A multi-item combined detection method for the urinary system

Abstract

The present invention relates to the technical field of urinary system joint detection, and in particular to a method for joint detection of multiple items of the urinary system. The method comprises: collecting and analyzing substances in a urine sample in a multivariate detection channel to obtain multivariate biochemical indicators; digitally converting the multivariate biochemical indicators to obtain converted multivariate biochemical indicator data; performing synchronous alignment processing on the converted multivariate biochemical indicator data, and then extracting multiscale features to obtain multiscale feature data; fusing the multiscale feature data to obtain fused feature data; introducing a multi-stage feedback control anomaly detection algorithm, performing anomaly detection on the fused feature data, and performing comprehensive discrimination evaluation to generate an evaluation result of the urine sample. The method solves the technical problems of the traditional urine detection method in single detection of the urinary system, inaccurate processing of indicator data during the detection process, and low detection accuracy.

Description

Technical Field

[0001] The present invention relates to the technical field of urinary system combined detection, and in particular to a urinary system multi-item combined detection method. Background Art

[0002] Among existing urinary system testing technologies, urine serves as an important clinical sample and is commonly used to monitor conditions such as kidney function, urinary tract infections, and abnormal protein metabolism. However, currently widely used urine tests often utilize urine test strips and semi-automated analyzers. Their core mechanisms rely on colorimetric reactions or simple conductivity measurements, covering only a limited number of basic indicators (such as protein, glucose, and urobilinogen). Furthermore, these indicators are often obtained independently, lacking a unified analytical framework. This makes it impossible to jointly assess multiple physiologically synergistic indicators in a single test. These methods present significant limitations, primarily manifested in technical issues such as a single detection dimension, inaccurate indicator data processing, poor timeliness, and low test result accuracy. These limitations can easily lead to missed detections and misdiagnosis in practical applications, particularly in clinical scenarios such as early renal injury and occult urinary tract infections, where they struggle to provide sufficient information support. Furthermore, with the advancement of multi-channel sensing technology, some studies have attempted to integrate multiple detection elements into microfluidic platforms to enable simultaneous detection of multiple indicators. Summary of the Invention

[0003] The present invention provides a multi-item combined detection method for the urinary system to solve the technical problems of traditional urine detection methods in single detection of the urinary system, inaccurate processing of indicator data during the detection process, and low detection accuracy.

[0004] The present invention provides a multi-item combined detection method for the urinary system, which specifically includes the following technical solutions:

[0005] A multi-item combined detection method for the urinary system comprises the following steps:

[0006] S1. Collecting and analyzing substances in a urine sample in a multivariate detection channel to obtain multivariate biochemical indicators; digitally converting the multivariate biochemical indicators to obtain converted multivariate biochemical indicator data; performing synchronous alignment processing on the converted multivariate biochemical indicator data, and then performing multiscale feature extraction to obtain multiscale feature data;

[0007] S2. Fuse the multi-scale feature data to obtain fused feature data; introduce a multi-stage feedback control anomaly detection algorithm to perform anomaly detection on the fused feature data, and perform comprehensive discrimination evaluation to generate an evaluation result of the urine sample.

[0008] Preferably, the S1 specifically includes:

[0009] The converted multivariate biochemical indicator data in each detection channel are synchronously aligned to obtain the synchronously aligned multivariate biochemical indicator data; and multi-scale feature extraction is performed on the synchronously aligned multivariate biochemical indicator data using a logarithmic wavelet energy spectrum enhancement algorithm.

[0010] Preferably, the S1 specifically includes:

[0011] In the implementation of the logarithmic wavelet spectrum enhancement algorithm, based on the synchronously aligned multivariate biochemical index data, the scale position weighting factor is introduced and combined with the logarithmic transformation to obtain multi-scale feature data.

[0012] Preferably, the S2 specifically includes:

[0013] After the multi-scale feature data is subjected to feature dimensionality reduction processing, it is normalized to obtain the multi-scale feature data after dimensionality reduction; the multi-scale feature data after dimensionality reduction is fused to obtain the fused feature data.

[0014] Preferably, the S2 specifically includes:

[0015] In the implementation process of the multi-stage feedback control anomaly detection algorithm, the overall weighted anomaly response intensity is calculated based on the fused feature data and combined with the feature data of the reference sample; the overall weighted anomaly response intensity is logarithmically compressed to obtain a preliminary anomaly score.

[0016] Preferably, the S2 specifically includes:

[0017] Based on the fused feature data and the feature data of the reference sample, the deviation degree of the current urine sample in the feature dimension is quantified, and combined with the Mahalanobis kernel suppression function, the overall weighted abnormal response intensity is obtained.

[0018] Preferably, the S2 specifically includes:

[0019] In the implementation process of the multi-stage feedback-controlled anomaly detection algorithm, a feedback mechanism based on reinforcement learning is introduced based on the fused feature data to adjust the preliminary anomaly score and obtain a comprehensive anomaly score.

[0020] Preferably, the S2 specifically includes:

[0021] Based on the comprehensive abnormality score, an abnormality discrimination threshold bias term is introduced to calculate the abnormality discrimination probability; the abnormality discrimination probability is compared with the preset threshold to generate the evaluation result of the urine sample: when the abnormality discrimination probability is greater than or equal to the threshold, it means that the urine sample represented by the current fused feature data is abnormal; when the abnormality discrimination probability is less than the threshold, it means that the urine sample represented by the current fused feature data is normal.

[0022] The beneficial effects of the technical solution of the present invention are:

[0023] 1. After obtaining the urine sample, the urine sample is introduced into the multiplex detection channel under the control of a micropump, and multiplex detection devices such as electrochemical arrays, colorimetric sensors, and redox electrodes are used to collect and analyze multiple target biochemical indicators (such as creatinine, white blood cells, red blood cells, nitrite, protein, etc.) in parallel, realizing real-time acquisition of multiple indicators under a single detection operation, effectively avoiding the problems of multiple batch detections required in traditional detection methods, which are time-consuming and have large errors.

[0024] 2. By applying the logarithmic wavelet energy spectrum enhancement algorithm to the synchronized and aligned multivariate biochemical indicator data, a scale-time two-dimensional feature space is constructed. The logarithmic wavelet energy spectrum enhancement algorithm combines the local sensitivity of the wavelet basis with the nonlinear enhancement characteristics of the logarithmic function, significantly improving the detection ability of weak, local fluctuations (such as mild leukocyte infiltration and low-concentration protein abnormalities), and can effectively capture micro-changes in urine samples in sub-healthy and early pathological conditions.

[0025] 3. A multi-stage feedback-controlled anomaly detection algorithm is introduced, which gradually completes preliminary anomaly scoring, feedback correction and comprehensive judgment in three stages. It has the ability to enhance outliers, suppress normal values, and perform multiple rounds of convergence, which can greatly improve the sensitivity and stability of anomaly recognition. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] Figure 1 This is a flow chart of the multi-item combined detection method for the urinary system described in the present invention. DETAILED DESCRIPTION

[0027] In order to further illustrate the technical means and effects adopted by the present invention to achieve the predetermined purpose of the invention, 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 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 making creative efforts shall fall within the scope of protection of the present invention.

[0028] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.

[0029] The specific scheme of the multi-item combined detection method of the urinary system provided by the present invention is described in detail below with reference to the accompanying drawings.

[0030] Refer to the attached Figure 1, which shows a flow chart of a multi-item combined detection method for the urinary system provided by one embodiment of the present invention, the method comprising the following steps:

[0031] S1. Collecting and analyzing substances in a urine sample in a multivariate detection channel to obtain multivariate biochemical indicators; digitally converting the multivariate biochemical indicators to obtain converted multivariate biochemical indicator data; performing synchronous alignment processing on the converted multivariate biochemical indicator data, and then performing multiscale feature extraction to obtain multiscale feature data;

[0032] After obtaining a urine sample through a urine collection device, the urine sample is injected into a multi-element detection channel under the control of a micropump. In the multi-element detection channel, different substances in the urine sample are collected and analyzed in parallel by multi-element detection devices such as an electrochemical array, a colorimetric sensor, and a redox electrode to obtain multi-element biochemical indicators. The collection and analysis methods are technical means well known to those skilled in the art and are not described in detail here. The multi-element biochemical indicators such as creatinine, white blood cells, red blood cells, nitrite, protein, etc. are further digitally converted by an existing analog-to-digital converter to obtain converted multi-element biochemical indicator data;

[0033] In order to solve the problem that the data of different channels cannot be synchronized due to the difference in acquisition time, which affects the analysis and fusion of subsequent data, the converted multivariate biochemical indicator data in each detection channel (the converted multivariate biochemical indicator data of the th detection channel is represented by ) is synchronized and aligned using the existing dynamic time warping and mutual information methods to obtain the synchronized multivariate biochemical indicator data;

[0034] Furthermore, the logarithmic wavelet spectrum enhancement algorithm is used to perform multi-scale feature extraction on the synchronized and aligned multivariate biochemical index data to obtain multi-scale feature data. The logarithmic wavelet spectrum enhancement algorithm is based on wavelet transform technology and combined with logarithmic transform to construct a feature extraction formula:

[0035] ;

[0036] in, It is Detection channels at scale and time The characteristic response value on represents the feature after wavelet transformation and logarithmic enhancement; It is The first detection channel Multiple biochemical index data after synchronous alignment; Indicates the The total number of multiplex biochemical index data after synchronous alignment of detection channels; It is the Morlet wavelet function, which is widely used in time-frequency analysis to extract features at specific scales and time positions. It is a technical means well known to those skilled in the art and will not be described in detail here. is the current time; is the scale parameter, which is used to control the width of the wavelet function; is a scale position weighting factor, which is used to suppress the weight of data at distant locations to enhance the local focus of the multivariate biochemical index data after synchronous alignment; the introduction of the logarithmic function can enhance weak feature responses while preventing high response values ​​from dominating the results; through the above process, we can construct multi-scale feature data. , where represents multi-scale feature data, Indicates the The multi-scale feature data of each detection channel can more accurately capture the subtle changes of each biochemical indicator in urine.

[0037] S2. Fuse the multi-scale feature data to obtain fused feature data; introduce a multi-stage feedback control anomaly detection algorithm to perform anomaly detection on the fused feature data, and perform comprehensive discrimination evaluation to generate an evaluation result of the urine sample.

[0038] After the multi-scale feature data is subjected to feature dimensionality reduction processing using a data dimensionality reduction processing technique such as principal component analysis, normalization processing is performed to eliminate the dimensionality effect to obtain multi-scale feature data after dimensionality reduction; the multi-scale feature data after dimensionality reduction is fused using Tucker decomposition to obtain fused feature data; the data dimensionality reduction processing technique, normalization processing method and Tucker decomposition are technical means well known to those skilled in the art and will not be described in detail here;

[0039] Anomaly detection is performed on the fused feature data using a multi-stage feedback control anomaly detection algorithm. The specific implementation process of the multi-stage feedback control anomaly detection algorithm is as follows:

[0040] In the first stage, we integrate the local anomaly detection concept, kernel function technology, dimension weighting mechanism, and compressed response method to measure anomaly and obtain a preliminary anomaly score. Specifically, based on the fused feature data, we introduce reference samples and combine them with logarithmic correction to obtain a preliminary anomaly score. The specific formula is as follows:

[0041] ;

[0042] in, It is The fused feature data Initial abnormality score; It is a regulating factor used to control the dynamic range of the preliminary abnormality score and reduce the noise amplification caused by abnormal peaks. It is determined according to expert experience and the reference value is ; The number of reference samples is set according to specific needs and is not limited here; is the total number of dimensions of the fused feature data; It is The characteristic data of a reference sample is obtained from a known sample reference database, wherein the dimension of the characteristic data of the reference sample is consistent; It is The first reference sample characteristic data eigenvalues; It is The fused feature data No. eigenvalues; represents transpose; Represents the dimension weighted intensity term, which is used to quantify the degree of deviation of the current urine sample (i.e., the sample to be tested) in the feature dimension; is the Mahalanobis kernel suppression function, which means that the farther away from the normal sample (i.e., reference sample) distribution center, the smaller the kernel function value is, and the greater the contribution to abnormality judgment is. The Mahalanobis kernel suppression function is a technical means well known to those skilled in the art and will not be described in detail here; The abnormal response intensity based on the weighting of reference samples and kernel function suppression, namely the overall weighted abnormal response intensity, is described. At the same time, the overall weighted abnormal response intensity is logarithmically compressed to prevent excessive response due to abnormal amplification of individual biochemical indicator data, while retaining nonlinear trends so that even weak abnormal samples can be discriminated.

[0043] In the second stage, the preliminary anomaly score is adjusted using existing reinforcement learning algorithms. These algorithms, such as Actor-Critic or Proximal Policy Optimization (PPO), consider the context and relevance of each fused feature data and adjust the preliminary anomaly score through a reinforcement learning-based feedback mechanism to obtain a comprehensive anomaly score.

[0044] In the third stage, a comprehensive discriminant formula is constructed based on the nonlinear mapping function of the Sigmoid structure. Combined with the comprehensive abnormality score, the abnormality discrimination probability is calculated to determine the abnormality of the urine sample. The specific expression of the comprehensive discriminant formula is as follows:

[0045] ;

[0046] in, is the abnormal discrimination probability, which represents the current fused feature data The confidence level that the urine sample represented is abnormal; is the sensitivity factor of nonlinear compression, which is learned using an automatic parameter adjustment strategy (such as Bayesian optimization). The reference value range is The automatic parameter adjustment strategy is a technical means well known to those skilled in the art and will not be described in detail here; is the comprehensive abnormality score of the current fused feature data; is the abnormality discrimination threshold bias item, which is determined according to expert experience and has a reference value range of ; According to the abnormality discrimination probability, it is possible to draw a conclusion on whether the urine sample is abnormal; specifically: the abnormality discrimination probability is combined with the threshold value preset according to the expert experience method Comparison is made to generate an assessment of the urine sample: When , it is considered that the urine sample represented by the current fused feature data is abnormal; when When , it is considered that the current fused feature data The urine sample represented is normal. The value of is determined according to the specific application scenario, such as sensitive judgment (high-risk screening): , suitable for early screening of kidney disease, early stage of urinary tract infection, subclinical state, etc., even if the abnormality is not high, early warning is required; conservative judgment (reducing false positives): set It is suitable for periodic testing of people with a confirmed medical history, with the goal of accurately identifying true pathological abnormalities and reducing the false alarm rate.

[0047] In summary, a multi-item combined detection method for the urinary system was completed.

[0048] The order in which the embodiments of the invention are presented is for illustrative purposes only and does not necessarily represent the superiority or inferiority of the embodiments. The processes depicted in the accompanying drawings do not necessarily require the specific order or sequential order shown to achieve the desired results. In certain embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0049] The various embodiments in this specification are described in a progressive manner, and the same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on the differences from other embodiments.

[0050] 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 above embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the above embodiments, or make equivalent replacements for some of the technical features therein. These modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be included in the scope of protection of the present invention.

Claims

1. A multi-item combined detection method for the urinary system, characterized in that: The following steps are involved: S1. Collect and analyze substances in the urine sample in the multiplex detection channel to obtain multiplex biochemical indicators; Digitally converting the multivariate biochemical indicators to obtain converted multivariate biochemical indicator data; performing synchronous alignment processing on the converted multivariate biochemical indicator data, and then performing multi-scale feature extraction to obtain multi-scale feature data; S2. Fuse the multi-scale feature data to obtain fused feature data; introduce a multi-stage feedback control anomaly detection algorithm to perform anomaly detection on the fused feature data; the specific implementation process of the multi-stage feedback control anomaly detection algorithm is as follows: based on the fused feature data and the feature data of the reference sample, quantify the degree of deviation of the current urine sample in the feature dimension, and combine it with the Mahalanobis kernel suppression function to obtain the overall weighted anomaly response intensity; combine the adjustment factor to perform logarithmic compression transformation on the overall weighted anomaly response intensity to obtain a preliminary anomaly score; based on the fused feature data, introduce a reinforcement learning-based feedback mechanism to adjust the preliminary anomaly score to obtain a comprehensive anomaly score; Based on the comprehensive anomaly score, an anomaly discrimination threshold bias term is introduced to calculate the anomaly discrimination probability; The abnormality discrimination probability is compared with the preset threshold to generate the evaluation result of the urine sample: when the abnormality discrimination probability is greater than or equal to the threshold, it means that the urine sample represented by the current fused feature data is abnormal; When the abnormality discrimination probability is less than the threshold, it means that the urine sample represented by the current fused feature data is normal.

2. A multi-item joint detection method for the urinary system according to claim 1, characterized in that: Said S1 specifically includes: The converted multivariate biochemical indicator data in each detection channel are synchronously aligned to obtain the synchronously aligned multivariate biochemical indicator data; and multi-scale feature extraction is performed on the synchronously aligned multivariate biochemical indicator data using a logarithmic wavelet energy spectrum enhancement algorithm.

3. A multi-item joint detection method for the urinary system according to claim 2, characterized in that: Said S1 specifically includes: In the implementation of the logarithmic wavelet spectrum enhancement algorithm, based on the synchronously aligned multivariate biochemical index data, the scale position weighting factor is introduced and combined with the logarithmic transformation to obtain multi-scale feature data.

4. A multi-item combined detection method for the urinary system according to claim 1, characterized in that: Said S2 specifically includes: After the multi-scale feature data is subjected to feature dimensionality reduction processing, it is normalized to obtain the multi-scale feature data after dimensionality reduction; the multi-scale feature data after dimensionality reduction is fused to obtain the fused feature data.

Images