Multi-channel one-core multi-detection integrated detection method

Abstract

The invention provides a multi-channel one-core multi-detection integrated detection method, which is applied to the technical field of in-vitro diagnosis and comprises the following steps: acquiring time sequence original signals of blood routine, liver function and kidney function channels; obtaining a baseline stable primary signal through cross-channel baseline trend coupling and adaptive correction; adopting a machine learning model to identify interference components and performing component offset to obtain a decoupling signal; mapping the decoupling signal into a physiological parameter value and carrying out rationality test; when the low credible parameter exists, a rechecking mechanism is triggered; performing weighted fusion on the basis of the rationality test result and the recheck signal to obtain a comprehensive parameter set; and finally generating a comprehensive detection report containing the credibility score. Cooperative processing and fusion analysis of multi-channel detection signals are realized, and the accuracy and reliability of a detection result are remarkably improved through interference identification and a dynamic recheck mechanism.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of in vitro diagnosis, and in particular to a multi-channel one-core multi-detection integrated detection method. BACKGROUND

[0002] Blood detection is a core means in clinical medical diagnosis and health management. In the prior art, different detection methods are often used for different detection indexes. For example, resistance impedance detection technology is used to realize counting and classification of blood cell components, electrochemical detection technology is used to analyze enzyme components in blood, and optical detection technology is used to quantitatively detect metabolites. The above detection means have significant differences in detection principles, detection conditions and detection environments. Usually, samples need to be taken separately and multiple independent devices are used to complete the detection, which not only increases the detection steps and time, but also leads to low sample utilization and detection result deviation due to inconsistent sample conditions.

[0003] With the development of microfluidic chips and integrated sensing technology, multiple detection modules are integrated on the same chip or the same device. However, even if the collection is performed on the same device, the signal characteristics of different detection channels are different. In addition, in other complex scenarios such as network reliability evaluation, there are similar problems: multiple types of index data from the device running layer, the link transmission layer and the application service layer need to be processed at the same time. The differences in time dimension and statistical characteristics of various data make it difficult to reveal the correlation between different source indexes by conventional processing, resulting in a lack of accurate quantitative basis for evaluating the overall running state of the system.

[0004] Therefore, a multi-type detection signal processing method capable of being implemented on the same platform is needed to enable collaborative analysis and result output of multiple types of signals under a unified framework, and to provide a reliable data basis for blood sample multi-index detection and other multi-source data joint analysis scenarios. SUMMARY

[0005] Embodiments of the present application provide a multi-channel one-core multi-detection integrated detection method, which realizes a multi-type detection signal processing method, enables collaborative analysis and result output of multiple types of signals under a unified framework, and achieves the above-mentioned purposes by adopting the following technical solutions: A multi-channel one-core multi-detection integrated detection method is applied to a detection device, and a microfluidic chip is integrated in a detection channel of the detection device; Timing original signals are obtained, which come from blood routine, liver function and kidney function channels for parallel detection of the same blood sample on the microfluidic chip; Based on the timing original signals, a stable time-varying baseline of each detection channel is calculated through cross-channel baseline trend coupling and adaptive correction, the timing original signals are corrected to obtain baseline stable primary signals; The decoupled signal is obtained by canceling the components of the baseline-stabilized primary signal. The decoupling signal is mapped to physiological parameter values, and the physiological parameter values ​​are checked for reasonableness. Reasonableness check results are generated based on the reasonableness check. When the reasonableness check results show that there are low-confidence parameters, a review mechanism is triggered to generate a review signal. A comprehensive parameter set is obtained by weighted fusion of the reasonableness test results and the verification signal; A comprehensive test report is generated based on a comprehensive parameter set.

[0006] The original time-series signal is segmented into sliding time windows. Statistical feature parameters are extracted within each time window, and local trend curves are fitted to obtain time-resolved baseline candidate curves. A cross-channel baseline trend coupling model is established among the detection channels. By comparing the correlation and offset between the baseline candidate curves of different detection channels, a cross-channel baseline trend is generated. Based on the cross-channel baseline trend, the baseline candidate curves of each detection channel are adaptively corrected to remove noise and drift components, thereby obtaining a stable time-varying baseline.

[0007] The time-series raw signal of each detection channel is subtracted point by point from the stable time-varying baseline to obtain a preliminary signal difference sequence; Smoothing and outlier removal algorithms are applied to signal difference sequences to suppress spikes caused by transient interference. Baseline stable primary signals are output on blood routine, liver function and kidney function channels.

[0008] Frequency domain transformation, time domain statistics, and phase analysis are performed on the baseline-stabilized primary signal to extract frequency domain features, time domain features, and phase coupling features, thereby obtaining the interference feature vector. The interference feature vector is input into a preset machine learning model to identify the fluid pulsation interference component, temperature gradient interference component, and electromagnetic coupling interference component. Based on each interference component, a component decoupling matrix is ​​constructed, and the interference components in the baseline stable primary signal are canceled one by one to obtain the decoupled signal.

[0009] For each detection channel, a nonlinear response mapping relationship between physical quantities and physiological parameters is established, and the decoupled signal is converted into candidate physiological parameter values; Candidate physiological parameter values ​​from each detection channel are compared across channels to eliminate outliers that deviate too much from physiological norms. The physiological parameter values, after mapping and comparison correction, are output in each detection channel.

[0010] Based on the gender, age group, characteristics, and medical history of the test subjects, obtain the corresponding reference range of physiological parameters; Physiological parameter values ​​are compared with reference ranges, and a reasonableness index is calculated based on the statistical deviation. Reasonableness test results are generated based on reasonableness indicators.

[0011] When the rationality test results indicate the presence of low-confidence parameters, the baseline stable primary signal and decoupling signal are re-called and repeatedly calculated under different time windows and parameter thresholds to obtain the verified physiological parameter values. Multiple consistency comparisons were performed on the verified physiological parameter values ​​to generate a verification signal.

[0012] The initial weights of the decoupling signals are assigned based on the results of the rationality test, and the initial weights of the verification signals are assigned based on the consistency level of the verification signals. The time-varying weighted coefficients are obtained by adjusting the initial weights using a nonlinear weighting function based on a time decay factor or a consistency index. The decoupled signal and the complex signal are weighted and fused using time-varying weighting coefficients to obtain a comprehensive parameter set.

[0013] Blood routine parameters, liver function parameters, and kidney function parameters were extracted from the comprehensive parameter set and classified into test results. A credibility calculation process is introduced into the detection results of each category, and a credibility score is generated based on the internal consistency of the comprehensive parameter set, cross-channel comparison results and historical review results. The classification test results are integrated with the credibility score to obtain a comprehensive test report that includes both the test results and the credibility score.

[0014] The machine learning model is one of the following: support vector machine model, random forest model, or convolutional neural network model.

[0015] As can be seen from the above technical solution, this application has the following beneficial effects: 1. This invention introduces collaborative technical steps such as microfluidic chip shunting, time-varying baseline correction, interference identification and component cancellation, nonlinear response mapping, and cross-channel comparison correction. It realizes end-to-end processing of electrical impedance signals, electrochemical signals, and optical signals from acquisition to parameter output. It effectively deducts and corrects system noise, environmental disturbances, and cross-interference between different detection channels that occur during blood sample testing, ensuring the stability of the three types of detection signals. By realizing the synchronous acquisition and fusion processing of multi-channel detection data on the same chip, and combining a hierarchical population knowledge base and an automated verification mechanism, the final comprehensive parameter set can accurately reflect the biochemical and physical characteristics of blood samples.

[0016] 2. This invention establishes a multi-dimensional signal processing and cross-channel data fusion framework, organically linking technologies such as joint feature extraction in the time, frequency, and phase domains, interference component identification and decoupling, cross-channel comparison and correction, rationality verification of stratified populations, and dynamic weighted fusion. It enables comprehensive processing of multi-source signals from different detection channels, such as routine blood tests, liver function tests, and kidney function tests, within the same system. It can dynamically adjust signal weights based on the time-varying nature of detection conditions and individual differences, improving the consistency of output parameters. By integrating consistency verification and historical review information, this invention generates a comprehensive test report containing a reliability score, providing medical institutions with a unified data output interface for convenient subsequent analysis and record management. Attached Figure Description

[0017] The invention will now be further described with reference to the accompanying drawings.

[0018] Figure 1 A first flowchart of a multi-channel, single-chip, multi-detection integrated detection method provided in this application embodiment; Figure 2 A second flowchart of a multi-channel, single-chip, multi-detection integrated detection method provided in this application embodiment; Figure 3 A third flowchart of a multi-channel, single-chip, multi-detection integrated detection method provided in this application embodiment; Figure 4 The fourth flowchart of a multi-channel, single-chip, multi-detection integrated detection method provided in this application embodiment is shown. Detailed Implementation

[0019] The terms "first," "second," and "third," etc., used in this application specification, claims, and drawings are used to distinguish different objects, not to limit a specific order.

[0020] In the embodiments of this application, the terms "exemplary" or "for example" are used to indicate that something is an example, illustration, or description. Any embodiment or design that is described as "exemplary" or "for example" in the embodiments of this application should not be construed as being more preferred or advantageous than other embodiments or design. Specifically, the use of the terms "exemplary" or "for example" is intended to present the relevant concepts in a specific manner.

[0021] Research has found that most existing chips only support a single detection mode, making it difficult to achieve a unified time reference and data fusion between the detection signals. This results in insufficient correlation of the detection results and makes it difficult to achieve a multi-dimensional joint assessment of the health status of blood samples. Due to the limitations of the detection channel's structural design and signal processing methods, it is difficult to ensure both high throughput and convenience of detection while maintaining detection sensitivity and specificity.

[0022] To address the aforementioned issues, this application provides a multi-channel, single-chip, multi-detection integrated detection method: Example 1

[0023] To solve the above problems, such as Figure 1 As shown, this embodiment provides an overall implementation of a multi-channel, single-chip, multi-detection integrated detection method. The detection device includes a detection channel component, a signal acquisition and processing module, a human-machine interface, and a report generation unit. The detection channel component integrates a microfluidic chip for blood sample distribution and signal acquisition.

[0024] The blood sample to be tested is introduced into the testing channel through the sampling mechanism. The microfluidic chip is equipped with a microfluidic distribution channel, which diverts the same blood sample to the routine blood test channel, the liver function test channel, and the kidney function test channel, enabling the above three types of channels to perform parallel testing on the same chip.

[0025] The signal acquisition and processing module receives the raw timing signals from each detection channel. To eliminate system noise and environmental disturbances, the system first performs time-varying baseline calculation on the raw timing signals of each detection channel. This process uses a sliding time window method to extract statistical parameters segment by segment and fit local trend curves. Then, through a cross-channel baseline trend adaptive correction model, a stable time-varying baseline that matches the detection characteristics of each channel is formed.

[0026] Subsequently, the system performs a differential operation between the original signal and a stable time-varying baseline, and combines this with smoothing and outlier removal steps to obtain a baseline-stabilized primary signal. This signal retains the main detection information and suppresses environmental drift and transient interference.

[0027] Decoupling signal extraction is achieved through interference identification and component cancellation steps. The system extracts features in the frequency, time, and phase domains, which are then input into a nonlinear transformation model to identify three main types of interference components: fluid pulsation interference, temperature gradient interference, and electromagnetic coupling interference. After processing with the component decoupling matrix, these interference components are canceled, thus obtaining the decoupling signal.

[0028] For decoupled signals, the system converts them into candidate physiological parameter values ​​according to the nonlinear response mapping relationship between physical quantities and physiological parameters preset in each detection channel, and performs cross-channel comparison to eliminate outliers and obtain corrected physiological parameter values.

[0029] Based on this, the system uses the stratified population information of the test subjects to call the corresponding stratified population physiological knowledge base to check the rationality of the physiological parameter values. When the test results show that there are low-confidence parameters, the system automatically starts the review mechanism, repeatedly calculating the baseline stable primary signal and decoupled signal under different time windows and parameter thresholds to obtain the reviewed physiological parameter values.

[0030] The verification signals and preliminary rationality test results are weighted and fused using nonlinear dynamic weights to obtain a comprehensive parameter set. Finally, the report generation unit extracts the classification detection results from the comprehensive parameter set, generates a credibility score based on internal consistency and historical verification, and outputs a comprehensive detection report containing the detection results and credibility score. The report also provides query, printing, and data export functions through a human-computer interaction interface. Example 2

[0031] like Figure 2 As shown, specifically: Based on the baseline stable primary signal obtained in Example 1, this example details how to achieve decoupling signal extraction and interference suppression in the multi-channel single-core multi-detector integrated detection method through component cancellation.

[0032] First, multidimensional feature extraction is performed on the obtained baseline-stabilized primary signal. For each detection channel, the acquired signal not only contains amplitudes that vary over time, but is also affected by various interference sources in the operating environment of the detection equipment. Therefore, it is necessary to describe the signal characteristics from multiple dimensions. This embodiment adopts a joint analysis method in the time domain, frequency domain, and phase domain.

[0033] In the time domain, features such as mean, standard deviation, peak factor, impulse factor, skewness, and kurtosis are calculated to reflect the overall level and fluctuation characteristics of the signal. In the frequency domain, indicators such as power spectral density distribution, main frequency components, and frequency band energy ratio are obtained through fast Fourier transform to identify systematic interference in specific frequency bands. In the phase domain, indices such as inter-channel coherence and phase delay distribution are calculated to reveal cross-channel coupling and potential electromagnetic synchronization interference.

[0034] The aforementioned multidimensional indicators are extracted in each sampling time period and then form a multidimensional interference feature vector.

[0035] The interference feature vector is input into a pre-constructed interference identification model. This model, based on a multi-layer nonlinear transform structure, is trained offline during equipment installation, commissioning, and calibration using standard reference samples and specific interference sources, such as fluid velocity pulsation sources, temperature control devices, and electromagnetic radiation generators. It can identify three main types of interference components mixed into the signal during actual detection: fluid pulsation interference components, temperature gradient interference components, and electromagnetic coupling interference components. The model calculates the contribution coefficient of each type of interference and outputs the amplitude and phase characteristics of each interference component.

[0036] A component decoupling matrix is ​​established based on the identified interference components. For each detection channel, the projection of the interference component onto the channel signal is defined within the decoupling matrix. Matrix operations are then used to subtract each interference component from the baseline-stabilized primary signal, thereby extracting the true response components. The signal after this processing is the decoupled signal, which contains information primarily related to the biochemical characteristics and physical response of the blood sample itself, while minimizing the influence of external non-target interference.

[0037] By employing multi-dimensional feature extraction and interference identification based on a trained model, this embodiment can simultaneously identify and counteract complex coupled interference from the environment and equipment, making it suitable for situations with significant cross-interference in multi-channel integrated detection systems. Compared to traditional filtering or simple smoothing algorithms, it can reduce the impact of noise and drift on measurement accuracy while preserving effective signal components, laying a reliable foundation for subsequent physiological parameter calculations. Example 3

[0038] like Figure 3 Specifically, in this embodiment, based on the decoupled signal obtained after interference suppression, a nonlinear response mapping model is established for different functional channels. To this end, during the device development and calibration phase, standard reference samples with known concentrations or specific values ​​are used to collect decoupled signals for the blood routine channel, electrochemical channel, and optical channel, and the corresponding standard physiological parameter values, such as hemoglobin concentration, alanine aminotransferase activity, and creatinine content, are recorded. Through fitting analysis, nonlinear response curves or mathematical expressions between the decoupled signals of each channel and the target physiological parameters are obtained to reflect the quantitative relationship between the detection signal and the physiological parameters.

[0039] In actual testing, the decoupled signal is input into the mapping model described above to obtain candidate physiological parameter values ​​for each channel. Due to differences in clinical samples and occasional changes in testing conditions, some candidate values ​​may deviate significantly from the conventional physiological range, thus requiring further correction.

[0040] After obtaining candidate values, this embodiment introduces a cross-channel comparison and correction mechanism. Utilizing known physiological correlations between parameters from different channels—for example, abnormal renal function is often accompanied by abnormalities in specific blood routine parameters—the mechanism performs joint analysis of interrelated parameters and calculates the degree of deviation. Outliers that clearly do not conform to physiological correlation patterns or deviate excessively from norms are either eliminated or corrected based on regression analysis of relevant parameters.

[0041] After the above mapping and cross-channel correction processing, the final physiological parameter values ​​of each channel can be obtained, providing a data basis for subsequent rationality testing and comprehensive analysis.

[0042] This embodiment establishes a nonlinear response relationship between decoupled signals and physiological parameters, realizing the scientific conversion of detection signals into medically usable indicators. Furthermore, cross-channel comparison correction reduces measurement errors caused by single-channel anomalies, improving the accuracy and robustness of the results. Example 4

[0043] like Figure 4 Specifically, after obtaining the physiological parameter values, in order to improve the relevance and accuracy of the test, this embodiment pre-constructs a stratified population physiological knowledge base in the system. This knowledge base stratifies the population according to gender, age group, and common medical history, and records the normal range and typical variation patterns of the main physiological indicators of the corresponding stratified population.

[0044] During the actual testing, the system automatically retrieves the matching sub-knowledge base from the main knowledge base based on the examinee's gender, age, and past medical history tags.

[0045] The physiological parameter values ​​obtained from the tests are compared with reference intervals in the sub-knowledge base, and a rationality index is calculated based on indicators such as statistical deviation and rate of change, combined with time series trends. The rationality index reflects the degree of deviation between the tested parameters and the normal physiological range.

[0046] When the rationality check results show that certain parameters have low credibility or deviate abnormally, the system automatically triggers a verification mechanism. This mechanism calls upon previously stored baseline stable primary signals and decoupling signals, and recalculates them under different time window divisions and threshold adjustments to obtain the verified physiological parameter values. Subsequently, multiple consistency comparisons are performed on the verified parameter values ​​to confirm whether errors are caused by noise or transient anomalies. If multiple calculations still yield inconsistent results, the parameters are marked as abnormal.

[0047] The review mechanism ultimately generates a review signal, which is used to provide auxiliary judgment criteria in subsequent comprehensive analysis.

[0048] By introducing a knowledge base that matches the characteristics of the test subjects and an automated review process, this embodiment can effectively identify and correct unreliable test results caused by individual differences, environmental interference or accidental noise, thereby improving the credibility and stability of the test conclusions. Example 5

[0049] like Figure 4 Specifically, based on the rationality test results, initial weights are assigned to the decoupling signals. Parameters that pass the rationality test are given higher weights, while parameters with questionable aspects are given lower weights. Simultaneously, the verification data is assigned corresponding weights based on the consistency level of the verification signals. To adapt to the characteristics of signal state changes over time during the detection process, a nonlinear weighting function is designed to dynamically adjust the above weights at different time scales, resulting in time-varying weighted coefficients. To adapt to the characteristics of signal state changing over time during the detection process, the system is designed with a nonlinear weighting function to dynamically adjust the weights at different time scales and obtain time-varying weighted coefficients. The basic definition of a nonlinear weighting function: Nonlinear weighting function Used to describe at each time point The dynamic characteristics of signal weights changing over time are described below. Their mathematical form can be expressed as:

[0050] in These are control parameters used to adjust the changing characteristics of the weighting function; The time center point is the reference time point where the signal change occurs; is the base of the natural logarithm.

[0051] The choice of this function form is based on the sigmoid function, which has the advantage of smoothly transitioning weight values, avoiding abrupt changes, and effectively simulating the smooth transition of signal changes over time. Specifically, the weight function in time... The nearest changes are the fastest, while those further away... The time tends to be stable, making it suitable for signal fusion requirements that are sensitive to time changes; To address the requirements of varying signal characteristics, time-varying control parameters were designed for the system. The parameters are adjusted according to changes in signal characteristics during the detection process; for example, during phases where the signal changes rapidly. Increase the value to enhance time sensitivity; while decrease it during periods of slow or stable signal change. This makes weight changes smoother and ensures signal fusion in a stable state; At each time point Next, by calculating the weight of each signal. The decoupled signal and the verification signal are combined to generate the corresponding weighted fusion result. The specific weighted fusion formula is as follows:

[0052] Indicates a fused signal; Indicates the decoupling signal; Indicates a verification signal; It is a time-varying weighting factor; and These are signals obtained based on preliminary signal processing and verification processing, respectively. The dynamic weighting coefficients that change over time determine the degree of importance attached to the decoupled signal and the verification signal during signal fusion. This nonlinear weighting function allows for adjustments to the signal fusion strategy based on the signal's temporal evolution characteristics. For example, during periods of signal instability or uncertainty, the system assigns higher weights to the verification signal to ensure that abnormal signals are effectively identified and corrected. Conversely, during periods of signal stability, the decoupling signal is used preferentially to improve detection accuracy. A time-varying weighting coefficient is applied to the decoupled signal and the verified signal for weighted fusion, resulting in a comprehensive parameter set. This parameter set includes multi-dimensional indicators such as blood routine related parameters, liver function related parameters, and kidney function related parameters.

[0053] Based on the comprehensive parameter set, various indicators are classified and integrated to form classification detection results. A credibility calculation module is introduced to calculate the credibility score by analyzing the consistency within the comprehensive parameter set, cross-channel comparison results, and historical review records.

[0054] The results of the classification tests are integrated with the credibility scores to form a comprehensive test report. The report content can be displayed in real time through a human-computer interaction interface, and it supports printing and data export to facilitate medical personnel to record and further analyze the data.

[0055] By weighted fusion of the rationality test results and the verification signals, this embodiment can comprehensively utilize the advantages of preliminary detection and verification data, dynamically adjust the weights to improve the credibility of the comprehensive results, and generate a test report containing a credibility score, providing more reliable and comprehensive data support for clinical diagnosis.

[0056] Example 6 like Figures 1 to 4 As shown in the figure, this embodiment provides an overall implementation process for a multi-channel, single-chip, multi-detection integrated detection method, which is used to realize parallel detection of multiple indicators of blood samples, signal decoupling processing, result verification, and comprehensive report output in an integrated detection device.

[0057] Blood samples are introduced into the detection device via a sampling mechanism, and then evenly distributed to various functional detection channels within the distribution channels of the microfluidic chip. Each detection channel detects the sample based on electrical impedance, electrochemical, and optical methods, respectively, and outputs the corresponding time-series raw signals in real time.

[0058] The signal acquisition and processing module first performs baseline processing on the raw time-series signals acquired from each detection channel. Using a sliding time window approach, statistical feature parameters are extracted within each window, and a local trend curve is fitted to obtain a time-resolved baseline candidate curve.

[0059] A cross-channel baseline trend coupling model is established across all channels. By comparing the correlation and offset of baseline candidate curves in different channels, a cross-channel baseline trend is formed. Based on this trend, the baseline candidate curve of each channel is adaptively corrected to eliminate drift and noise effects, resulting in a stable time-varying baseline that matches the characteristics of the detection channel.

[0060] The raw signals from each detection channel are differentially analyzed point-by-point with a stable time-varying baseline to obtain a preliminary signal difference sequence. This sequence is then combined with smoothing and outlier removal algorithms to eliminate spike interference and transient anomalies, resulting in a stable primary baseline signal. This signal effectively preserves target response information and suppresses low-frequency drift and occasional noise caused by environmental changes.

[0061] Multidimensional feature extraction is performed on the baseline-stable primary signal, including: Time-domain characteristics include mean, standard deviation, kurtosis factor, impulse factor, skewness, and kurtosis; Frequency domain characteristics, including power spectral density, main frequency components, and frequency band energy distribution; Phase domain characteristics, including cross-channel coherence and phase delay.

[0062] The aforementioned multidimensional features form an interference feature vector, which is input into an interference identification model based on multi-layer nonlinear transformation. This model is trained using standard samples and specific interference sources during equipment calibration. It can identify fluid pulsation interference components, temperature gradient interference components, and electromagnetic coupling interference components, and output the amplitude and phase information of each type of interference.

[0063] A component decoupling matrix is ​​established based on the identified interference components. Matrix operations are then performed on the baseline stable primary signal to cancel out each interference component one by one, obtaining the decoupled signal. The decoupled signal can accurately reflect the biochemical and physical response characteristics of the blood sample itself.

[0064] For the blood routine test channel, liver function test channel, and kidney function test channel, a nonlinear response mapping model between decoupled signals and corresponding physiological parameters was established using standard reference samples during the equipment development and calibration phase.

[0065] In actual testing, the decoupled signal is input into the corresponding mapping model to obtain candidate physiological parameter values.

[0066] To improve the clinical validity of the results, a cross-channel comparison and correction mechanism was introduced. By analyzing the correlation between physiological parameters, such as the fact that abnormal renal function is often accompanied by changes in specific blood routine parameters, the deviation of candidate values ​​was calculated, and parameters with obvious abnormalities were removed or corrected by regression to obtain the final physiological parameter values ​​for each channel.

[0067] The system has a built-in hierarchical population physiological knowledge base, which is divided into categories by gender, age group, and medical history, and records the normal range and variation patterns of the main physiological indicators of various population groups.

[0068] During the test, the corresponding sub-knowledge base is called according to the characteristics of the examinee. The obtained physiological parameter values ​​are compared with the reference range and change trend in the sub-knowledge base. The rationality index is calculated based on the statistical deviation and the rate of change.

[0069] If the rationality test results show that certain parameters have low credibility or deviate abnormally, the system will automatically trigger the review mechanism, re-call the baseline stable primary signal and decoupling signal, repeat the calculation under the condition of changing the time window and parameter threshold, and perform a consistency comparison on the parameter values ​​obtained from the review, and finally generate a review signal.

[0070] Initial weights are assigned to the decoupling signals based on the rationality test results, and weights are assigned to the verification data based on the consistency level of the verification signals.

[0071] By designing a nonlinear weighting function, the weights are dynamically adjusted at different time scales to obtain time-varying weighted coefficients.

[0072] The decoupled signal and the verification signal are weighted and fused using time-varying weighting coefficients to obtain a comprehensive parameter set. This comprehensive parameter set includes multi-dimensional indicators such as blood routine parameters, liver function parameters, and kidney function parameters.

[0073] The detection results for each category are extracted based on the comprehensive parameter set, and the credibility score is calculated by combining the internal consistency of the comprehensive parameter set, cross-channel comparison results, and historical review records.

[0074] The system ultimately generates a comprehensive test report that includes classification test results and credibility scores, and provides query, printing, and data export functions through a human-computer interaction interface.

[0075] By employing multi-channel parallel detection and unified baseline correction within the same chip, the incomparability of multi-source signals due to time and environmental differences is effectively resolved. Interference identification and component decoupling reduce coupling interference between the device and the environment, improving the purity and stability of the original signal. Nonlinear mapping and cross-channel comparison correction enable accurate conversion of detection signals into medical indicators, reducing measurement deviations caused by single-channel anomalies. A hierarchical population knowledge base and verification mechanism enhance the clinical rationality and credibility of the test results. The system can automatically identify anomalies and perform secondary calculations. A dynamic weighted fusion method is used to integrate preliminary detection and verification results, further improving the robustness and credibility of the test conclusions. The comprehensive test report can directly provide integrated data support for medical personnel, reducing human judgment errors and enhancing clinical application value.

[0076] The foregoing has shown and described the basic principles, main features, and advantages of this application. Those skilled in the art should understand that this application is not limited to the above embodiments. The embodiments and descriptions in the specification are merely illustrative of the principles of this application. Various changes and modifications can be made to this application without departing from the spirit and scope thereof, and all such changes and modifications fall within the scope of this application as claimed. The scope of protection of this application is defined by the appended claims and their equivalents.

Claims

1. A multi-channel one-core multi-collection integrated detection method, characterized in that, The application is applied to a detection device integrated with a microfluidic chip in a detection channel; the method comprises: acquiring a time sequence original signal from blood routine, liver function and kidney function channels for parallel detection on the same blood sample on the microfluidic chip; based on the time sequence original signal, calculating a stable time-varying baseline of each detection channel through cross-channel baseline trend coupling and adaptive correction, correcting the time sequence original signal to obtain a baseline stable primary signal; performing component cancellation on the baseline stable primary signal to obtain a decoupled signal; mapping the decoupled signal to a physiological parameter value, reasonably testing the physiological parameter value, generating a reasonableness test result based on the reasonableness test, triggering a review mechanism to generate a review signal when the reasonableness test result indicates that there is a low credible parameter; based on the reasonableness test result and the review signal, performing weighted fusion to obtain a comprehensive parameter set; generating a comprehensive detection report based on the comprehensive parameter set.

2. The method of claim 1, wherein, The method comprises: performing sliding time window segmentation processing on the time sequence original signal, extracting statistical feature parameters and fitting a local trend curve in each time window to obtain a time-resolved baseline candidate curve; establishing a cross-channel baseline trend coupling model between the detection channels, comparing the correlation and offset between the baseline candidate curves of different detection channels to generate a cross-channel baseline trend; based on the cross-channel baseline trend, adaptively correcting the baseline candidate curves of each detection channel to remove noise and drift components to obtain a stable time-varying baseline.

3. The method of claim 2, wherein, The method comprises: subtracting the time sequence original signal of each detection channel from the stable time-varying baseline point by point to obtain a preliminary signal difference sequence; applying a smoothing algorithm and an outlier rejection algorithm to the signal difference sequence to suppress sharp peaks caused by transient interference; outputting the baseline stable primary signal on the blood routine, liver function and kidney function channels.

4. The method of claim 3, wherein, The method comprises: performing frequency domain transformation, time domain statistics and phase analysis on the baseline stable primary signal to extract frequency domain features, time domain features and phase coupling features to obtain an interference feature vector; inputting the interference feature vector into a preset machine learning model to identify fluid pulsation interference components, temperature gradient interference components and electromagnetic coupling interference components; based on each interference component, constructing a component decoupling matrix to cancel each interference component in the baseline stable primary signal to obtain a decoupled signal.

5. The method of claim 1, wherein, The method comprises: establishing a nonlinear response mapping relationship between physical quantities and physiological parameters for each detection channel to convert the decoupled signal to a candidate physiological parameter value; performing cross-channel comparison on the candidate physiological parameter values of each detection channel to remove abnormal values with large deviations from the physiological norm; outputting the physiological parameter values corrected through mapping and comparison in each detection channel.

6. The method of claim 5, wherein, The rationality test on the physiological parameter value, generating a rationality test result based on the rationality test, comprises: According to the gender, age, characteristics and medical history of the detection object, the corresponding physiological parameter reference range is obtained; The physiological parameter value is compared with the reference range, and a rationality index is calculated based on the statistical deviation; Generate a rationality test result based on the rationality index.

7. The method of claim 1, wherein, When the rationality test result indicates that there is a low credible parameter, trigger the review mechanism to generate a review signal, comprising: When the rationality test result indicates that there is a low credible parameter, the baseline stability primary signal and the decoupling signal are re-called, and repeated calculation is performed under different time windows and parameter thresholds to obtain a review physiological parameter value; The review physiological parameter value is compared for multiple times, and a review signal is generated.

8. The method of claim 7, wherein, The comprehensive parameter set is obtained by weighted fusion based on the rationality test result and the review signal, comprising: According to the rationality test result, the initial weight of the decoupling signal is allocated, and according to the consistency level of the review signal, the initial weight of the review signal is allocated; The initial weight is adjusted by a nonlinear weight function based on a time decay factor or a consistency index to obtain a time-varying weighting coefficient; The decoupling signal and the review signal are weighted and fused by using the time-varying weighting coefficient to obtain the comprehensive parameter set.

9. The method of claim 1, wherein, The comprehensive detection report is generated based on the comprehensive parameter set, comprising: Extract blood routine related parameters, liver function related parameters and kidney function related parameters in the comprehensive parameter set to form classified detection results respectively; Introduce a credibility calculation process in each classified detection result, generate a credibility score based on the internal consistency of the comprehensive parameter set, cross-channel comparison results and historical review results; Integrate the classified detection results and the credibility score to obtain a comprehensive detection report containing the detection results and the credibility score.

10. The method of claim 4, wherein, The method further comprises: The machine learning model is one of support vector machine model, random forest model or convolutional neural network model.

Images