A method and system for detecting leakage of hysteroscopic fluid management

Abstract

The present invention discloses a hysteroscopic liquid management leakage detection method and system, which specifically relates to the field of hysteroscopic surgery liquid management, and is used to solve the problem of real-time monitoring leakage. The method is to set the sampling frequency and the moving average method, collect pressure data in real time, dynamically update the lag time using the cross-correlation function and the sliding window, build a prediction model to improve the pressure prediction accuracy, dynamically adjust the perfusion pump output using the PID algorithm, continuously monitor and calculate the pressure difference and lag time in real time, generate corresponding signals, divide the lag time series to calculate the lag time complexity index, combine the pressure fluctuation asymmetry index, and comprehensively quantify the system disorder and abnormal pressure fluctuation, provide in-depth risk assessment, and ensure the safety of surgery. Through multi-level evaluation and real-time adjustment, it is possible to accurately monitor the liquid management status, identify potential leakage risks, ensure the smooth progress of surgery and patient safety, and improve system reliability and surgery quality.

Description

Technical Field

[0001] The present invention relates to the field of hysteroscopic surgery liquid management, and more specifically, to a hysteroscopic liquid management leakage detection method and system. Background Art

[0002] During hysteroscopic surgery, accurate fluid management is critical to the success of the operation and the safety of the patient. The operation requires dilating the uterine cavity by infusing fluid to ensure a clear surgical field of view and operating space. However, problems such as fluid leakage, line blockage, or equipment failure may occur during the operation, which can cause abnormal changes in fluid flow and affect surgical results and patient safety. Therefore, it is crucial to monitor the status of the fluid management system in real time and identify potential risks in a timely manner.

[0003] The existing method mainly calculates the fluid balance by the weight of the injected fluid and the collected fluid. Although this method is simple and direct, its accuracy and timeliness are limited in a dynamic surgical environment. Complex situations such as fluid seepage, local retention, and instantaneous flow changes cannot be accurately reflected by a simple weight difference. In addition, this method does not adequately identify the disorder and abnormal pressure fluctuations within the system, making it difficult to provide a comprehensive and in-depth risk assessment, and unable to promptly warn of potential leakage risks, resulting in difficulties in timely handling of sudden problems that may arise during surgery, thereby affecting the smooth progress of the operation and the safety of the patient.

[0004] In order to solve the above problems, a technical solution is now provided. Summary of the invention

[0005] In order to overcome the above-mentioned defects of the prior art, the embodiment of the present invention provides a hysteroscopic fluid management leakage detection method and system, which collects and smoothes pressure data in real time by setting the sampling frequency and moving average method, dynamically updates the lag time by using the cross-correlation function and the sliding window method, builds an adaptive prediction model and adopts the extended Kalman filter algorithm to improve the pressure prediction accuracy, dynamically adjusts the output of the perfusion pump by the proportional-integral-differential control algorithm, continuously monitors and calculates the pressure difference and lag time in real time, generates corresponding signals, divides the lag time series and calculates the approximate entropy and multi-scale entropy, obtains the lag time complexity index, combines the pressure fluctuation asymmetry index, calculates the leakage risk warning index by cross entropy and geometric mean, comprehensively quantifies the system disorder and abnormal pressure fluctuation, provides in-depth risk assessment, and ensures the safety of surgery. Through multi-level and multi-dimensional comprehensive evaluation and real-time adjustment, the system can accurately monitor the state of fluid management, identify potential leakage risks, ensure the smooth progress of surgery and patient safety, improve system reliability and surgical quality, so as to solve the problems raised in the above background technology.

[0006] To achieve the above object, the present invention provides the following technical solutions:

[0007] S1, before filling the liquid, perform self-check and calibration on the sensors and filling system at the ports of the inlet and outlet pipes, and record the calibration data each time;

[0008] S2, set the sampling frequency, collect and smooth the pressure data of the inlet pipe and the outlet pipe, calculate the cross-correlation function to determine the lag time, build an adaptive prediction model to predict the outlet pipe pressure, adjust the perfusion pump output through the PID control algorithm, continuously monitor and update the model, generate a corresponding indication signal, and if the obtained indication signal is a signal to be processed, synchronously execute step S3 and step S4;

[0009] S3, splits the lag time series according to the set window size, calculates the approximate entropy and multi-scale entropy of each subsequence, calculates the lag time complexity index by geometric mean and normalization, and quantifies the disorder and complexity of the series;

[0010] S4, real-time acquisition and smoothing of pressure data, dividing the time series into multiple windows, calculating the difference between positive and negative fluctuations in each window, normalizing it to obtain the pressure fluctuation asymmetry index, and quantifying the symmetry of pressure changes;

[0011] S5, based on the hysteresis time complexity index and the pressure fluctuation asymmetry index, uses the cross entropy combined with the geometric mean method to comprehensively calculate the leakage risk warning index, quantify the instability of the system and abnormal pressure fluctuations, evaluate the leakage risk and take corresponding measures.

[0012] In a preferred embodiment, S2 includes the following contents:

[0013] S2.1, set the sampling frequency, collect the pressure data of the liquid inlet pipe and the liquid outlet pipe in real time through the sensor, and form a time series;

[0014] S2.2, using a moving average method to perform noise filtering on the collected pressure data to obtain smoothed inlet pipe pressure data and smoothed outlet pipe pressure data;

[0015] S2.3, calculating the cross-correlation function of the inlet and outlet pressures to determine the lag time between the two;

[0016] The sliding window method is used to select the real-time lag time corresponding to the maximum cross-correlation coefficient as the time lag of the pressure of the inlet pipe and the outlet pipe;

[0017] S2.4, based on the real-time lag time and real-time pressure data, an adaptive prediction model is constructed and an extended Kalman filter algorithm is used to predict the outlet pipe pressure;

[0018] S2.5, calculating the difference between the actual outlet pipe pressure and the predicted pressure;

[0019] The output of the perfusion pump is adjusted according to the pressure difference. The proportional-integral-differential control algorithm is adopted to achieve dynamic regulation and precise control of the liquid flow by adjusting the proportional gain, integral gain and differential gain.

[0020] S2.6, if the difference between the actual outlet pipe pressure and the predicted pressure is within the pressure difference threshold range, a normal signal is generated;

[0021] If the difference between the actual outlet pipe pressure and the predicted pressure exceeds the lower limit of the pressure difference threshold but does not reach the upper limit, a signal to be analyzed is generated;

[0022] If the difference between the actual outlet pipe pressure and the predicted pressure exceeds the upper limit of the pressure difference threshold, an abnormal signal is generated.

[0023] In a preferred embodiment, step S3 includes the following contents:

[0024] For the signal to be analyzed, the generated lag time series is divided according to the set window size to form multiple subsequences;

[0025] S3.1, for each subsequence, calculate the approximate entropy;

[0026] S3.2, for each subsequence, calculate the multi-scale entropy;

[0027] S3.3, the approximate entropy and multiscale entropy are calculated and normalized using the geometric mean to obtain the lag time complexity index.

[0028] In a preferred embodiment, step S4 includes the following contents:

[0029] For the signal to be analyzed, the moving average method is used to perform noise filtering and smoothing on the data to obtain smoothed pressure data;

[0030] S4.1, split the pressure time series into multiple time windows;

[0031] For each time window, the positive fluctuations in the pressure series are calculated, i.e., the sum of the rising parts of the data points;

[0032] For each time window, calculate the negative fluctuations in the pressure series, i.e., the sum of the falling parts of the data points;

[0033] S4.2, calculate the difference between positive and negative fluctuations in each time window to obtain the fluctuation asymmetry;

[0034] The fluctuation asymmetry of all time windows is averaged and normalized to obtain the overall pressure fluctuation asymmetry index.

[0035] In a preferred embodiment, step S5 includes the following contents:

[0036] Compare the leakage risk warning index with the risk threshold. If the leakage risk warning index is greater than or equal to the risk threshold, corresponding inspection and intervention measures need to be taken immediately; otherwise, if the leakage risk warning index is less than the risk threshold.

[0037] In a preferred embodiment, the specific calculation steps of step S3.1 are as follows:

[0038] For each embedding vector, calculate its distance to other embedding vectors;

[0039] If the distance between each embedding vector and other embedding vectors is less than the preset similarity threshold, the two are considered to be similar vectors;

[0040] Count the number of other embedding vectors similar to each embedding vector, divide it by the total number of embedding vectors, and get the similarity ratio;

[0041] The approximate entropy is obtained by taking the logarithm of the similarity ratio of all embedding vectors and then averaging them.

[0042] In a preferred embodiment, the specific calculation steps of step S3.2 are as follows:

[0043] Perform multi-scale decomposition on the lagged time series. According to a given time scale, the series is segmented according to the time scale, and the average value of each segment is taken to form a new subsequence.

[0044] For each subsequence obtained by multi-scale decomposition, calculate its entropy value. The specific steps are as follows:

[0045] Count the probability of each different state appearing in the subsequence;

[0046] For each state, calculate its probability multiplied by its log probability and mark it as the first result;

[0047] Sum the first results of all states and take the negative value, which is the entropy value of the corresponding subsequence, that is, the multi-scale entropy.

[0048] A hysteroscopic fluid management leakage detection system, comprising: a self-check calibration module, a data acquisition module, a hysteresis analysis module, a complexity analysis module, a fluctuation analysis module and a risk assessment module;

[0049] Before perfusing the liquid, the self-checking and calibration module performs self-checking and calibration on the sensors and perfusion system at the ports of the inlet and outlet pipes, and records each calibration data, which is then transmitted to the data acquisition module;

[0050] The data acquisition module sets the sampling frequency, collects and smoothes the pressure data of the liquid inlet pipe and the liquid outlet pipe, and transmits the smoothed pressure data to the lag analysis module and the fluctuation analysis module;

[0051] The hysteresis analysis module calculates the cross-correlation function to determine the hysteresis time, constructs an adaptive prediction model to predict the pressure of the liquid outlet pipe, adjusts the output of the perfusion pump through the PID control algorithm, continuously monitors and updates the model, generates a corresponding indication signal, and if the obtained indication signal is a signal to be processed, the data is passed to the complexity analysis module and the fluctuation analysis module;

[0052] The complexity analysis module divides the lag time series according to the set window size, calculates the approximate entropy and multi-scale entropy of each subsequence, calculates the lag time complexity index through geometric mean and normalization, quantifies the disorder and complexity of the sequence, and passes the lag time complexity index to the risk assessment module;

[0053] The fluctuation analysis module collects and smoothes the pressure data in real time, divides the time series into multiple windows, calculates the difference between positive and negative fluctuations in each window, normalizes it to obtain the pressure fluctuation asymmetry index, quantifies the symmetry of pressure changes, and passes the pressure fluctuation asymmetry index to the risk assessment module;

[0054] The risk assessment module is based on the hysteresis time complexity index and the pressure fluctuation asymmetry index. It uses the cross entropy combined with the geometric mean method to comprehensively calculate the leakage risk warning index, quantify the instability of the system and abnormal pressure fluctuations, assess the leakage risk and take corresponding measures.

[0055] The technical effects and advantages of the hysteroscopic liquid management leakage detection method and system of the present invention are as follows:

[0056] 1. The present invention collects and smoothes the pressure data of the inlet and outlet tubes in real time by setting the sampling frequency and the moving average method to ensure the accuracy and consistency of the data. The cross-correlation function and the sliding window method are used to dynamically calculate and update the lag time to accurately reflect the real-time pressure changes. Based on this, an adaptive prediction model is constructed, and the extended Kalman filter algorithm is used to improve the accuracy of pressure prediction. Through the proportional-integral-differential control algorithm, the output of the perfusion pump is dynamically adjusted according to the difference between the actual pressure and the predicted pressure to achieve precise control and regulation of the liquid flow. The pressure difference and lag time are continuously monitored and calculated in real time, and the prediction model and feedback adjustment parameters are updated. According to the difference between the actual outlet tube pressure and the predicted pressure, the system generates normal, to-be-analyzed and abnormal signals to ensure timely identification and response to potential problems, and effectively guarantee the accuracy of intrauterine fluid management and the safety of surgery.

[0057] 2. The present invention divides the lag time series and calculates the approximate entropy and multi-scale entropy to comprehensively obtain the lag time complexity index, and then combines the real-time acquisition and smoothing of pressure data, calculates the difference between positive and negative fluctuations by dividing the time window, and obtains the pressure fluctuation asymmetry index. Finally, the cross entropy combined with the geometric mean method is used to comprehensively calculate the leakage risk warning index, comprehensively quantify the disorder and abnormal pressure fluctuations of the system, and provide a deeper risk assessment. This series of processing steps and comprehensive analysis can not only monitor and adjust the system state in real time, but also identify potential leakage risks through more comprehensive data analysis and comprehensive evaluation. By further calculating the leakage risk warning index, it is possible to combine data from multiple dimensions to quantify the complex dynamic changes of the system, thereby providing a deeper understanding of the system state and the ability to provide early warning. This multi-level and multi-dimensional comprehensive evaluation mechanism helps medical staff to identify and respond to potential risks in a timely manner, take corresponding measures, ensure the smooth progress of the operation and the safety of the patient, and further improve the reliability of the system and the quality of surgery. BRIEF DESCRIPTION OF THE DRAWINGS

[0058] Figure 1 A schematic diagram of a flow chart of a hysteroscopic liquid management leakage detection method of the present invention;

[0059] Figure 2 The present invention is a structural schematic diagram of a hysteroscopic liquid management leakage detection system. DETAILED DESCRIPTION

[0060] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0061] Embodiment 1: Figure 1 The present invention provides a hysteroscopic liquid management leakage detection method, comprising:

[0062] S1, before filling the liquid, perform self-check and calibration on the sensors and filling system at the ports of the inlet and outlet pipes, and record the calibration data each time;

[0063] S2, set the sampling frequency, collect and smooth the pressure data of the inlet pipe and the outlet pipe, calculate the cross-correlation function to determine the lag time, build an adaptive prediction model to predict the outlet pipe pressure, adjust the perfusion pump output through the PID control algorithm, continuously monitor and update the model, generate a corresponding indication signal, and if the obtained indication signal is a signal to be processed, synchronously execute step S3 and step S4;

[0064] S3, splits the lag time series according to the set window size, calculates the approximate entropy and multi-scale entropy of each subsequence, calculates the lag time complexity index by geometric mean and normalization, and quantifies the disorder and complexity of the series;

[0065] S4, real-time acquisition and smoothing of pressure data, dividing the time series into multiple windows, calculating the difference between positive and negative fluctuations in each window, normalizing it to obtain the pressure fluctuation asymmetry index, and quantifying the symmetry of pressure changes;

[0066] S5, based on the hysteresis time complexity index and the pressure fluctuation asymmetry index, uses the cross entropy combined with the geometric mean method to comprehensively calculate the leakage risk warning index, quantify the instability of the system and abnormal pressure fluctuations, evaluate the leakage risk and take corresponding measures.

[0067] Step S1 includes the following contents:

[0068] Before perfusing the liquid, the sensors and perfusion system at the ports of the inlet and outlet pipes are self-checked and calibrated, and the calibration data is recorded each time. Self-checking ensures the normal working state of the sensors and perfusion system, while calibration ensures the accuracy and consistency of the pressure data. Recording the calibration data facilitates subsequent analysis and tracking of equipment performance. This process provides a reliable data basis for subsequent operations, ensures accurate control and monitoring of the system during surgery, and further improves the safety and effectiveness of surgery.

[0069] During hysteroscopic surgery, accurate fluid management is crucial for surgical success and patient safety. The operation requires maintaining an appropriate amount of perfusion fluid in the uterine cavity to ensure a clear surgical field of view and prevent tissue damage. However, the flow characteristics of the fluid and the possible leakage or bleeding of body fluids during surgery make fluid management complicated. By real-time monitoring of the pressure of the inlet and outlet tubes, the state of liquid flow can be grasped, but due to the time lag in the flow of liquid in the pipeline, direct observation of pressure data may not reflect the actual situation in a timely manner. For this reason, the calculation and analysis of the lag time is introduced to more accurately predict the pressure of the outlet tube, so as to achieve early warning and timely adjustment.

[0070] By calculating and analyzing the lag time of the inlet and outlet pipes, we can understand the time difference of the liquid from the inlet pipe to the outlet pipe, so as to establish a more accurate liquid flow model. This prediction model, combined with actual pressure data, can identify abnormal situations in advance, such as minor body fluid leakage or pipe blockage, as well as serious leakage or equipment failure. By setting a preset threshold, the system can issue an early warning when the difference exceeds the threshold, which is divided into three situations: normal, further analysis required, and failure.

[0071] The significance of this treatment method is to improve the accuracy and response speed of fluid management, ensure the smooth progress of the operation and the safety of the patient. By identifying potential problems in advance, the risk of surgical interruption and complications can be reduced, and the success rate of the operation can be improved. In addition, this method can provide clear guidance to medical staff, so that they can quickly take appropriate countermeasures during the operation to avoid serious problems. Through scientific fluid management methods, not only the quality of surgery is improved, but also the patient's treatment experience and safety are improved.

[0072] S2 includes the following:

[0073] S2.1, set sampling frequency , collect the pressure data of the liquid inlet pipe in real time through the sensor And the outlet pipe pressure data , forming a time series.

[0074] Sampling time interval , get the time series .

[0075] S2.2, using the moving average method, the collected pressure data is noise filtered to obtain the smoothed pressure data of the liquid inlet pipe And the smoothed outlet pipe pressure data ;

[0076] ;

[0077] ;

[0078] in, Represents the index of the data point within the moving average window, that is, the index of each data point used to calculate the moving average at the current time The relative position of the index. From 0 to change.

[0079] Indicates the size of the moving average window, that is, the number of data points included in the time window used for smoothing.

[0080] Normalization is performed to map the data to the interval [0,1] for easy subsequent analysis.

[0081] S2.3, calculate the cross-correlation function of the inlet and outlet pressures , to determine the lag time between the two; ;

[0082] Using the sliding window method, find the real-time lag time corresponding to the maximum cross-correlation coefficient , as the time lag of the pressure in the inlet and outlet pipes:

[0083] ;

[0084] in, is the window size.

[0085] S2.4, based on the real-time lag time and real-time pressure data, build an adaptive prediction model and use the extended Kalman filter algorithm to predict the outlet pipe pressure:

[0086] ;

[0087] in, and are model parameters, which are iteratively updated by minimizing the prediction error.

[0088] The model parameters are iteratively updated by minimizing the prediction error to improve the prediction accuracy.

[0089] S2.5, calculate the difference between the actual outlet pipe pressure and the predicted pressure : ;

[0090] The output of the perfusion pump is adjusted according to the pressure difference. The proportional-integral-differential (PID) control algorithm is adopted to achieve dynamic regulation and precise control of the liquid flow by adjusting the proportional gain, integral gain and differential gain: ;

[0091] in, Indicates at time The adjustment amount of the irrigation pump output at all times is based on the real-time pressure difference and lag time changes, and is calculated by the PID control algorithm to adjust the parameters used to adjust the irrigation pump flow. Its purpose is to dynamically adjust the liquid irrigation rate according to the real-time monitored pressure changes to maintain the liquid balance in the uterine cavity during the operation, ensuring a clear view of the surgical area and the safety of the patient.

[0092] in, is the proportional gain, is the integral gain, is the differential gain.

[0093] S2.6, continuously monitor the pressure of the liquid inlet and outlet pipes, calculate the pressure difference and lag time in real time, and update the prediction model and feedback adjustment parameters.

[0094] If the difference between the actual outlet pipe pressure and the predicted pressure is within the pressure difference threshold range, it means that the system is operating normally, the liquid flow and pressure are within the controllable range, no additional intervention is required, and a normal signal is generated.

[0095] If the difference between the actual outlet pipe pressure and the predicted pressure exceeds the lower limit of the pressure difference threshold but does not reach the upper limit, it indicates that there may be a slight abnormality or unstable state. It is recommended to conduct detailed data analysis to determine potential problems and take corresponding measures to generate a signal to be analyzed.

[0096] If the difference between the actual outlet pipe pressure and the predicted pressure exceeds the upper limit of the pressure difference threshold, it indicates a serious abnormality or fault, such as leakage or equipment failure. It is necessary to immediately shut down the machine for inspection and take emergency repair measures, generating an abnormal signal.

[0097] The present invention collects and smoothes the pressure data of the inlet and outlet pipes in real time by setting the sampling frequency and the moving average method, ensuring the accuracy and consistency of the data. The cross-correlation function and the sliding window method are used to dynamically calculate and update the lag time to accurately reflect the real-time pressure changes. Based on this, an adaptive prediction model is constructed, and the extended Kalman filter algorithm is adopted to improve the accuracy of pressure prediction. Through the proportional-integral-differential control algorithm, the output of the perfusion pump is dynamically adjusted according to the difference between the actual pressure and the predicted pressure, so as to achieve accurate control and regulation of the liquid flow. The pressure difference and lag time are continuously monitored and calculated in real time, and the prediction model and feedback adjustment parameters are updated. According to the difference between the actual outlet pipe pressure and the predicted pressure, the system generates normal, to-be-analyzed and abnormal signals to ensure timely identification and response to potential problems, and effectively guarantee the accuracy of intrauterine fluid management and the safety of surgery.

[0098] During hysteroscopic surgery, accurate fluid management is critical to the success of the operation and the safety of the patient. The operation requires maintaining an appropriate amount of perfusion fluid in the uterine cavity to ensure a clear surgical field of view and prevent tissue damage. However, the flow characteristics of the fluid in the tube, as well as the possible leakage of body fluids or bleeding during the operation, make fluid management complicated. By real-time monitoring of the pressure of the inlet and outlet tubes, the state of fluid flow can be grasped, but due to the time lag of the flow of fluid in the tube, direct observation of the pressure data may not reflect the real situation in a timely manner.

[0099] The calculation method of the hysteresis time complexity index is introduced to quantify the dynamic characteristics of the liquid management system by analyzing the disorder and complexity of the pressure data. The hysteresis time complexity index provides a quantitative indicator for evaluating the disorder and complexity of the system by combining approximate entropy and multiscale entropy. The larger the index, the higher the disorder and complexity of the sequence. Conversely, the smaller the index, the lower the disorder and complexity of the sequence.

[0100] In practical applications, a higher hysteresis complexity index may mean that there are more unstable factors in the system, such as increased fluid leakage or pipeline blockage, which requires more detailed analysis and processing; a lower index may indicate that the system is running smoothly and the liquid flow and pressure are within the controllable range. Through this index, the dynamic characteristics of the liquid management system during surgery can be evaluated, helping medical staff to identify potential problems in a timely manner and take corresponding measures to ensure the smooth progress of the operation and the safety of the patient.

[0101] This approach not only improves the accuracy and responsiveness of fluid management, but also provides a scientific basis for abnormality detection and system optimization during surgery, ultimately improving the quality of surgery and the patient's treatment experience.

[0102] Step S3 includes the following contents:

[0103] The generated lagged time series is split according to the set window size to form multiple subsequences. For each subsequence, a disorder analysis is performed and the approximate entropy is calculated to quantify the disorder and unpredictability of the data series. At the same time, a complexity analysis is performed and multi-scale entropy is calculated to quantify the complexity of the data series at different scales. The benefit of this stage is that through multi-dimensional analysis, the characteristics of the lagged time series can be comprehensively evaluated and potential anomalies can be identified.

[0104] For the signal to be analyzed, the generated lag time series is divided according to the set window size to form multiple subsequences.

[0105] S3.1, for each subsequence, calculate the approximate entropy. The specific calculation formula is as follows:

[0106] For the length , defining the embedding dimension and similarity threshold .

[0107] The construction length is The embedding vector of : ;

[0108] in, The index of the data point.

[0109] For each embedding vector , calculate it and other embedding vectors distance.

[0110] If the distance between each embedding vector and other embedding vectors is less than a preset similarity threshold, the two are considered to be similar vectors.

[0111] Count the number of other embedding vectors similar to each embedding vector, divided by the total number of embedding vectors , get the similarity ratio ;

[0112] The approximate entropy is obtained by taking the logarithm of the similarity ratio of all embedding vectors and then averaging them. : ;

[0113] The calculation formula of approximate entropy quantifies the disorder and unpredictability of the data sequence by counting and processing the number of similar vectors, reflecting the complexity of the sequence.

[0114] S3.2, for each subsequence, calculate the multi-scale entropy. The specific calculation formula is as follows:

[0115] Multi-scale decomposition of lagged time series. , the sequence is segmented according to the time scale, and the average value of each segment is taken to form a new subsequence. Specifically, for each new data point , calculate the original sequence from arrive this The average of the data points.

[0116] ;

[0117] in, The new subsequence index.

[0118] For each subsequence obtained by multi-scale decomposition, calculate its entropy value. The entropy value reflects the complexity of the subsequence. The specific steps are as follows:

[0119] Count the probability of each different state appearing in the subsequence ;

[0120] For each state, calculate its probability multiplied by its log probability and mark it as the first result;

[0121] Sum the first results of all states and take the negative value, which is the entropy value of the subsequence, i.e., the multi-scale entropy .

[0122] Multiscale entropy quantifies the complexity of data series by decomposing and analyzing lagged time series at different time scales, reflecting the dynamic characteristics of the series at multiple time scales.

[0123] S3.3, using geometric mean to comprehensively calculate the approximate entropy and the multi-scale entropy and normalize them to obtain the lag time complexity index. For example, the product of the approximate entropy and the multi-scale entropy is calculated and square rooted, and then the lag time complexity index is calculated by max-min normalization.

[0124] The lag time complexity index is obtained by calculating the geometric mean of the approximate entropy and the multiscale entropy and by the maximum-minimum normalization method. The lag time complexity index is used to quantify the disorder and complexity of the lag time series. Specifically, the larger the index, the higher the disorder and complexity of the sequence, and vice versa, the smaller the index, the lower the disorder and complexity of the sequence. In the scenario of hysteroscopic fluid management, a higher lag time complexity index may mean that there are more unstable factors in the system, such as increased fluid exudation or pipeline blockage, which requires more detailed analysis and processing; while a lower index may indicate that the system is running smoothly and the fluid flow and pressure are within the controllable range. Through this index, the dynamic characteristics of the fluid management system during surgery can be evaluated, which helps to identify potential problems in a timely manner and take corresponding measures to ensure the smooth progress of the operation and the safety of the patient.

[0125] During hysteroscopic surgery, accurate fluid management is crucial for the success of the operation and the safety of the patient. The pressure fluctuation asymmetry index is used to quantify the asymmetry of the pressure fluctuations in the inlet and outlet tubes, providing important information about the dynamic characteristics of fluid flow. During surgery, the volatility of fluid flow may be affected by a variety of factors such as fluid exudation, pipeline obstruction, and equipment failure. These potential abnormalities can be identified by real-time monitoring and calculation of the pressure fluctuation asymmetry index. The larger the index, the stronger the asymmetry of the pressure fluctuation, which may mean that there is an abnormality in the system; the smaller the index, the more symmetrical the pressure fluctuation and the smoother the system operation. The introduction and application of this indicator can help medical staff to promptly detect problems in the fluid management system, take corresponding measures, ensure the smooth progress of the operation and the safety of the patient, and improve the overall surgical quality and treatment effect. The calculation and analysis of the pressure fluctuation asymmetry index provides a scientific basis for abnormality detection and system optimization during surgery, and has important clinical application prospects.

[0126] Step S4 includes the following contents:

[0127] For the signal to be analyzed, the sampling frequency is set, and the pressure data of the liquid inlet pipe and the liquid outlet pipe are collected in real time through the sensor to form a time series;

[0128] The moving average method is used to filter and smooth the data to obtain smoothed pressure data;

[0129] S4.1, split the pressure time series into multiple time windows, each with a length of ;

[0130] For each time window, calculate the positive fluctuations in the pressure series, that is, the sum of the rising parts of the data points ;

[0131] in, Indicates time window.

[0132] For each time window, calculate the negative fluctuations in the pressure series, that is, the sum of the falling parts of the data points ;

[0133] S4.2, calculate the difference between positive and negative fluctuations in each time window to obtain the fluctuation asymmetry ;

[0134] The fluctuation asymmetry of all time windows is averaged and normalized to obtain the overall pressure fluctuation asymmetry index.

[0135] The pressure fluctuation asymmetry index is used to quantify the symmetry of pressure changes. The larger the index, the stronger the asymmetry of the pressure fluctuation, which may mean that there is an abnormality in the system; the smaller the index, the more symmetrical the pressure fluctuation and the smoother the system operation. Through this index, potential problems in the fluid management system can be identified and evaluated in a timely manner to improve the safety and effectiveness of surgery.

[0136] Based on multi-sensor data fusion and adaptive delay compensation algorithm, the pressure of the inlet and outlet pipes is monitored in real time, the extended Kalman filter is used to predict the pressure of the outlet pipe, and the flow is adjusted through proportional-integral-differential control. After dynamically updating the lag time and presenting the corresponding indication signal, the leakage risk warning index is further calculated to analyze the leakage risk. The significance lies in providing a more comprehensive and in-depth risk assessment. Although the above methods can monitor and adjust the system status in real time, they mainly focus on the current operating parameters and immediate response. By calculating the lag time complexity index and the pressure fluctuation asymmetry index, and comprehensively forming the leakage risk warning index, the disorder of the system and the abnormal situation of pressure fluctuation can be quantified, thereby identifying potential leakage risks. The necessity of doing so is that simple real-time monitoring and immediate adjustment cannot fully capture the complex dynamic changes within the system. The leakage risk warning index provides a deeper understanding of the system status and the ability to warn in advance through more comprehensive data analysis and comprehensive evaluation, ensuring the smooth progress of the operation and the safety of the patient, and further improving the reliability of the system and the quality of surgery.

[0137] Step S5 includes the following contents:

[0138] Let the lag time complexity index be , the pressure fluctuation asymmetry index is In order to evaluate the leakage risk, the two indexes are analyzed comprehensively, and the leakage risk warning index is obtained by combining the cross entropy and geometric mean method. , for example, the comprehensive calculation formula is: ;

[0139] Cross entropy is used to measure the difference between two probability distributions, and is used here to combine two different physical quantities. The geometric mean is used to balance the influence of the two indicators to avoid a single indicator dominating the comprehensive result.

[0140] The purpose of taking a negative value is to convert the indicator into a risk assessment index, which is more in line with practical significance. High risk corresponds to a high index value, and low risk corresponds to a low index value.

[0141] Cross entropy part The logarithmic values ​​of two different indicators are combined to reflect the interaction and joint impact between the two indicators.

[0142] Using Square Root A geometric mean is performed to balance the impact of the two indicators and ensure that the combined result is not overly dependent on any single indicator.

[0143] The leakage risk warning index is used to evaluate and indicate the leakage risk of the liquid management system during hysteroscopic surgery. The index combines the hysteresis time complexity index and the pressure fluctuation asymmetry index to quantify the instability of the system and abnormal pressure fluctuations. The larger the index, the stronger the instability and abnormal pressure fluctuations of the system, the higher the risk of leakage, and there may be problems such as pipeline blockage, increased fluid exudation or equipment failure; the smaller the index, the smoother the system operation, the lower the risk of leakage, and the liquid flow and pressure are within the controllable range. Through this index, medical staff can promptly identify and evaluate potential problems in the liquid management system, take corresponding measures to ensure the smooth progress of the operation and the safety of the patient, and improve the quality of surgery and treatment effect.

[0144] Compare the leakage risk warning index with the risk threshold. If the leakage risk warning index is greater than or equal to the risk threshold, it means that the system has a high risk of leakage, and there may be problems such as pipeline obstruction, increased body fluid exudation, or equipment failure. Appropriate inspection and intervention measures need to be taken immediately to ensure the smooth progress of the operation and the safety of the patient. On the contrary, if the leakage risk warning index is less than the risk threshold, it means that the system is running relatively smoothly, the liquid flow and pressure are within the controllable range, the leakage risk is low, the operation can continue, and medical staff only need to perform routine monitoring. This risk assessment mechanism helps medical staff identify and respond to potential risks in a timely manner by comparing the index and threshold in real time, thereby improving the quality of surgery and patient safety.

[0145] The present invention divides the lag time series and calculates the approximate entropy and multi-scale entropy to comprehensively obtain the lag time complexity index, and then combines the real-time acquisition and smoothing of pressure data, calculates the difference between positive and negative fluctuations by dividing the time window, and obtains the pressure fluctuation asymmetry index. Finally, the cross entropy combined with the geometric mean method is used to comprehensively calculate the leakage risk warning index, comprehensively quantify the disorder and abnormal pressure fluctuations of the system, and provide a deeper risk assessment. This series of processing steps and comprehensive analysis can not only monitor and adjust the system state in real time, but also identify potential leakage risks through more comprehensive data analysis and comprehensive evaluation. By further calculating the leakage risk warning index, it is possible to combine data from multiple dimensions to quantify the complex dynamic changes of the system, thereby providing a deeper understanding of the system state and the ability to warn in advance. This multi-level and multi-dimensional comprehensive evaluation mechanism helps medical staff to identify and respond to potential risks in a timely manner, take corresponding measures, ensure the smooth progress of the operation and the safety of the patient, and further improve the reliability of the system and the quality of surgery.

[0146] Embodiment 2: Figure 2 The present invention provides a hysteroscopic liquid management leakage detection system, comprising: a self-check calibration module, a data acquisition module, a hysteresis analysis module, a complexity analysis module, a fluctuation analysis module and a risk assessment module;

[0147] Before perfusing the liquid, the self-checking and calibration module performs self-checking and calibration on the sensors and perfusion system at the ports of the inlet and outlet pipes, and records each calibration data, which is then transmitted to the data acquisition module;

[0148] The data acquisition module sets the sampling frequency, collects and smoothes the pressure data of the liquid inlet pipe and the liquid outlet pipe, and transmits the smoothed pressure data to the lag analysis module and the fluctuation analysis module;

[0149] The hysteresis analysis module calculates the cross-correlation function to determine the hysteresis time, constructs an adaptive prediction model to predict the pressure of the liquid outlet pipe, adjusts the output of the perfusion pump through the PID control algorithm, continuously monitors and updates the model, generates a corresponding indication signal, and if the obtained indication signal is a signal to be processed, the data is passed to the complexity analysis module and the fluctuation analysis module;

[0150] The complexity analysis module divides the lag time series according to the set window size, calculates the approximate entropy and multi-scale entropy of each subsequence, calculates the lag time complexity index through geometric mean and normalization, quantifies the disorder and complexity of the sequence, and passes the lag time complexity index to the risk assessment module;

[0151] The fluctuation analysis module collects and smoothes the pressure data in real time, divides the time series into multiple windows, calculates the difference between positive and negative fluctuations in each window, normalizes it to obtain the pressure fluctuation asymmetry index, quantifies the symmetry of pressure changes, and passes the pressure fluctuation asymmetry index to the risk assessment module;

[0152] The risk assessment module is based on the hysteresis time complexity index and the pressure fluctuation asymmetry index. It uses the cross entropy combined with the geometric mean method to comprehensively calculate the leakage risk warning index, quantify the instability of the system and abnormal pressure fluctuations, assess the leakage risk and take corresponding measures.

[0153] The above formulas are all dimensionless and numerical calculations. The formula is a formula for the most recent real situation obtained by collecting a large amount of data and performing software simulation. The preset parameters in the formula are set by technicians in this field according to actual conditions.

[0154] The above description is only by way of illustration of certain exemplary embodiments of the present invention. It is undoubted that those skilled in the art can modify the described embodiments in various ways without departing from the spirit and scope of the present invention. Therefore, the above drawings and descriptions are illustrative in nature and should not be construed as limiting the scope of protection of the claims of the present invention.

[0155] It should be noted that, in this article, if there are relational terms such as first and second, etc., they are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "including a..." do not exclude the existence of other identical elements in the process, method, article or device including the elements.

[0156] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Claims

1. A method for detecting leakage of hysteroscopic fluid management, characterized in that: S1, before filling the liquid, perform self-check and calibration on the sensors and filling system at the ports of the inlet and outlet pipes, and record the calibration data each time; S2, set the sampling frequency, collect and smooth the pressure data of the inlet pipe and the outlet pipe, calculate the cross-correlation function to determine the lag time, build an adaptive prediction model to predict the outlet pipe pressure, adjust the perfusion pump output through the PID control algorithm, continuously monitor and update the model, generate a corresponding indication signal, and if the obtained indication signal is a signal to be processed, synchronously execute step S3 and step S4; S3, splits the lag time series according to the set window size, calculates the approximate entropy and multi-scale entropy of each subsequence, calculates the lag time complexity index by geometric mean and normalization, and quantifies the disorder and complexity of the series; S4, real-time acquisition and smoothing of pressure data, dividing the time series into multiple windows, calculating the difference between positive and negative fluctuations in each window, normalizing it to obtain the pressure fluctuation asymmetry index, and quantifying the symmetry of pressure changes; S5, based on the hysteresis time complexity index and the pressure fluctuation asymmetry index, uses the cross entropy combined with the geometric mean method to comprehensively calculate the leakage risk warning index, quantify the instability of the system and abnormal pressure fluctuations, evaluate the leakage risk and take corresponding measures.

2. A hysteroscopic fluid management leakage detection method according to claim 1, characterized in that: S2 includes the following: S2.1, set the sampling frequency, collect the pressure data of the liquid inlet pipe and the liquid outlet pipe in real time through the sensor, and form a time series; S2.2, using a moving average method to perform noise filtering on the collected pressure data to obtain smoothed inlet pipe pressure data and smoothed outlet pipe pressure data; S2.3, calculating the cross-correlation function of the inlet and outlet pressures to determine the lag time between the two; The sliding window method is used to select the real-time lag time corresponding to the maximum cross-correlation coefficient as the time lag of the pressure of the inlet pipe and the outlet pipe; S2.4, based on the real-time lag time and real-time pressure data, an adaptive prediction model is constructed and an extended Kalman filter algorithm is used to predict the outlet pipe pressure; S2.5, calculating the difference between the actual outlet pipe pressure and the predicted pressure; The output of the perfusion pump is adjusted according to the pressure difference. The proportional-integral-differential control algorithm is adopted to achieve dynamic regulation and precise control of the liquid flow by adjusting the proportional gain, integral gain and differential gain. S2.6, if the difference between the actual outlet pipe pressure and the predicted pressure is within the pressure difference threshold range, a normal signal is generated; If the difference between the actual outlet pipe pressure and the predicted pressure exceeds the lower limit of the pressure difference threshold but does not reach the upper limit, a signal to be analyzed is generated; If the difference between the actual outlet pipe pressure and the predicted pressure exceeds the upper limit of the pressure difference threshold, an abnormal signal is generated.

3. A hysteroscopic fluid management leakage detection method according to claim 2, characterized in that: Step S3 includes the following contents: For the signal to be analyzed, the generated lag time series is divided according to the set window size to form multiple subsequences; S3.1, for each subsequence, calculate the approximate entropy; S3.2, for each subsequence, calculate the multi-scale entropy; S3.3, the approximate entropy and multiscale entropy are calculated and normalized using the geometric mean to obtain the lag time complexity index.

4. A hysteroscopic fluid management leakage detection method according to claim 3, characterized in that: Step S4 includes the following contents: For the signal to be analyzed, the moving average method is used to perform noise filtering and smoothing on the data to obtain smoothed pressure data; S4.1, split the pressure time series into multiple time windows; For each time window, the positive fluctuations in the pressure series are calculated, i.e., the sum of the rising parts of the data points; For each time window, calculate the negative fluctuations in the pressure series, i.e., the sum of the falling parts of the data points; S4.2, calculate the difference between positive and negative fluctuations in each time window to obtain the fluctuation asymmetry; The fluctuation asymmetry of all time windows is averaged and normalized to obtain the overall pressure fluctuation asymmetry index.

5. A hysteroscopic fluid management leakage detection method according to claim 4, characterized in that: Step S5 includes the following contents: Compare the leakage risk warning index with the risk threshold. If the leakage risk warning index is greater than or equal to the risk threshold, appropriate inspection and intervention measures need to be taken immediately. On the contrary, if the leakage risk warning index is less than the risk threshold.

6. A hysteroscopic fluid management leakage detection method according to claim 3, characterized in that: The specific calculation steps of step S3.1 are as follows: For each embedding vector, calculate its distance to other embedding vectors; If the distance between each embedding vector and other embedding vectors is less than the preset similarity threshold, the two are considered to be similar vectors; Count the number of other embedding vectors similar to each embedding vector, divide it by the total number of embedding vectors, and get the similarity ratio; The approximate entropy is obtained by taking the logarithm of the similarity ratio of all embedding vectors and then averaging them.

7. A hysteroscopic fluid management leakage detection method according to claim 3, characterized in that: The specific calculation steps of step S3.2 are as follows: Perform multi-scale decomposition on the lagged time series. According to a given time scale, the series is segmented according to the time scale, and the average value of each segment is taken to form a new subsequence. For each subsequence obtained by multi-scale decomposition, calculate its entropy value. The specific steps are as follows: Count the probability of each different state appearing in the subsequence; For each state, calculate its probability multiplied by its log probability and mark it as the first result; Sum the first results of all states and take the negative value, which is the entropy value of the corresponding subsequence, that is, the multi-scale entropy.

8. A hysteroscopic liquid management leakage detection system, used to implement a hysteroscopic liquid management leakage detection method according to any one of claims 1 to 7, characterized in that: include: Self-check calibration module, data acquisition module, hysteresis analysis module, complexity analysis module, fluctuation analysis module and risk assessment module; Before filling the liquid, the self-checking and calibration module performs self-checking and calibration on the sensors and the filling system at the ports of the liquid inlet and outlet pipes, and records each calibration data, which is then transmitted to the data acquisition module; The data acquisition module sets the sampling frequency, collects and smoothes the pressure data of the liquid inlet pipe and the liquid outlet pipe, and transmits the smoothed pressure data to the lag analysis module and the fluctuation analysis module; The hysteresis analysis module calculates the cross-correlation function to determine the hysteresis time, constructs an adaptive prediction model to predict the pressure of the liquid outlet pipe, adjusts the output of the perfusion pump through the PID control algorithm, continuously monitors and updates the model, generates a corresponding indication signal, and if the obtained indication signal is a signal to be processed, the data is passed to the complexity analysis module and the fluctuation analysis module; The complexity analysis module divides the lag time series according to the set window size, calculates the approximate entropy and multi-scale entropy of each subsequence, calculates the lag time complexity index through geometric mean and normalization, quantifies the disorder and complexity of the sequence, and passes the lag time complexity index to the risk assessment module; The fluctuation analysis module collects and smoothes the pressure data in real time, divides the time series into multiple windows, calculates the difference between positive and negative fluctuations in each window, normalizes it to obtain the pressure fluctuation asymmetry index, quantifies the symmetry of pressure changes, and passes the pressure fluctuation asymmetry index to the risk assessment module; The risk assessment module is based on the hysteresis time complexity index and the pressure fluctuation asymmetry index. It uses the cross entropy combined with the geometric mean method to comprehensively calculate the leakage risk warning index, quantify the instability of the system and abnormal pressure fluctuations, assess the leakage risk and take corresponding measures.

Images