A blood vessel detection control method and device based on multi-modal signal quality evaluation, equipment and storage medium

Abstract

The present application relates to the technical field of blood vessel detection signal processing and intelligent control, and discloses a blood vessel detection control method and device based on multi-modal signal quality evaluation, equipment and storage medium. The method obtains multiple signal channel data, respectively performs single-channel quality evaluation, and combines inter-channel consistency to perform comprehensive effectiveness determination, and then performs control operations such as continuing measurement, eliminating invalid data, adjusting measurement parameters, prompting re-measurement, re-measuring or terminating measurement according to the determination result. The single-channel quality evaluation, comprehensive effectiveness determination and control operation are cyclically executed in the detection process to form a real-time quality monitoring and control closed loop. The scheme can improve the effectiveness determination accuracy, adaptive control capability and measurement success rate of the detection process, and is suitable for blood vessel detection process control scenes such as blood pressure detection, ankle-brachial index detection, pulse wave conduction velocity detection and blood vessel hardness evaluation.

Description

Technical Field

[0001] This invention relates to the field of vascular detection signal processing and intelligent control technology, and in particular to a vascular detection control method, device, equipment and storage medium based on multimodal signal quality assessment. Background Technology

[0002] In blood pressure monitoring, ankle-brachial index (ABI) testing, pulse wave velocity (PVR) testing, and vascular stiffness assessment, the reliability of the results is highly dependent on the quality of the original physiological signals. For vascular monitoring devices used in home or portable settings, pressure pulse wave (BPPV) and photoplethysmography (PPG) pulse wave (PPG) signals are typically acquired simultaneously to obtain blood pressure parameters, propagation time difference (PTD), and vascular stiffness-related indicators.

[0003] However, in actual testing, the aforementioned signals are easily affected by factors such as body movement interference, abnormal cuff wearing, poor probe contact, insufficient perfusion, fluctuations during the deflation process, and sampling timing deviations, resulting in waveform distortion, unstable feature points, abnormal rhythms, or inconsistencies between channels. If these signals are directly used in parameter calculations when the signal quality is poor, it can easily lead to distorted results, or even render the test results meaningless.

[0004] In existing technologies, some solutions judge detection effectiveness based on a single signal amplitude or a single threshold, which is difficult to accurately reflect the true quality status in multi-channel joint detection scenarios. Other solutions can score the quality of a single channel, but they lack a comprehensive judgment mechanism that combines the consistency relationship between channels, and they also lack an adaptive control strategy that adjusts the detection process in real time based on the quality results, which makes it difficult to meet the requirements of home scenarios for reliability, interpretability and measurement success rate.

[0005] Therefore, a solution is needed that can perform single-channel quality assessment and inter-channel consistency assessment of multimodal physiological signals, and control the detection process in real time based on the comprehensive validity judgment results. Summary of the Invention

[0006] (a) Purpose of the invention The purpose of this invention is to provide a method, device, equipment and storage medium for blood vessel detection and control based on multimodal signal quality assessment, so as to solve the problems of inaccurate single threshold judgment, insufficient multi-channel joint judgment and weak detection and control adaptive capability in the prior art.

[0007] (II) Technical Solution To achieve the above objectives, the present invention adopts the following technical solution: On the one hand, a method for controlling vascular detection based on multimodal signal quality assessment is provided, comprising the following steps: 1. Acquire data from multiple signal channels during the vascular detection process, wherein the multiple signal channel data includes at least a pressure pulse wave related signal channel and a photoplethysmography pulse wave related signal channel; 2. Preprocess the data for each signal channel separately, and extract the single-channel quality evaluation index corresponding to each signal channel; 3. Based on the single-channel quality evaluation index corresponding to each signal channel, generate the single-channel quality results for each signal channel respectively; 4. Based on the single-channel quality results of each signal channel and the consistency evaluation results between different signal channels, a comprehensive validity judgment result for the current detection stage is generated; 5. Based on the comprehensive validity determination result, perform adaptive control on the current detection process.

[0008] The adaptive control includes at least one of continuing measurement, removing invalid data, adjusting measurement parameters, prompting for retesting, re-measuring, and terminating measurement; the preprocessing, single-channel quality assessment, comprehensive validity determination, and adaptive control are executed cyclically during the detection process to achieve real-time quality monitoring and control of the detection process.

[0009] In one embodiment, the pressure pulse wave correlated signal channel includes at least one of a cuff static pressure signal channel and a cuff pressure oscillation signal channel; the photoplethysmography pulse wave correlated signal channel includes at least one of a red photoplethysmography pulse wave signal channel and an infrared photoplethysmography pulse wave signal channel.

[0010] In one embodiment, the single-channel quality evaluation index includes at least one of waveform period consistency, peak-valley discernibility, waveform amplitude stability, signal-to-noise ratio, rising edge clarity, derivative significance, and body motion noise level; in the dual-wavelength photoplethysmography pulse wave scenario, it may also include a dual-wavelength correlation index; in the cuff pressure oscillation wave scenario, it may also include a cuff oscillation envelope stability index and a pressure change slope anomaly index.

[0011] In one implementation, the single-channel quality result can be generated by normalizing multiple single-channel quality evaluation indicators, calculating the single-channel quality score according to a preset weight, and comparing it with a quality threshold; or by generating a quality judgment model trained based on labeled samples.

[0012] In one embodiment, the consistency evaluation results include at least one of inter-channel rhythm consistency, inter-channel timing consistency, inter-channel waveform change consistency, and timing deviation stability between different heartbeats.

[0013] In one implementation, the comprehensive validity determination result includes at least one of the following: current data is valid, current data is valid with low confidence, current data is partially invalid, and current data is entirely invalid. Based on the comprehensive validity determination result, a control conclusion can be further generated that the current test needs to be retested or the current test should be terminated. Specifically, partially invalid current data means that some heartbeats, some channels, or some time windows do not meet the result calculation conditions but can still be detected; entirely invalid current data means that the current detection stage as a whole does not meet the result calculation conditions.

[0014] In one implementation, the adaptive control performs different control operations based on different overall validity determination results: for valid data, it performs continued measurement and / or participates in result calculation; for valid data with low confidence, it performs continued measurement and adds a confidence mark; for partially invalid data, it performs at least one of the following: removing invalid heartbeats, removing invalid time windows, extending sampling duration, continuing measurement, and adjusting measurement parameters; for overall invalid data, it performs at least one of the following: prompting retesting, re-measuring, restarting the inflation / deflation measurement process, or prompting to re-wear the sensor; for safety anomalies, it terminates measurement and / or rapidly depressurizes.

[0015] In one implementation, when a preset number of consecutive heartbeats or data within a preset duration are all judged to be partially invalid, the overall validity judgment result can be upgraded to overall invalid.

[0016] In one embodiment, adjusting the measurement parameters includes at least one of adjusting the deflation rate, adjusting the target inflation pressure, adjusting the sampling duration, adjusting the feature extraction threshold, and adjusting the quality judgment threshold.

[0017] In one implementation, when abnormal cuff pressure, persistent strong body motion interference, continuous loss of photoplethysmography (PPG) signal, or multiple consecutive invalid heartbeats are detected, the measurement is terminated, and rapid decompression and alarm can be further triggered.

[0018] In one embodiment, the method can be used for at least one of blood pressure detection, ankle-brachial index detection, pulse wave velocity detection, and vascular stiffness assessment.

[0019] In one implementation, the method outputs at least one of the following information in addition to the overall validity determination result: single-channel quality score, multimodal overall quality score, invalidity cause classification, retest suggestion, and result confidence level.

[0020] On the other hand, the present invention also provides an apparatus, electronic device, and computer-readable storage medium for implementing the above method. The apparatus includes: a signal acquisition module for acquiring data from multiple signal channels during the vascular detection process; a preprocessing module for preprocessing the data from each signal channel; an index extraction module for extracting single-channel quality evaluation indices corresponding to each signal channel; a single-channel evaluation module for generating single-channel quality results for each signal channel based on the single-channel quality evaluation indices; a comprehensive evaluation module for generating a comprehensive validity determination result for the current detection stage based on the single-channel quality results of each signal channel and the consistency evaluation results between different signal channels; and a control execution module for performing at least one of the following operations according to the comprehensive validity determination result: continuing measurement, removing invalid data, adjusting measurement parameters, prompting for retesting, re-measuring, and terminating measurement.

[0021] Upon meeting the safety termination conditions, the control execution module may also trigger a pressure relief or alarm operation. The electronic device includes a processor and a memory, the memory storing a computer program that, when executed by the processor, implements the above-described method. A computer-readable storage medium stores a computer program that, when executed by the processor, implements the above-described method.

[0022] (III) Beneficial Effects 1. By employing a two-tiered mechanism of "single-channel quality assessment + inter-channel consistency assessment" for comprehensive evaluation, the accuracy of validity assessment in multi-channel detection scenarios is improved; 2. By cyclically executing quality monitoring and control decisions during the testing process, the real-time adaptive capability of the testing process is enhanced; 3. By distinguishing between different states such as valid, low-confidence valid, partially invalid, and globally invalid, the control actions are matched one-to-one with the data states, thereby improving the accuracy and executability of the detection and control strategy; 4. By combining control operations such as invalid heartbeat rejection, extended sampling time, parameter adjustment, and retest prompts, the impact of adverse signals on the final detection results can be reduced; 5. Applicable to home self-testing, home-based chronic disease follow-up, community screening, and process control of blood pressure detection, ankle-brachial index detection, pulse wave velocity detection, and vascular stiffness assessment in portable vascular testing devices. Attached Figure Description

[0023] Figure 1 This is a schematic diagram of the overall process of the method of the present invention.

[0024] Figure 2 This is a schematic diagram of the single-channel quality assessment process of the present invention.

[0025] Figure 3This is a schematic diagram of the multimodal integrated quality assessment process of the present invention.

[0026] Figure 4 This is a schematic diagram of the detection and control decision-making process of the present invention.

[0027] Figure 5 This is a block diagram of the device structure of the present invention. Detailed Implementation

[0028] The present invention will be further described below with reference to the accompanying drawings and embodiments, but the scope of protection of the present invention is not limited to the following embodiments.

[0029] Example 1: Overall Method Flow. (For example...) Figure 1 As shown, the method of the present invention includes: signal acquisition 1, preprocessing 2, index extraction 3, single-channel quality assessment 4, multimodal integrated quality assessment 5, and control execution 6.

[0030] The process includes: Signal Acquisition 1, which acquires data from multiple signal channels during the detection process; Preprocessing 2, which performs filtering, drift removal, artifact suppression, and time-base unification on the data from each signal channel; Index Extraction 3, which extracts single-channel quality evaluation indicators; Single-Channel Quality Assessment 4, which generates single-channel quality results for each signal channel; Multimodal Comprehensive Quality Assessment 5, which generates a comprehensive validity judgment result for the current detection stage; and Control Execution 6, which controls the detection process based on the comprehensive validity judgment result. These steps are executed cyclically during the detection process to form a real-time quality monitoring closed loop.

[0031] Example 2: Single-channel quality assessment. For example... Figure 2 As shown, the single-channel quality assessment process includes: inputting the single-channel waveform 7, extracting the period consistency index 8, extracting the peak-valley discernibility index 9, extracting the amplitude stability index 10, extracting the noise-related index 11, calculating the single-channel quality score 12, and generating the single-channel quality result 13.

[0032] The periodic consistency index 8 is used to characterize the stability of the length of multiple consecutive heartbeat cycles; the peak-valley discernibility index 9 is used to characterize the prominence of key feature points in local waveforms; the amplitude stability index 10 is used to characterize the amplitude dispersion among multiple heartbeats; and the noise correlation index 11 is used to characterize the degree of high-frequency noise, baseline disturbance, or body motion interference. When calculating the single-channel quality score 12, each index can be normalized first, and then linearly combined according to preset weights; alternatively, each index can be input into a quality judgment model trained with labeled samples to output quality results. The generated single-channel quality results 13 preferably include: single-channel effective, single-channel low quality, and single-channel invalid.

[0033] Example 3: Multimodal integrated quality assessment. For example... Figure 3As shown, the multimodal integrated quality assessment can receive the single-channel quality results of pressure pulse wave 14, the single-channel quality results of PPG signal 15, and other channel quality results 16, and further calculate the inter-channel rhythm consistency 17, inter-channel timing consistency 18, and waveform change consistency 19, and then generate the integrated validity judgment result 21 through integrated assessment 20.

[0034] The comprehensive evaluation 20 may employ rule fusion, weighted scoring fusion, statistical model fusion, or machine learning model fusion methods. The comprehensive validity determination result 21 may include one or more of the following: the current data is valid, the current data is valid with low confidence, the current data is partially invalid, the current data is entirely invalid, the current test needs to be retested, and the current test should be terminated.

[0035] Example 4: Detection and Control Decisions. For example... Figure 4 As shown, based on the comprehensive validity judgment result 22, the control execution process can distinguish different states such as current data validity 23, low confidence validity 24, partial invalidity 25, overall invalidity 26, and safety anomaly 27. When the current data is valid 23, measurement continues or it participates in the result calculation; when the low confidence validity is valid 24, measurement continues and a confidence mark is added; when the partial invalidity is 25, invalid heartbeats are removed, invalid time windows are removed, the sampling time is extended, or the measurement parameters are adjusted; when the overall invalidity is 26, a prompt for retesting is executed, and it can further trigger the restart of the inflation / deflation measurement process or a prompt to re-wear the sensor; when the safety anomaly is 27, the measurement is terminated, and it can further trigger rapid depressurization and alarms.

[0036] In a specific application, when the cuff pressure exceeds the safety threshold, the deflation process exhibits an abnormal slope, the photoplethysmography (PPG) signal is continuously lost, there is continuous strong body motion interference, or multiple consecutive heartbeats are deemed invalid, the control execution process can directly enter the safety anomaly handling branch to avoid continuing to collect invalid data and improve detection safety.

[0037] Example 5: Device Structure. (For example...) Figure 5 As shown, the blood vessel detection and control device of the present invention includes a signal acquisition module 28, a preprocessing module 29, an index extraction module 30, a single-channel evaluation module 31, a comprehensive evaluation module 32, and a control execution module 33.

[0038] The signal acquisition module 28 is used to acquire data from multiple signal channels during the blood vessel detection process; the preprocessing module 29 is used to perform filtering, drift removal, artifact suppression, unified resampling, and time alignment; the index extraction module 30 is used to extract the single-channel quality evaluation index corresponding to each signal channel; the single-channel evaluation module 31 is used to generate single-channel quality results; the comprehensive evaluation module 32 is used to generate a comprehensive validity judgment result based on the single-channel quality results and the consistency evaluation results; the control execution module 33 is used to perform control operations such as continuing measurement, removing invalid data, adjusting parameters, prompting retesting, re-measuring, and terminating measurement according to the comprehensive validity judgment result.

[0039] When the safety termination conditions are met, the control execution module 33 can also trigger a pressure relief or alarm operation 34. This device can be deployed in home monitoring devices, host computers, mobile terminals, or embedded processors, and can be used in conjunction with blood pressure detection, ankle-brachial index detection, pulse wave velocity detection, and vascular stiffness assessment processes.

[0040] This invention is not limited to the embodiments described above. Any equivalent substitutions or modifications made by those skilled in the art to the single-channel quality evaluation indicators, multimodal consistency evaluation methods, comprehensive judgment strategies, and detection control actions without departing from the spirit and essence of this invention should fall within the protection scope of this invention.

Claims

1. A method for blood vessel detection and control based on multimodal signal quality assessment, characterized in that, The process includes the following steps: S1, acquiring data from multiple signal channels during the vascular detection process, wherein the multiple signal channel data includes at least a pressure pulse wave related signal channel and a photoplethysmography pulse wave related signal channel; S2, preprocessing the data from each signal channel and extracting the single-channel quality evaluation index corresponding to each signal channel; S3, generating the single-channel quality result for each signal channel based on the single-channel quality evaluation index corresponding to each signal channel. S4. Based on the single-channel quality results of each signal channel and the consistency evaluation results between different signal channels, generate a comprehensive validity judgment result for the current detection stage; S5. Based on the comprehensive validity determination result, perform adaptive control on the current detection process; wherein, the adaptive control includes at least one of continuing measurement, removing invalid data, adjusting measurement parameters, prompting retesting, re-measuring, and terminating measurement; steps S2 to S5 are executed cyclically during the detection process to achieve real-time quality monitoring and control of the detection process.

2. The method according to claim 1, characterized in that, The pressure pulse wave related signal channel includes at least one of the cuff static pressure signal channel and the cuff pressure oscillation signal channel, and the photoplethysmography pulse wave related signal channel includes at least one of the red photoplethysmography pulse wave signal channel and the infrared photoplethysmography pulse wave signal channel.

3. The method according to claim 1, characterized in that, The preprocessing in step S2 includes at least one of bandpass filtering, baseline drift removal, DC component removal, abnormal pulse rejection, motion artifact suppression, uniform resampling, and time alignment.

4. The method according to claim 1, characterized in that, The single-channel quality evaluation index includes at least one of the following: waveform period consistency index, peak-valley discernibility index, waveform amplitude stability index, signal-to-noise ratio index, rising edge sharpness index, derivative significance index, and body motion noise level index. When there are red light photoplethysmography pulse wave signal channels and infrared light photoplethysmography pulse wave signal channels, the single-channel quality evaluation index also includes a dual-wavelength correlation index. When there is a cuff pressure oscillation signal channel, the single-channel quality evaluation index also includes at least one of the following: cuff oscillation envelope stability index and pressure change slope anomaly index.

5. The method according to claim 1, characterized in that, In step S3, the single-channel quality result is obtained by normalizing multiple single-channel quality evaluation indicators of the corresponding signal channel and calculating the single-channel quality score according to a preset weight, and then comparing the single-channel quality score with a preset quality threshold. Alternatively, the output can be obtained through a quality assessment model trained based on labeled samples; the single-channel quality result includes at least one of single-channel valid, single-channel low quality, and single-channel invalid.

6. The method according to claim 1, characterized in that, The consistency evaluation results include at least one of the following: rhythm consistency, timing consistency, waveform change consistency, and timing deviation stability of characteristic points between different signal channels; the comprehensive validity judgment results are generated by rule fusion, weighted scoring fusion, statistical model fusion, or machine learning model fusion, and include at least one of the following: current data is valid, current data is valid with low confidence, current data is partially invalid, current data is completely invalid, current detection needs to be retested, and current detection should be terminated.

7. The method according to claim 1, characterized in that, Different control operations are performed based on different comprehensive validity determination results: when the comprehensive validity determination result is that the current data is valid, continue measurement and / or participate in the result calculation; when the comprehensive validity determination result is that the current data is valid with low confidence, continue measurement and add a confidence level label to the result output. When the overall validity determination result indicates that the current data is partially invalid, at least one of the following actions is performed: removing invalid heartbeats, removing invalid time windows, extending the sampling duration, continuing measurement, and adjusting measurement parameters. When a preset number of consecutive heartbeats or data within a preset duration are all judged to be partially invalid, the overall validity determination result is upgraded to the current data being entirely invalid. When the overall validity determination result indicates that the current data is entirely invalid, at least one of the following actions is performed: prompting for retesting, re-measuring, restarting the inflation / deflation measurement process, or prompting for re-wearing the sensor. When abnormal cuff pressure is detected, body motion interference exceeds a preset threshold, photoplethysmography pulse wave signal is continuously lost, or multiple consecutive heartbeats are judged to be invalid, measurement termination and / or rapid depressurization control are performed, and alarm information is output.

8. The method according to claim 1, characterized in that, In addition to the overall validity determination result, the method outputs at least one of the following information: single-channel quality score, multimodal overall quality score, invalidity cause classification, retest suggestion, and result confidence level.

9. A blood vessel detection and control device based on multimodal signal quality assessment, characterized in that, include: The signal acquisition module is used to acquire data from multiple signal channels during the vascular detection process; The preprocessing module is used to preprocess the data from each signal channel; The index extraction module is used to extract the single-channel quality evaluation indexes corresponding to each signal channel. The single-channel evaluation module is used to generate single-channel quality results for each signal channel based on single-channel quality evaluation indicators. The comprehensive evaluation module is used to generate a comprehensive validity judgment result for the current detection stage based on the single-channel quality results of each signal channel and the consistency evaluation results between different signal channels; The control execution module is used to perform at least one of the following operations based on the comprehensive validity determination result: continue measurement, remove invalid data, adjust measurement parameters, prompt for retesting, remeasure, and terminate measurement, and trigger pressure relief or alarm when the safe termination condition is met.

10. An electronic device or computer-readable storage medium, characterized in that, It includes a processor and a memory, or thereon storing a computer program, which, when executed by the processor, implements the method of any one of claims 1 to 8.

Images