Aortic septal defect detection method based on self-adjusting dynamic Gaussian mixture model

Abstract

The invention discloses an aortic septal defect detection method based on a self-adjusting dynamic Gaussian mixture model. The aortic septal defect detection method comprises the following steps: creating a second heart sound self-adaptive segmentation algorithm; constructing a second heart sound model based on the self-adjusting dynamic Gaussian mixture model; and calculating a second heart sound splitting coefficient S2split and a second heart sound component time interval # imgabs0 # based on a second heart sound model, and effectively fusing the splitting coefficient and the time interval statistic to realize aortic septal defect detection. By adopting the method, the accuracy and timeliness of aortic septal defect detection are improved based on effective fusion of the splitting coefficient and the time interval statistic of the self-adjusting dynamic Gaussian mixture model.

Description

Technical Field

[0001] The present invention relates to the technical field of medical signal processing and analysis, and in particular to an aortic septal defect detection method based on a self-regulating dynamic Gaussian mixture model. Background Art

[0002] In medicine, heart sounds are important physiological signals reflecting cardiac function, and their analysis is crucial for the diagnosis of heart disease. Second heart sound splitting is a typical clinical manifestation of some heart diseases, and accurate estimation of the second heart sound splitting coefficient plays a key role in early detection and timely intervention.

[0003] However, existing heart sound analysis methods are mostly based on static modeling, lacking the ability to respond to dynamic changes and failing to meet the demands of real-time analysis. In particular, the commonly used least-squares Gaussian fitting method for modeling the second heart sound signal has significant limitations. On the one hand, this method suffers from a complex modeling process and high computational overhead, resulting in high overall time complexity and difficulty in efficiently processing massive amounts of heart sound data. On the other hand, the model updates lag, making it unable to effectively adapt to the dynamic changes in the heart sound signal. This affects the accuracy of the second heart sound splitting coefficient estimation and limits its application in real-time clinical monitoring.

[0004] Therefore, it is urgent to propose a second heart sound splitting coefficient estimation method with real-time, high precision and low computational complexity to meet the actual needs of dynamic monitoring and intelligent diagnosis of heart disease. Summary of the Invention

[0005] The purpose of the present invention is to provide an aortic septal defect detection method based on a self-adjusting dynamic Gaussian mixture model to solve the problems of poor modeling real-time performance, high time complexity and insufficient estimation accuracy when estimating the second heart sound splitting coefficient of the Gaussian fitting method based on the least squares method.

[0006] To achieve the above objectives, the present invention provides a method for detecting aortic septal defect based on a self-adjusting dynamic Gaussian mixture model, comprising the following steps:

[0007] S1. Collect heart sound data, pre-process the heart sound data, adaptively segment the heart sound data, and adaptively identify and extract the second heart sound;

[0008] S2. Based on the second heart sound generation mechanism and components, combined with the characteristics of the dynamic incremental Gaussian mixture model, a second heart sound model based on the self-regulating dynamic Gaussian mixture model is constructed; based on the heart sound splitting generation mechanism and shape characteristics, the second heart sound splitting coefficient S is calculated based on the second heart sound model. 2split and the interval between components of the second heart sound

[0009] S3. Based on the splitting characteristics of the second heart sound, such as wide splitting, physiological splitting and reverse splitting, the splitting coefficient and the time interval statistics are fused to realize the aortic septal defect detection.

[0010] Preferably, the constructing of the second heart sound model based on the self-adjusting dynamic Gaussian mixture model includes:

[0011] A Gaussian mixture model with two fixed components is established, and the formula is:

[0012]

[0013] in, is a random time coordinate, is the corresponding model amplitude value, π k , μ k , are the weight coefficient, center and divergence, and probability density function of the kth Gaussian component respectively;

[0014] Based on the second heart sound data Generate random data Among them, x i is the time coordinate, y i is the corresponding amplitude value, [x i ,x i+1 ] to generate equal sample points between in N i =y i / max(y i ,i=1,…,N)×N, where N represents the number of second heart sound samples;

[0015] By introducing the self-regulating factor and combining the characteristics of the dynamic incremental Gaussian mixture model, a second heart sound model based on the self-regulating dynamic Gaussian mixture model is constructed, including:

[0016] Initialize the self-regulating factor α and obtain the initial model parameters

[0017] For sample points The self-regulating factor α is used to determine the Mth Gaussian component to be updated. The formula is:

[0018]

[0019] If M is empty and the current number of components is less than 2, a new component is created. The formula is:

[0020] sp2=1;

[0021] If M is empty and the current number of components is equal to 2, then decrease α = α - δα , ensure that the number of output components is 2;

[0022] If M is non-empty, update component k∈M.

[0023] Preferably, the component updating process includes:

[0024] Calculate its conditional probability density, the formula is:

[0025]

[0026] Calculate the posterior probability estimate, the formula is:

[0027]

[0028] Secondly, update the cumulative probability SP k And calculate the cumulative rate of change ω k , the formula is:

[0029]

[0030] Finally, update the model parameters and generate the model The formula is:

[0031]

[0032] Preferably, in step S2, the second heart sound splitting coefficient S is calculated based on the second heart sound model. 2split and the interval between components of the second heart sound include:

[0033] According to the second heart sound model The minimum sum between the two Gaussian components The ratio of the lowest amplitude values in the components defines the second heart sound splitting coefficient S 2split , its minimum value and ratio are complementary, and the calculation formula is:

[0034]

[0035] The time interval between the second heart sound components is determined by using the amplitude value corresponding to the center point of the aortic component to be greater than the amplitude value corresponding to the center point of the pulmonary artery component. The center position of the component corresponding to the second heart sound, and then the time interval between the components of the second heart sound To calculate, the formula is:

[0036]

[0037] Preferably, step S3 specifically includes:

[0038] Calculate the statistics of the splitting coefficient within the respiratory cycle, the formula is:

[0039]

[0040] Where, Respectively represent the mean and variance of the splitting coefficient within the respiratory cycle; S 2split [i] represents the splitting coefficient of the i-th second heart sound in the respiratory cycle;

[0041] Compute statistics for intervals within a respiratory cycle:

[0042]

[0043] Where, It represents the mean time interval within the respiratory cycle; represents the i-th time interval in the respiratory cycle;

[0044] The second heart sound splitting coefficient and the time interval statistics between the second heart sound components are integrated to establish a detection scheme for detecting aortic septal defects. The formula is:

[0045]

[0046] Therefore, the present invention adopts the above-mentioned aortic septal defect detection method based on the self-adjusting dynamic Gaussian mixture model, which has the following beneficial effects:

[0047] (1) Improve detection accuracy: By constructing a second heart sound model based on a self-regulating dynamic Gaussian mixture model, the heart sound data can be fitted more accurately, thereby more accurately estimating the second heart sound splitting coefficient S2 and the time interval between the second heart sound aortic valve closure sound A2 and the pulmonary valve closure sound P2. These two parameters play a key role in the diagnosis of aortic septal defect (ASD), improving the accuracy and reliability of detection;

[0048] (2) Real-time monitoring: The incremental mode is used to achieve real-time modeling of the second heart sound, which enables the method to process and analyze heart sound data in real time and meet the needs of clinical real-time monitoring. This is of great significance for the timely detection and intervention of heart diseases, especially in emergency situations or in the management of patients who require long-term monitoring.

[0049] (3) Reduced time complexity: By dynamically updating the parameters of the Gaussian mixture model, the amount of computation and time complexity are reduced, which makes the method more efficient in processing large amounts of heart sound data, can quickly provide diagnostic results, and improves the feasibility and efficiency of clinical applications;

[0050] (4) Adaptation to individual differences: This method can automatically adjust the model parameters according to the input data to adapt to the changes in heart sound characteristics of different individuals and different physiological states. This makes the method widely applicable and can accurately detect aortic septal defects under various physiological conditions.

[0051] The technical solution of the present invention is further described in detail below through the accompanying drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] Figure 1 is a flow chart of a method according to an embodiment of the present invention;

[0053] Figure 2 A comparison diagram of the heart sound filtering effect in an embodiment of the present invention;

[0054] Figure 3 This is a modeling diagram of the second heart sound model according to an embodiment of the present invention;

[0055] Figure 4 This is a diagram of the process of modeling the second heart sound model according to an embodiment of the present invention;

[0056] Figure 5 The best classifier trained for this embodiment of the present invention is used to diagnose ASD, other splitting and third heart sounds. DETAILED DESCRIPTION

[0057] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all of the embodiments. The components of the embodiments of the present invention generally described and shown in the drawings herein can be arranged and designed in various different configurations. In the description of the present invention, it should be noted that the orientation or position relationship indicated by the terms "upper", "lower", "inside", "outside", etc. is based on the orientation or position relationship shown in the drawings, or is the orientation or position relationship in which the product of the invention is usually placed when in use. It is only for the convenience of describing the present invention and simplifying the description, and does not indicate or imply that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as a limitation on the present invention.

[0058] Example

[0059] Reference Figure 1 The present invention provides a method for detecting aortic septal defect based on a self-adjusting dynamic Gaussian mixture model, the steps comprising:

[0060] S1. Collect heart sound data, pre-process the heart sound data, adaptively segment the heart sound data, and adaptively identify and extract the second heart sound.

[0061] S2. Based on the second heart sound generation mechanism and components, combined with the characteristics of the dynamic incremental Gaussian mixture model, a second heart sound model based on the self-regulating dynamic Gaussian mixture model is constructed; based on the heart sound splitting generation mechanism and shape characteristics, the second heart sound splitting coefficient S is calculated based on the second heart sound model. 2split and the interval between components of the second heart sound

[0062] S3. Based on the splitting characteristics of the second heart sound, such as wide splitting, physiological splitting and reverse splitting, the splitting coefficient and the time interval statistics are fused to realize the aortic septal defect detection.

[0063] In this embodiment, step S1 specifically includes:

[0064] An electronic stethoscope is used to collect periodic sound wave data (heart sound signals) caused by myocardial contraction, heart valve closure, and blood hitting the ventricular wall and aorta wall.

[0065] Wavelet filtering algorithm is used to realize heart sound preprocessing. Figure 2 As shown in the figure, it is a comparison diagram of the heart sound effects before and after filtering, where A is the time domain waveform before filtering, B is the frequency domain waveform before filtering, C is the time domain waveform after filtering, and D is the frequency domain waveform after filtering.

[0066] Adaptive recognition and extraction of the second heart sound are achieved based on the shape features of the components of the periodic heart sound signal (the first heart sound and the second heart sound).

[0067] In this embodiment, refer to Figure 3 , step S2 specifically includes:

[0068] Construct a second heart sound model based on a self-regulating dynamic Gaussian mixture model, including:

[0069] A Gaussian mixture model with two fixed components is established, and the formula is:

[0070]

[0071] in, is a random time coordinate, is the corresponding model amplitude value, π k , μ k , σ k , are the weight coefficient, center and divergence, and probability density function of the kth Gaussian component respectively;

[0072] Based on the second heart sound data Generate random data Among them, x i is the time coordinate, y i is the corresponding amplitude value, [xi ,x i+1 ] to generate equal sample points between in N i =y i / max(y i ,i=1,…,N)×N, where N represents the number of second heart sound samples;

[0073] By introducing the self-regulating factor and combining the characteristics of the dynamic incremental Gaussian mixture model, a second heart sound model based on the self-regulating dynamic Gaussian mixture model is constructed, such as Figure 4 The figure shows the process of the self-regulating factor α regulating the fraction.

[0074] The specific process includes:

[0075] Initialize the self-regulating factor α and obtain the initial model parameters

[0076] For sample points The self-regulating factor α is used to determine the Mth Gaussian component to be updated. The formula is:

[0077]

[0078] If M is empty and the current number of components is less than 2, a new component is created. The formula is:

[0079] sp2=1;

[0080] If M is empty and the current number of components is equal to 2, then decrease α = α - δ α , ensure that the number of output components is 2;

[0081] If M is non-empty, update the component k∈M. The component update process includes:

[0082] Calculate its conditional probability density, the formula is:

[0083]

[0084] Calculate the posterior probability estimate, the formula is:

[0085]

[0086] Secondly, update the cumulative probability SP k And calculate the cumulative rate of change ω k , the formula is:

[0087]

[0088] Finally, update the model parameters and generate the model The formula is:

[0089]

[0090] In this embodiment, the second heart sound splitting coefficient S is calculated based on the second heart sound model. 2split and the interval between components of the second heart sound include:

[0091] According to the second heart sound model The minimum sum between the two Gaussian components The ratio of the lowest amplitude values in the components defines the second heart sound splitting coefficient S 2split , its minimum value and ratio are complementary, and the calculation formula is:

[0092]

[0093] The time interval between the second heart sound components is determined by using the amplitude value corresponding to the center point of the aortic component to be greater than the amplitude value corresponding to the center point of the pulmonary artery component. The center position of the component corresponding to the second heart sound, and then the time interval between the components of the second heart sound To calculate, the formula is:

[0094]

[0095] In this embodiment, Figure 5 As shown, for the aortic septal defect detection classifier, step S3 specifically includes:

[0096] Calculate the statistics of the splitting coefficient within the respiratory cycle, the formula is:

[0097]

[0098] Where, Respectively represent the mean and variance of the splitting coefficient within the respiratory cycle; S 2split [i] represents the splitting coefficient of the i-th second heart sound in the respiratory cycle;

[0099] Compute statistics for intervals within a respiratory cycle:

[0100]

[0101] Where, It represents the mean time interval within the respiratory cycle; represents the i-th time interval in the respiratory cycle;

[0102] The second heart sound splitting coefficient and the time interval statistics between the second heart sound components are integrated to establish a detection scheme for detecting aortic septal defects. The formula is:

[0103]

[0104] Therefore, the present invention adopts the above-mentioned aortic septal defect detection method based on the self-adjusting dynamic Gaussian mixture model, and effectively integrates the splitting coefficient and time interval statistics of the self-adjusting dynamic Gaussian mixture model to improve the accuracy and timeliness of aortic septal defect detection.

[0105] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit the same. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that they can still modify or replace the technical solutions of the present invention with equivalents, and these modifications or equivalent replacements cannot cause the modified technical solutions to deviate from the spirit and scope of the technical solutions of the present invention.

Claims

1. A method for detecting aortic septal defect based on a self-regulating dynamic Gaussian mixture model, characterized in that the steps include: S1. Collect heart sound data, pre-process the heart sound data, adaptively segment the heart sound data, and adaptively identify and extract the second heart sound; S2. Based on the second heart sound generation mechanism and components, combined with the characteristics of the dynamic incremental Gaussian mixture model, a second heart sound model based on the self-regulating dynamic Gaussian mixture model is constructed; based on the heart sound splitting generation mechanism and shape characteristics, the second heart sound splitting coefficient S is calculated based on the second heart sound model. 2split and the interval between components of the second heart sound S3. Based on the splitting characteristics of the second heart sound, namely, wide splitting, physiological splitting, and reverse splitting, the second heart sound splitting coefficient and the time interval statistics between the second heart sound components are fused to realize aortic septal defect detection.

2. The aortic septal defect detection method based on a self-adjusting dynamic Gaussian mixture model according to claim 1, characterized in that: The constructing of the second heart sound model based on the self-adjusting dynamic Gaussian mixture model includes: A Gaussian mixture model with two fixed components is established, and the formula is: in, is a random time coordinate, is the corresponding model amplitude value, π k , μ k , are the weight coefficient, center and divergence, and probability density function of the kth Gaussian component respectively; Based on the second heart sound data Generate random data Among them, x i is the time coordinate, y i is the corresponding amplitude value, [x i ,x i+1 ] to generate equal sample points between in N i =y i / max(y i ,i=1,…,N)×N, where N represents the number of second heart sound samples; By introducing the self-regulating factor and combining the characteristics of the dynamic incremental Gaussian mixture model, a second heart sound model based on the self-regulating dynamic Gaussian mixture model is constructed, including: Initialize the self-regulating factor α and obtain the initial model parameters For sample points The self-regulating factor α is used to determine the Mth Gaussian component to be updated. The formula is: If M is empty and the current number of components is less than 2, a new component is created. The formula is: If M is empty and the current number of components is equal to 2, then decrease α = α - δ α , ensure that the number of output components is 2; If M is non-empty, update component k∈M.

3. The aortic septal defect detection method based on a self-adjusting dynamic Gaussian mixture model according to claim 2, characterized in that: The update component process includes: Calculate its conditional probability density, the formula is: Calculate the posterior probability estimate, the formula is: Secondly, update the cumulative probability SP k And calculate the cumulative rate of change ω k , the formula is: Finally, update the model parameters and generate the model The formula is:

4. The aortic septal defect detection method based on a self-adjusting dynamic Gaussian mixture model according to claim 3, characterized in that: In step S2, the second heart sound splitting coefficient S is calculated based on the second heart sound model. 2split and the interval between components of the second heart sound include: According to the second heart sound model The minimum sum between the two Gaussian components The ratio of the lowest amplitude values in the components defines the second heart sound splitting coefficient S 2split , its minimum value and ratio are complementary, and the calculation formula is: The time interval between the second heart sound components is determined by using the amplitude value corresponding to the center point of the aortic component to be greater than the amplitude value corresponding to the center point of the pulmonary artery component. The center position of the component corresponding to the second heart sound, and then the time interval between the components of the second heart sound To calculate, the formula is:

5. The aortic septal defect detection method based on a self-adjusting dynamic Gaussian mixture model according to claim 4, characterized in that: Step S3 specifically includes: Calculate the statistics of the splitting coefficient within the respiratory cycle, the formula is: Where, Respectively represent the mean and variance of the splitting coefficient within the respiratory cycle; S 2split [i] represents the splitting coefficient of the i-th second heart sound in the respiratory cycle; Compute statistics for intervals within a respiratory cycle: Where, It represents the mean time interval within the respiratory cycle; represents the i-th time interval in the respiratory cycle; The second heart sound splitting coefficient and the time interval statistics between the second heart sound components are integrated to establish a detection scheme for detecting aortic septal defects. The formula is:

Images