flow change curve. If a target curve segment in the voiceprint blood flow change curve with a similarity greater than a set value to the actual blood flow change trend is obtained, the blood flow change trend of the target curve segment in the future set time period is taken as the predicted blood flow change trend of the target object in the future set time period; wherein, the voiceprint blood flow mapping relationship is the voiceprint blood flow change curve.

[0058] According to a third aspect of this disclosure, a wearable device is provided for implementing the method for predicting stoma blood flow as described in the first aspect of this disclosure.

[0059] According to a fourth aspect of this disclosure, a stoma blood flow monitoring platform is provided, the stoma blood flow platform including the stoma blood flow prediction system described in the second aspect of this disclosure.

[0060] According to a fifth aspect of this disclosure, an electronic device is provided, including a memory, a processor, and a computer program stored in the memory and for running on the processor, wherein the processor executes the computer program to implement the method for predicting stoma blood flow according to a first aspect of this disclosure.

[0061] According to a sixth aspect of this disclosure, a computer-readable storage medium is provided having a computer program stored thereon, which, when executed by a processor, implements the method for predicting stoma blood flow as described in the first aspect of this disclosure.

[0062] According to a seventh aspect of this disclosure, a computer program product is provided, comprising a computer program that, when executed by a processor, implements the method for predicting stoma blood flow as described in the first aspect of this disclosure.

[0063] Based on common knowledge in the field, the above-mentioned preferred conditions can be combined arbitrarily to obtain various preferred embodiments of this disclosure.

[0064] The positive advancements of this disclosure lie in the following: by constructing a voiceprint-blood flow mapping relationship between voiceprint features and blood flow, it can accurately, continuously, and in real-time display blood flow change trends without invasive surgery. This provides doctors and nurses with more data, helping them to more accurately assess the health of the fistula and take timely measures to address the problem. Furthermore, it enables remote monitoring and data analysis, improving the user experience. Attached Figure Description

[0065] Figure 1 This is a first flowchart of the method for predicting stoma blood flow according to Embodiment 1 of this disclosure;

[0066] Figure 2 This is a time-domain comparison diagram of preoperative and postoperative stoma voiceprint signals collected by the stoma voiceprint sensor in Embodiment 1 of this disclosure.

[0067] Figure 3 This is a second flowchart of the method for predicting stoma blood flow according to Embodiment 1 of this disclosure;

[0068] Figure 4 This is the third flowchart of the method for predicting stoma blood flow in Embodiment 1 of this disclosure;

[0069] Figure 5 This is a schematic diagram of the preoperative and postoperative MFCC characteristic changes in Embodiment 1 of this disclosure;

[0070] Figure 6 This is a graph showing the trend of changes in blood flow and MFCC voiceprint characteristics before and after dialysis in a maintenance dialysis patient according to Embodiment 1 of this disclosure;

[0071] Figure 7 This is a graph showing the weighted changes in blood flow and MFCC characteristics before and after surgery in a patient with reconstructed fistula according to Embodiment 1 of this disclosure;

[0072] Figure 8 This is a schematic diagram of the modules of the fistula blood flow prediction system of Embodiment 2 of this disclosure;

[0073] Figure 9 This is a schematic diagram of the structure of an electronic device according to Embodiment 5 of this disclosure. Detailed Implementation

[0074] The present disclosure is further illustrated below by way of embodiments, but the present disclosure is not limited to the scope of the embodiments described herein.

[0075] The prefixes such as "first" and "second" used in this disclosure are merely for distinguishing different descriptive objects and do not limit the position, order, priority, quantity, or content of the described objects. The use of ordinal numbers and other prefixes used to distinguish descriptive objects in this disclosure does not constitute a limitation on the described objects. The description of the described objects is given in the claims or the context of the embodiments, and should not be construed as an unnecessary limitation. Furthermore, in the description of this embodiment, unless otherwise stated, "multiple" means two or more.

[0076] In this embodiment of the disclosure, the collection, storage, use, processing, transmission, provision, and disclosure of user personal information comply with relevant laws and regulations and do not violate public order and good morals.

[0077] Example 1

[0078] In a specific embodiment of this disclosure, a method for predicting stoma blood flow is provided, such as... Figure 1 As shown, the prediction method includes:

[0079] S1. Obtain stoma sample data from different sample users at several different times. Each stoma sample data includes stoma acoustic signature signal and blood flow at the same time.

[0080] S2. Process each stoma voiceprint signal to obtain the corresponding voiceprint features;

[0081] S3. Based on each set of voiceprint features and blood flow, a voiceprint-blood flow mapping relationship is constructed;

[0082] S4. Collect the target fistula acoustic signal of the target object at the time of collection;

[0083] S5. Process the target fistula voiceprint signal to obtain the corresponding target voiceprint features;

[0084] S6. Based on the target voiceprint features and the voiceprint blood flow mapping relationship, the fistula blood flow of the target object is predicted.

[0085] Specifically, patients undergoing dialysis (reconstructed fistula, newly created fistula, and maintenance dialysis) can be used as sample users. Step S1 involves using a fistula acoustic signature sensor to collect the fistula acoustic signature signals (cephalic vein vibration data) of these three types of dialysis patients at different times, and using real-time ultrasound monitoring of blood flow data at the corresponding times to obtain fistula sample data. For example, ... Figure 2 As shown, preoperative and postoperative analgesic voiceprint signals and corresponding blood flow data were collected from three types of dialysis patients as analgesic sample data. Figure 2 The image above shows the preoperative stoma voiceprint signal and corresponding blood flow data. Figure 2 The figure below shows the postoperative stoma voiceprint signal and corresponding blood flow data; then, step S2 is used to process the stoma voiceprint signal in each stoma sample data to extract the corresponding voiceprint features.

[0086] Analysis of multiple sets of stoma sample data revealed that the changes in blood flow and voiceprint features before and after surgery in three types of dialysis patients were largely consistent. Specifically, the higher the blood flow value, the higher the voiceprint feature value, and vice versa. Therefore, step S3 maps the extracted voiceprint features to the corresponding blood flow data to construct a voiceprint-blood flow mapping relationship, which is then used to predict stoma blood flow.

[0087] After constructing the voiceprint-blood flow mapping relationship, the target ostomy voiceprint signal of the target object at the time of acquisition can be collected in step S4, and the target ostomy voiceprint signal can be processed in step S5 to extract the corresponding target voiceprint features. Then, based on the constructed voiceprint-blood flow mapping relationship, the blood flow corresponding to the target voiceprint features can be obtained in step S6, thereby realizing the prediction of the ostomy blood flow of the target object.

[0088] This specific implementation establishes a voiceprint-blood flow mapping relationship between voiceprint features and blood flow, enabling accurate, continuous, and real-time visualization of blood flow trends without invasive surgery. This provides doctors and nurses with more data to help them more accurately assess the health of the fistula and take timely measures to address the problem. Furthermore, it enables remote monitoring and data analysis, enhancing the user experience.

[0089] In one specific implementation, such as Figure 3 As shown, step S2 includes:

[0090] S21. Select the voiceprint signal segments that meet the preset conditions from each stoma voiceprint signal;

[0091] S22. Extract voiceprint features from the voiceprint signal segment.

[0092] Specifically, in order to improve the accuracy of blood flow prediction, it is necessary to filter the stoma voiceprint signals collected by the sensor, obtain the voiceprint signal segments corresponding to the effective period in the stoma voiceprint signals, and extract the corresponding voiceprint features to construct the voiceprint blood flow mapping relationship.

[0093] In one specific embodiment, step S21 includes:

[0094] Determine whether the length of each stoma voiceprint signal meets the preset number of acquisition points. If so, the stoma voiceprint signal is taken as a voiceprint signal segment. Otherwise, the stoma voiceprint signal is filled with a preset filling method to obtain a voiceprint signal segment that meets the preset number of acquisition points.

[0095] Among them, the preset number of sampling points represents the number of sampling points that the sensor has set in advance for a single sampling.

[0096] Specifically, the length L of the stoma acoustic signature signal acquired in a single acquisition can be calculated. By determining whether L is equal to the number of acquisition points set by the sensor in a single acquisition, if they are not equal, the stoma acoustic signature signal is linearly filled to the number of acquisition points set by the sensor in a single acquisition. For example, if the number of acquisition points set by the sensor in a single acquisition is 40,000, and L ≠ 40,000, then the stoma acoustic signature signal is linearly filled once, where the number of points to be filled is a = 40,000 - L.

[0097] In one specific embodiment, step S21 further includes:

[0098] Each stoma voiceprint signal is filtered using a preset autocorrelation function to obtain voiceprint signal segments that meet preset conditions.

[0099] Specifically, when screening the voiceprint signal segments corresponding to the effective period in the analgesia voiceprint signal, a preset autocorrelation function can be used, such as the ACF autocorrelation function, to analyze the autocorrelation and periodicity of the analgesia voiceprint signal, and then extract the voiceprint signal segments corresponding to the effective period.

[0100] In one specific implementation, the step of filtering each stoma voiceprint signal using a preset autocorrelation function to obtain voiceprint signal segments that meet preset conditions includes:

[0101] Obtain the ACF autocorrelation function of the fistula acoustic signature signal under each sliding window;

[0102] Iterate through each sliding window and obtain the maximum autocorrelation value and hysteresis value of the current sliding window based on the ACF autocorrelation function of the current sliding window;

[0103] Determine whether the maximum autocorrelation value and hysteresis value meet the preset thresholds. If so, determine that the voiceprint signal segment under the current sliding window meets the preset conditions.

[0104] Voiceprint signal segments that meet the preset conditions are selected.

[0105] Specifically, to avoid suboptimal acquisition of stoma acoustic signature signals due to human intervention, vibration, or poor contact, the ACF autocorrelation function can be used to filter the stoma acoustic signature signals. This is achieved by setting... Sliding window size W_s, square wave threshold Value S_t = (the sum of the maximum amplitude W_max and the minimum amplitude W_min) The parameters are: minimum period length (min_period_threshold), minimum value difference (min_value_diff). For example, the sliding window size W_s is set to 15000, the square wave threshold S_t is set to 4000, the minimum period length (min_period_threshold) is set to 1000, and the minimum value difference (min_value_diff) is set to 50. Then, the autocorrelation function array of the time series data of the fistula voiceprint signal is calculated by iterating through each sliding window. Based on the preset square wave threshold and minimum value difference, the calculated maximum autocorrelation value and hysteresis value are filtered and judged to determine whether the data of the current window meets the requirements. If it does not meet the requirements, it is removed, thereby filtering out the voiceprint signal segments that meet the preset conditions.

[0106] In one specific implementation, such as Figure 4 As shown, step S22 includes:

[0107] S221. Normalize and perform wavelet denoising on the voiceprint signal segment to obtain the target voiceprint signal segment;

[0108] S222. Extract the voiceprint features from the target voiceprint signal segment.

[0109] Specifically, to eliminate the influence of data scales from different devices, the voiceprint signal obtained in step S21 can be normalized. The length of the voiceprint signal segment filtered in step S21 is L′, and the normalization method can be... Then, wavelet denoising is used to remove noisy data. The wavelet type can be db4 wavelet, and the wavelet decomposition level can be set to 3.

[0110] In one specific embodiment, step S222 includes:

[0111] The target voiceprint signal segment is processed using the MFCC function to obtain the voiceprint feature matrix;

[0112] MFCC coefficients at different times are obtained based on the voiceprint feature matrix;

[0113] The MFCC coefficients at several different times are weighted to obtain the voiceprint features.

[0114] Specifically, the target voiceprint signal segment obtained in step S221 can be processed using the MFCC function, utilizing mel(f) = 2595*log 10 (1+f / 700) maps the frequency of the target voiceprint signal segment to the Mel nonlinear spectrum, where f represents the frequency. Then, the Mel nonlinear spectrum is filtered using a Mel filter bank, which typically consists of 26 triangular filter banks. These 26 filters can be represented by a matrix Hm, where the rows of the matrix indicate the filter number and the columns indicate the number of points in the Fourier transform, thus obtaining the voiceprint feature matrix.

[0115] Since most of the signal data is concentrated in the low-frequency region after transformation, only the first 13 dimensions of each frame are taken for discrete cosine transform to obtain the MFCC coefficients at different times. After weighting, the voiceprint features can be obtained. The calculation formula for the voiceprint features is as follows:

[0116]

[0117] Where H represents the voiceprint feature matrix, M represents the number of Mel filters, i represents the data of the i-th frame, n represents the n-th column of the i-th frame, and m represents the m-th Mel filter.

[0118] In a specific example, the target voiceprint signal segment L″ obtained in step S221 can be used as input, and the MFCC function in the python_speech_features module of Python (a programming language) can be used for feature representation. Here, fs can be set to 1300, winlen = 0.02, winstep = 0.01, nfilt = 26, and nfft = 26. Then, the column average of the obtained M(rows, cols) feature matrix is ​​performed to obtain the data M'(1, 13), representing the different MFCC coefficients at each time step. Figure 5 As shown, each small graph represents different columns of MFCC features. The horizontal axis represents the patient's name, and the vertical axis represents the normalized MFCC coefficient. Preoperative and postoperative features are connected by a short line. If the feature changes before and after surgery are consistent with the changes in blood flow, the slope is non-negative.

[0119] Then, the MFCC coefficients are weighted; for example, the sum of the data in columns 9, 10, 11, and 12 is used as the final MFCC voiceprint feature.

[0120] In one specific implementation, the voiceprint blood flow mapping relationship includes a voiceprint blood flow mapping table and / or a voiceprint blood flow change curve.

[0121] Specifically, the voiceprint blood flow mapping relationship can be represented by a mapping table, a change curve, or by inputting it into a preset model to train the preset model, and then using the model to predict blood flow. This embodiment does not make specific limitations on this.

[0122] In one specific implementation, step S4 includes:

[0123] S41. Collect multiple target fistula voiceprint signals of the target object at different times;

[0124] When acquiring multiple target fistula acoustic signature signals of the target object at multiple consecutive acquisition times, after step S6, the prediction method further includes:

[0125] S7. Based on multiple stoma blood flow rates corresponding to multiple target stoma acoustic signals, the actual blood flow change trend of the target object at multiple acquisition times in continuous time is obtained.

[0126] Specifically, when predicting blood flow, multiple target stoma voiceprint signals of the target object at different times can be continuously collected over a period of time. By predicting the actual blood flow corresponding to each target stoma voiceprint signal, the actual blood flow change trend of the target object over a period of time can be obtained.

[0127] In one specific implementation, step S7 further includes:

[0128] Based on multiple stoma blood flow rates corresponding to multiple target stoma voiceprint signals, the actual blood flow change trend of the target object at multiple consecutive acquisition times is obtained. The actual blood flow change trend is compared with the voiceprint blood flow change curve. If a target curve segment with a similarity greater than a set value to the actual blood flow change trend is found in the voiceprint blood flow change curve, the blood flow change trend of the target curve segment within a set time period in the future is taken as the predicted blood flow change trend of the target object within the set time period in the future. The voiceprint blood flow mapping relationship is the voiceprint blood flow change curve.

[0129] Specifically, after obtaining the actual blood flow change trend of the target object over a period of time, the actual blood flow change trend can be compared with the constructed voiceprint blood flow change curve. If the similarity between the actual blood flow change trend and a certain segment of the voiceprint blood flow change curve (i.e., the target curve segment) is greater than a set value, it indicates that the blood flow change trend of the target object is consistent with the trend of the voiceprint blood flow change curve. The change trend after the target curve segment can be used as the predicted blood flow change trend of the target object thereafter, thereby achieving early warning of fistula abnormalities, enabling doctors to intervene in a timely manner, and further ensuring the patient's life safety.

[0130] In a specific experiment, a total of 53 patients participated, including 22 maintenance dialysis patients, with 276 cases collected (pre-dialysis and post-dialysis as one group); 16 patients underwent new fistula construction, with 16 groups (pre-operative and post-operative as one group); and 15 patients underwent fistula reconstruction, with 15 groups (pre-operative and post-operative as one group). The comparative experiment included: (1) changes in blood flow volume and MFCC voiceprint characteristics before and after dialysis in maintenance dialysis patients (results are shown in the figure). Figure 6 (As shown), if the changes in the upper and lower slopes are consistent, it is determined to be a positive correlation, which means that the characteristic value is large before dialysis and small after dialysis, corresponding to a large blood flow before dialysis and a small blood flow after dialysis. Among them, a total of 260 groups were positively correlated. The correlation between the characteristic value and blood flow changes before and after dialysis in maintenance dialysis patients was 94%, and the negative correlation was 6%. (2) The pattern of characteristic value and blood flow changes before and after surgery in patients with reconstructed fistulas (results are shown in the figure). Figure 7 As shown in Table 1), the slope is positive, indicating consistent changes, that is, the preoperative blood flow is less than the postoperative blood flow and the characteristic value changes are consistent. There are 14 groups with the same changes, and the positive correlation is 14 / 15 = 93%. (3) The characteristic values ​​of maintenance dialysis patients before dialysis are much greater than those of patients with reconstructed fistulas before surgery. The characteristics of maintenance dialysis patients before surgery and the characteristics of reconstructed fistulas before surgery are clearly distinguishable, with the mean being about 2 times (as shown in Table 1). (4) The characteristics of newly created fistulas after surgery are clearly distinguishable from those of maintenance dialysis patients before dialysis (as shown in Table 2).

[0131] Table 1. Statistical analysis of t-test results before fistula reconstruction versus before maintenance dialysis.

[0132]

[0133] Table 2. t-test statistics of patients with newly formed fistulas vs. those before maintenance dialysis.

[0134]

[0135] In summary, the changes in blood flow and characteristic features before and after surgery were largely consistent across the three types of dialysis patients. Data from 432 cases involving 22 maintenance patients showed an accuracy rate of 94.3%, while data from 32 cases involving 16 reconstructed fistula patients showed an accuracy rate of 93%. A t-test comparing postoperative blood flow in newly constructed fistula patients with preoperative blood flow in maintenance dialysis patients (as shown in Table 1) and preoperative blood flow in reconstructed fistula patients with preoperative blood flow in maintenance dialysis patients (as shown in Table 2) showed that postoperative blood flow in both groups was lower than that in maintenance dialysis patients. Postoperative characteristic values ​​were also significantly lower in these two groups, according to the t-test. Therefore, the method for predicting fistula blood flow provided in any of the above embodiments can effectively reflect changes in blood flow in fistula patients, providing an auxiliary method for clinical diagnosis.

[0136] This embodiment establishes a voiceprint-blood flow mapping relationship between voiceprint features and blood flow, enabling accurate, continuous, and real-time visualization of blood flow trends without invasive surgery. This provides doctors and nurses with more data to help them more accurately assess the health of the fistula and take timely measures to address the problem. Furthermore, it enables remote monitoring and data analysis, enhancing the user experience.

[0137] Example 2

[0138] In one specific embodiment of this disclosure, a system for predicting stoma blood flow is provided, such as... Figure 8 As shown, the prediction system includes an acquisition module 100, a processing module 200, a construction module 300, a data collection module 400, and a prediction module 500.

[0139] The acquisition module 100 is used to acquire stoma sample data of different sample users at several different times. Each stoma sample data includes stoma acoustic signature signal and blood flow at the same time.

[0140] The processing module 200 is used to process each stoma voiceprint signal to obtain the corresponding voiceprint features;

[0141] The construction module 300 is used to construct a voiceprint-blood flow mapping relationship based on each set of voiceprint features and blood flow.

[0142] The acquisition module 400 is used to acquire the target fistula acoustic signal of the target object at the acquisition time;

[0143] The processing module 200 is also used to process the target fistula voiceprint signal to obtain the corresponding target voiceprint features;

[0144] The prediction module 500 is used to predict the stoma blood flow of the target object based on the target voiceprint features and the voiceprint blood flow mapping relationship.

[0145] In one specific embodiment, the processing module 200 includes a filtering unit and an extraction unit;

[0146] The filtering unit is used to filter out the voiceprint signal segments that meet the preset conditions from each fistula voiceprint signal;

[0147] The extraction unit is used to extract voiceprint features from the voiceprint signal segment.

[0148] In one specific implementation, the filtering unit is specifically used to determine whether the length of the voiceprint signal of each stoma voiceprint signal meets the preset number of acquisition points condition. If so, the stoma voiceprint signal is taken as a voiceprint signal segment; otherwise, the stoma voiceprint signal is filled with a preset filling method to obtain a voiceprint signal segment that meets the preset number of acquisition points condition.

[0149] Among them, the preset number of sampling points represents the number of sampling points that the sensor has set in advance for a single sampling.

[0150] In one specific implementation, the screening unit is further configured to use a preset autocorrelation function to screen each fistula voiceprint signal to obtain voiceprint signal segments that meet preset conditions.

[0151] In one specific implementation, the filtering unit is specifically used to obtain the autocorrelation function (ACF) of the ostomy voiceprint signal under each sliding window; traverse each sliding window in a loop, and obtain the maximum autocorrelation value and hysteresis value under the current sliding window based on the ACF autocorrelation function under the current sliding window; determine whether the maximum autocorrelation value and hysteresis value meet the preset threshold; if so, determine that the voiceprint signal segment under the current sliding window meets the preset condition; and filter out the voiceprint signal segments that meet the preset condition.

[0152] In one specific embodiment, the extraction unit is further configured to perform normalization and wavelet denoising on the voiceprint signal segment to obtain the target voiceprint signal segment; and extract the voiceprint features in the target voiceprint signal segment.

[0153] In one specific implementation, the extraction unit is specifically used to process the target voiceprint signal segment using the MFCC function to obtain a voiceprint feature matrix; obtain MFCC coefficients at different times based on the voiceprint feature matrix; and perform weighted processing on the MFCC coefficients at several different times to obtain voiceprint features.

[0154] In one specific implementation, the voiceprint blood flow mapping relationship includes a voiceprint blood flow mapping table and / or a voiceprint blood flow change curve.

[0155] In one specific embodiment, the acquisition module 400 is also used to acquire multiple target fistula voiceprint signals of the target object at different times;

[0156] When multiple target stoma voiceprint signals of the target object are continuously acquired at multiple acquisition times in a continuous time period, the prediction module 500 is also used to obtain the actual blood flow change trend of the target object at multiple acquisition times in a continuous time period based on multiple stoma blood flow corresponding to the multiple target stoma voiceprint signals; and / or, to obtain the actual blood flow change trend of the target object at multiple acquisition times in a continuous time period, and to compare the actual blood flow change trend with the voiceprint blood flow change curve. If a target curve segment in the voiceprint blood flow change curve with a similarity greater than a set value to the actual blood flow change trend is obtained, the blood flow change trend of the target curve segment in the future set time period is taken as the predicted blood flow change trend of the target object in the future set time period; wherein, the voiceprint blood flow mapping relationship is the voiceprint blood flow change curve.

[0157] For the system embodiments, since they basically correspond to the method embodiments, the relevant parts can be referred to in the description of the method embodiments. The system embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components 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 modules can be selected to achieve the purpose of this disclosure according to actual needs.

[0158] This embodiment establishes a voiceprint-blood flow mapping relationship between voiceprint features and blood flow, enabling accurate, continuous, and real-time visualization of blood flow trends without invasive surgery. This provides doctors and nurses with more data to help them more accurately assess the health of the fistula and take timely measures to address the problem. Furthermore, it enables remote monitoring and data analysis, enhancing the user experience.

[0159] Example 3

[0160] In one specific embodiment of this disclosure, a wearable device is provided for implementing the method for predicting stoma blood flow provided in any of the above embodiments.

[0161] Wearable devices can include various types such as wristbands, wrist straps, and blood pressure cuffs.

[0162] This embodiment establishes a voiceprint-blood flow mapping relationship between voiceprint features and blood flow, enabling accurate, continuous, and real-time visualization of blood flow trends without invasive surgery. This provides doctors and nurses with more data to help them more accurately assess the health of the fistula and take timely measures to address the problem. Furthermore, it enables remote monitoring and data analysis, enhancing the user experience.

[0163] Example 4

[0164] In one specific embodiment of this disclosure, a stoma blood flow monitoring platform is provided, which includes the stoma blood flow prediction system provided in any of the above embodiments.

[0165] This embodiment establishes a voiceprint-blood flow mapping relationship between voiceprint features and blood flow, enabling accurate, continuous, and real-time visualization of blood flow trends without invasive surgery. This provides doctors and nurses with more data to help them more accurately assess the health of the fistula and take timely measures to address the problem. Furthermore, it enables remote monitoring and data analysis, enhancing the user experience.

[0166] Example 5

[0167] Figure 9 This is a schematic diagram of the structure of an electronic device according to an example embodiment of the present disclosure. The electronic device includes a memory, a processor, and a computer program stored in the memory and used to run on the processor. When the processor executes the computer program, it implements the method for predicting stoma blood flow provided in any of the above embodiments. Figure 9 The electronic device 30 shown is merely an example and should not impose any limitation on the functionality and scope of use of the embodiments disclosed herein.

[0168] like Figure 9 As shown, the electronic device 30 can be manifested as a general-purpose computing device, such as a server device. The components of the electronic device 30 may include, but are not limited to: at least one processor 31, at least one memory 32, and a bus 33 connecting different system components (including memory 32 and processor 31).

[0169] Bus 33 includes a data bus, an address bus, and a control bus.

[0170] The memory 32 may include volatile memory, such as random access memory (RAM) 321 and / or cache memory 322, and may further include read-only memory (ROM) 323.

[0171] The memory 32 may also include a program tool 325 (or utility) having a set (at least one) program module 324, such program module 324 including but not limited to: an operating system, one or more application programs, other program modules, and program data, each or some combination of these examples may include an implementation of a network environment.

[0172] The processor 31 executes various functional applications and data processing by running computer programs stored in the memory 32, such as the method for predicting stoma blood flow provided in any of the above embodiments.

[0173] Electronic device 30 can also be connected to one or more external devices 34 (e.g., keyboard, pointing device, etc.). Communication is possible via input / output (I / O) interface 35. Furthermore, electronic device 30 can also communicate with one or more networks (e.g., local area networks (LANs), wide area networks (WANs), and / or public networks, such as the Internet) via network adapter 36. As shown, network adapter 36 communicates with other modules of electronic device 30 via bus 33. It should be understood that, although not shown in the figures, other hardware and / or software modules can be used in conjunction with electronic device 30, including but not limited to: microcode, device drivers, redundant processors, external disk drive arrays, RAID (disk array) systems, tape drives, and data backup storage systems.

[0174] It should be noted that although several units / modules or sub-units / modules of the electronic device have been mentioned in the detailed description above, this division is merely exemplary and not mandatory. In fact, according to embodiments of this disclosure, the features and functions of two or more units / modules described above can be embodied in one unit / module. Conversely, the features and functions of one unit / module described above can be further divided and embodied by multiple units / modules.

[0175] Example 6

[0176] This disclosure also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the method for predicting stoma blood flow provided in any of the above embodiments.

[0177] The readable storage medium may be more specifically adopted, including but not limited to: portable disk, hard disk, random access memory, read-only memory, erasable programmable read-only memory, optical storage device, magnetic storage device, or any suitable combination thereof.

[0178] Example 7

[0179] This disclosure also provides a computer program product, including a computer program that, when executed by a processor, implements the method for predicting stoma blood flow provided in any of the above embodiments.

[0180] The program code for executing the computer program product of this disclosure can be written in any combination of one or more programming languages, and the program code can be executed entirely on a user device, partially on a user device, as a stand-alone software package, partially on a user device and partially on a remote device, or entirely on a remote device.

[0181] While specific embodiments of this disclosure have been described above, those skilled in the art should understand that these are merely illustrative examples, and the scope of protection of this disclosure is defined by the appended claims. Those skilled in the art can make various changes or modifications to these embodiments without departing from the principles and essence of this disclosure, but all such changes and modifications fall within the scope of protection of this disclosure.

Claims

1. A method for predicting stoma blood flow, characterized in that, The prediction method includes: Acquire stoma sample data from different sample users at several different times, each stoma sample data including stoma acoustic signature signal and blood flow at the same time; Each of the fistula-forming voiceprint signals is processed to obtain the corresponding voiceprint features; Based on the voiceprint features and blood flow in each group, a voiceprint-blood flow mapping relationship is constructed. Collect the target fistula acoustic signal of the target object at the time of collection; The target fistula voiceprint signal is processed to obtain the corresponding target voiceprint features; Based on the target voiceprint features and the voiceprint blood flow mapping relationship, the stoma blood flow of the target object is predicted.

2. The prediction method according to claim 1, characterized in that, The step of processing each of the fistula-forming voiceprint signals to obtain the corresponding voiceprint features includes: The voiceprint signal segments that meet the preset conditions are selected from each of the fistula voiceprint signals; Extract the voiceprint features from the voiceprint signal segment.

3. The prediction method according to claim 2, characterized in that, The step of filtering to obtain the voiceprint signal segment that meets the preset conditions from each of the fistula voiceprint signals includes: Determine whether the length of each of the stoma voiceprint signals meets the preset number of acquisition points. If so, the stoma voiceprint signal is taken as the voiceprint signal segment. Otherwise, the stoma voiceprint signal is filled with a preset filling method to obtain the voiceprint signal segment that meets the preset number of acquisition points. The preset number of sampling points represents the number of sampling points that the sensor pre-sets for a single sampling.

4. The prediction method according to claim 2 or 3, characterized in that, The step of filtering to obtain the voiceprint signal segment that meets the preset conditions in each of the fistula voiceprint signals further includes: Each of the fistula-forming voiceprint signals is filtered using a preset autocorrelation function to obtain the voiceprint signal segments that meet the preset conditions.

5. The prediction method according to claim 4, characterized in that, The step of filtering each of the stoma voiceprint signals using a preset autocorrelation function to obtain the voiceprint signal segments that meet the preset conditions includes: Obtain the ACF autocorrelation function of the fistula acoustic signature signal under each sliding window; Iterate through each sliding window and obtain the maximum autocorrelation value and hysteresis value of the current sliding window based on the ACF autocorrelation function of the current sliding window; Determine whether the maximum autocorrelation value and the hysteresis value meet the preset threshold. If so, determine that the voiceprint signal segment under the current sliding window meets the preset condition. The voiceprint signal segments that meet the preset conditions are selected.

6. The prediction method according to claim 2, characterized in that, The step of extracting the voiceprint features from the voiceprint signal segment includes: The voiceprint signal segment is normalized and denoised using wavelet denoising to obtain the target voiceprint signal segment. Extract the voiceprint features from the target voiceprint signal segment.

7. The prediction method according to claim 6, characterized in that, The step of extracting the voiceprint features from the target voiceprint signal segment includes: The target voiceprint signal segment is processed using the MFCC function to obtain the voiceprint feature matrix; Based on the aforementioned voiceprint feature matrix, the MFCC coefficients at different times are obtained; The MFCC coefficients at several different times are weighted to obtain the voiceprint features.

8. The prediction method according to any one of claims 1 to 7, characterized in that, The voiceprint blood flow mapping relationship includes a voiceprint blood flow mapping table and / or a voiceprint blood flow change curve. And / or, The step of collecting the target fistula acoustic signature signal of the target object at the time of collection includes: Collect multiple target fistula voiceprint signals of the target object at different times; When acquiring multiple target stoma voiceprint signals of the target object at multiple consecutive acquisition times, after the step of predicting the stoma blood flow of the target object based on the target voiceprint features and the voiceprint blood flow mapping relationship, the prediction method further includes: Based on multiple stoma blood flow rates corresponding to multiple target stoma voiceprint signals, the actual blood flow change trend of the target object at multiple consecutive acquisition times is obtained; and / or, the actual blood flow change trend of the target object at multiple consecutive acquisition times is obtained, and the actual blood flow change trend is compared with the voiceprint blood flow change curve. If a target curve segment in the voiceprint blood flow change curve with a similarity greater than a set value to the actual blood flow change trend is obtained, the blood flow change trend of the target curve segment within a future set time period is taken as the predicted blood flow change trend of the target object within the future set time period; wherein, the voiceprint blood flow mapping relationship is the voiceprint blood flow change curve.

9. A system for predicting stoma blood flow, characterized in that, The prediction system includes an acquisition module, a processing module, a construction module, a data collection module, and a prediction module; The acquisition module is used to acquire stoma sample data of different sample users at several different times. Each stoma sample data includes stoma acoustic signature signal and blood flow at the same time. The processing module is used to process each of the fistula voiceprint signals to obtain the corresponding voiceprint features; The construction module is used to construct a voiceprint-blood flow mapping relationship based on each group of voiceprint features and blood flow. The acquisition module is used to acquire the target fistula acoustic signature signal of the target object at the acquisition time; The processing module is also used to process the target fistula voiceprint signal to obtain the corresponding target voiceprint features; The prediction module is used to predict the stoma blood flow of the target object based on the target voiceprint features and the voiceprint blood flow mapping relationship.

10. A wearable device, characterized in that, The wearable device is used to implement the method for predicting stoma blood flow as described in any one of claims 1 to 8.

11. A stoma blood flow monitoring platform, characterized in that, The stoma blood flow platform includes the stoma blood flow prediction system as described in claim 9.

12. An electronic device comprising a memory, a processor, and a computer program stored in the memory and for running on the processor, characterized in that, When the processor executes the computer program, it implements the method for predicting stoma blood flow as described in any one of claims 1 to 8.

13. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the method for predicting stoma blood flow as described in any one of claims 1 to 8.

14. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by the processor, it implements the method for predicting stoma blood flow as described in any one of claims 1 to 8.

Images