Multi-algorithm comprehensive analysis method and system for detecting test strip anomaly

Abstract

The application discloses a multi-algorithm comprehensive analysis method and system for detecting test strip abnormalities. The method comprises the following steps: obtaining B-line and C-line related data of the test strip; determining whether the test strip has a test strip position offset by comparing the average value difference at the end of the B-line and the C-line according to the B-line and C-line related data; if not, analyzing the photoelectric signal change caused by sample addition sequence error or excess, extracting and verifying the abnormal interval to obtain an abnormal area; repairing the abnormal area to restore the normal shape of the curve to obtain a correction result; determining a value window with no intersection with all abnormal areas according to the correction result to obtain a value position calculation result; constructing a model according to the abnormal area, the correction result and the value position calculation result and matching the model with historical cases to calculate a final detection result and a confidence degree. Through the method, various types of abnormal conditions can be effectively identified, and the value position can be dynamically optimized to improve the data confidence degree.

Description

Technical Field

[0001] This invention relates to a method for analyzing test strip anomalies, and more specifically to a multi-algorithm integrated analysis method and system for detecting test strip anomalies. Background Technology

[0002] In the field of modern biomedical testing, test strips are widely used as a convenient and rapid testing tool. However, their use often faces several challenges. For example, in practical applications, test strips used with electronic testing instruments frequently experience issues such as positional misalignment and operational errors, affecting the final test results. Furthermore, users may make mistakes in the sample addition order or add too much sample, which can cause duplicate chromatography on the test strip, further interfering with the normal testing process. Existing technologies often only identify and handle specific types of abnormalities, lacking a systematic and comprehensive solution to address all the aforementioned potential problems.

[0003] Several technical solutions have been developed to address these issues. For example, some devices employ sensor monitoring technology to track the position of the test strip in real time, immediately stopping detection upon detecting any abnormalities. Other systems analyze trends in photoelectric sensor data to identify anomalies caused by user error. While these methods improve detection accuracy to some extent, they generally suffer from drawbacks such as low accuracy, high false alarm rates, and an inability to effectively repair identified abnormal areas. Furthermore, existing repair algorithms are mostly based on simple interpolation or smoothing, making it difficult to fully recover the true characteristics of the original data, thus affecting subsequent data analysis and result interpretation.

[0004] Therefore, it is necessary to design a new method that can not only effectively identify various types of abnormal situations, but also dynamically optimize the value position to improve data confidence, thereby significantly improving the accuracy and reliability of biomedical test strip detection results. This would address the shortcomings of existing technologies in dealing with issues such as positional deviation and user operation errors during biomedical test strip detection, which include low accuracy, high false alarm rate, and limited data repair capabilities. Summary of the Invention

[0005] The purpose of this invention is to overcome the shortcomings of the prior art and provide a multi-algorithm integrated analysis method and system for detecting anomalies in test strips.

[0006] To achieve the above objectives, the present invention adopts the following technical solution: a multi-algorithm integrated analysis method for detecting anomalies in test strips, comprising:

[0007] Obtain relevant data for lines B and C of the test strip to be tested;

[0008] Based on the relevant data of lines B and C, the difference in the average value at the end of lines B and C is compared to determine whether there is a test strip position shift in the test strip to be tested;

[0009] If the test strip is misaligned, the test result is deemed invalid.

[0010] If the test strip to be tested does not have a test strip position offset, analyze the changes in photoelectric signal caused by incorrect sample addition sequence or excessive amount, extract and verify the abnormal interval to obtain the abnormal area;

[0011] The abnormal areas are repaired to restore the normal shape of the curve, thus obtaining the corrected result;

[0012] Based on the correction results, a value window that does not intersect with all the abnormal regions is determined to obtain the value location calculation result;

[0013] A model is constructed based on the abnormal region, the correction result, and the value location calculation result, and matched with historical cases to calculate the final detection result and its confidence level.

[0014] This invention also provides a multi-algorithm integrated analysis system for detecting anomalies in test strips, including:

[0015] The acquisition unit is used to acquire relevant data of the B line and C line of the test paper to be tested;

[0016] The difference calculation unit is used to determine whether the test strip has a positional shift by comparing the average difference at the end of the B line and the C line based on the relevant data of the B line and the C line.

[0017] The invalidation unit is used to determine that the test result is invalid if the test strip under test has a positional offset.

[0018] The analysis unit is used to analyze the changes in photoelectric signals caused by incorrect sample addition sequence or excessive sample addition if the test strip to be tested does not have a test strip position offset, and to extract and verify the abnormal interval to obtain the abnormal area;

[0019] The repair unit is used to repair the abnormal area and restore the curve to its normal shape in order to obtain the correction result;

[0020] The location calculation unit is used to determine the value window that does not intersect with all the abnormal regions based on the correction result, so as to obtain the value location calculation result;

[0021] The multidimensional analysis unit is used to construct a model based on the abnormal region, the correction result, and the value location calculation result, and match it with historical cases to calculate the final detection result and its confidence level.

[0022] The beneficial effects of this invention compared to existing technologies are as follows: This invention obtains data from lines B and C of the test strip and compares the difference in the average values ​​at the ends of these two lines to determine whether there is an anomaly in line C, thereby determining the validity of the test results. For non-line C anomalies, it further analyzes the changes in photoelectric signals caused by incorrect sample addition order or excessive addition, extracts and verifies abnormal intervals to identify abnormal regions. Next, the identified abnormal regions are repaired to restore the normal shape of the curve, resulting in a corrected result. Based on this corrected result, a value window that does not intersect with any abnormal regions is determined to ensure the accuracy of the value location calculation. Finally, a model is constructed based on the abnormal regions, the corrected result, and the value location and matched with historical cases to calculate the final test result and its confidence level. This method can not only effectively identify various types of anomalies but also dynamically optimize the value location to improve data confidence, significantly improving the accuracy and reliability of biomedical test strip test results. It solves the shortcomings of existing technologies, such as low accuracy, high false alarm rate, and limited data repair capabilities, when dealing with issues like positional shifts and user operation errors during biomedical test strip testing.

[0023] The present invention will be further described below with reference to the accompanying drawings and specific embodiments. Attached Figure Description

[0024] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0025] Figure 1 A flowchart illustrating the multi-algorithm integrated analysis method for detecting test strip anomalies provided in this embodiment of the invention;

[0026] Figure 2 This is a schematic block diagram of a multi-algorithm integrated analysis system for detecting test strip anomalies provided in an embodiment of the present invention;

[0027] Figure 3 A schematic block diagram of a computer device provided in an embodiment of the present invention;

[0028] Figure 4 This is a C-line anomaly detection curve provided in an embodiment of the present invention;

[0029] Figure 5 A graph showing the calculation of the value position provided in an embodiment of the present invention;

[0030] Figure 6 A graph illustrating the repair of abnormal regions provided in an embodiment of the present invention;

[0031] Figure 7 A graph illustrating anomaly identification provided in an embodiment of the present invention. Detailed Implementation

[0032] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0033] It should be understood that, when used in this specification and the appended claims, the terms "comprising" and "including" indicate the presence of the described features, integrals, steps, operations, elements and / or components, but do not exclude the presence or addition of one or more other features, integrals, steps, operations, elements, components and / or collections thereof.

[0034] It should also be understood that the terminology used in this specification is for the purpose of describing particular embodiments only and is not intended to limit the invention. As used in this specification and the appended claims, the singular forms “a,” “an,” and “the” are intended to include the plural forms unless the context clearly indicates otherwise.

[0035] It should also be further understood that the term "and / or" as used in this specification and the appended claims refers to any combination of one or more of the associated listed items and all possible combinations, and includes such combinations.

[0036] Please see Figure 1 , Figure 1 This is a flowchart illustrating a multi-algorithm comprehensive analysis method for detecting test strip anomalies provided in an embodiment of the present invention. This method is applied in a server. It determines whether line C is abnormal by comparing the difference in average values ​​at the ends of lines B and C. After confirming no abnormality in line C, it further analyzes changes in photoelectric signals caused by incorrect sample addition order or excessive sample addition, extracting and verifying abnormal intervals to obtain abnormal regions. For the identified abnormal regions, a dynamic adjustment and repair strategy is used to restore the curve's normal shape, and a value window that does not intersect with all abnormal regions is determined to optimize data confidence. A model is constructed based on the abnormal regions, correction results, and value positions, and matched with historical cases to finally calculate the detection results and their confidence levels. This method not only effectively identifies various types of anomalies but also significantly improves the accuracy and reliability of biomedical test strip detection results by dynamically optimizing value positions, solving the problems of low accuracy, high false alarm rate, and limited data repair capabilities in existing technologies.

[0037] Figure 1 This is a flowchart illustrating the multi-algorithm integrated analysis method for detecting anomalies in test strips provided in an embodiment of the present invention. Figure 1 As shown, the method includes the following steps S110 to S170.

[0038] S110. Obtain relevant data for the B and C lines of the test paper to be tested.

[0039] In this embodiment, the data related to lines B and C refer to the photoelectric signal intensity values ​​or corresponding numerical representations collected from the test strip under test. These data are obtained by reading the changes in light intensity reflected or transmitted from different detection lines (i.e., lines B and C) on the test strip using a specific instrument. Specifically:

[0040] Line B (background line): Represents the background area between lines T and C. Its signal intensity is directly related to the chromatography process of the test strip. Therefore, changes in the data of line B can reflect whether the chromatography process is normal.

[0041] C-line (control line): As an internal quality control standard, the C-line should show a certain signal intensity regardless of the presence of the target analyte. If the C-line signal is too low, it indicates that the test strip may be ineffective.

[0042] The specific steps for obtaining relevant data include, but are not limited to:

[0043] The user starts the electronic testing instrument, and the instrument calibrates automatically.

[0044] The user adds the sample to the sample application port of the device and waits for the test strip in the instrument to complete the reaction before the device outputs the test result.

[0045] The instrument collects photoelectric signal timing data of the test strip during the step-by-step reaction process.

[0046] After the test strip in the instrument completes the reaction, data acquisition ends and the data set is output.

[0047] Data analysis preparation: Organize the collected data into a dataset suitable for subsequent analysis. For example, calculate the average signal strength at each time point and build a time series.

[0048] The data obtained from the B and C lines in the above process form the basis for subsequent algorithm execution, including but not limited to the C line anomaly detection algorithm, anomaly identification algorithm, and process value curve repair algorithm. This aims to ensure the accuracy and reliability of the detection results, while providing effective solutions for potential problems.

[0049] S120. Based on the relevant data of lines B and C, determine whether there is a test strip position shift by comparing the difference in the average values ​​at the ends of lines B and C.

[0050] In one embodiment, step S120 described above may include steps S121 to S122.

[0051] S121. Calculate the average value of several data points at the end of lines B and C based on the relevant data of lines B and C.

[0052] In this embodiment, as Figure 4 As shown, in order to determine whether the positional shift of the test strip in the electronic testing instrument affects the accuracy of the photoelectric signal, it is necessary to first calculate the average value of the data at the end of each line (B and C). Specifically:

[0053] Select the last 10 data points from the B-line dataset (denoted as Bs), calculate the average of these data points, and denot it as Baverage: ;

[0054] Similarly, from the C-line dataset (denoted as Cs), select the last 10 data points and calculate the average of these data points, denoted as the Average: ;

[0055] Here, the Avg() function calculates the average of a given dataset.

[0056] S122. When the absolute value of the difference between the data of lines B and C at the initial moment and the average value of several data points at the end of lines B and C is not greater than the set difference threshold, it is determined that the test strip has a test strip position offset.

[0057] In this embodiment, after obtaining the average values ​​of the last portions of lines B and C, the next step is to compare them with the data of lines B and C at the initial time (i.e., the 0th data point) to determine whether there is an anomaly in line C. If so, it indicates that there is a shift in the test strip position. The specific operation is as follows:

[0058] Obtain initial data points: Line B Initial data points: B s0 Initial data points for line C: C s0 ;

[0059] Calculate the difference between the initial variance and the final mean: Initial variance: B s0 - C s0 The difference between the last two average values: Average - Average;

[0060] when This indicates an abnormality in the C line of the test strip. This is because the positional shift of the test strip causes the signal change trends of the B and C lines to become more consistent, resulting in very close differences in the initial and final data between the two.

[0061] in, : The data with index 0 in the B-line data set.

[0062] : The data with index 0 of the C-line data set.

[0063] : A function that calculates the average of the last 10 data points in a set.

[0064] : A function that calculates the absolute value.

[0065] During this process, if an abnormality is detected in the C line, it indicates that the test strip may be causing inaccurate readings due to improper assembly or other physical reasons, requiring appropriate processing of the results or retesting. This inspection mechanism helps ensure the reliability of test results and enables the timely detection and correction of potential problems.

[0066] S130. If the test strip is misaligned, the test result is deemed invalid.

[0067] In this embodiment, when the algorithm in step S120 determines that the test strip has an abnormality in line C (i.e., the initial and final differences between lines B and C are very close, indicating that the test strip position may have shifted), the result of this test is directly determined to be invalid. This is because positional shift will cause inaccurate photoelectric signals to be collected, thus affecting the final test result.

[0068] S140. If the test strip to be tested does not have a test strip position offset, analyze the changes in photoelectric signal caused by incorrect sample addition sequence or excessive amount, extract and verify the abnormal interval to obtain the abnormal area.

[0069] In this embodiment, the abnormal region refers to the interval in the photoelectric signal data where the signal exhibits a 'decline-rebound' characteristic in a short period of time due to incorrect sampling order or excessive sampling.

[0070] After confirming that the test strip did not show any abnormal C-line, the next step is to further analyze the changes in photoelectric signal caused by incorrect sample addition order or excessive sample amount by the user, and extract all valid abnormal intervals from the analysis.

[0071] In one embodiment, step S140 described above may include steps S141 to S148.

[0072] S141. Obtain relevant data for the test strip to be tested.

[0073] First, collect all relevant data for the test strip to be tested, including the data sets for the T line (detection line), B line, and C line, specifically the photoelectric signal intensity values ​​or their corresponding numerical representations.

[0074] S142. Set the abnormal interval set, time span, and threshold key parameters, and initialize the abnormal state to be off.

[0075] In this embodiment, the necessary variables and sets are defined:

[0076] Abnormal interval set: Tes, Bes, Ces;

[0077] Time span: DTS = 2 seconds / data point;

[0078] Maximum time span: MaxInterval = 400 seconds;

[0079] Minimum time span: MinInterval = 50 seconds;

[0080] Decline threshold for enabling anomaly detection: DeclineThreshold=0.0026;

[0081] Maximum acceptable rate of variation: DiscrepancyRate = 0.005;

[0082] Anomaly detection threshold: TDecision=15;

[0083] Boundary anomaly detection threshold: TBoundaryDecision=4;

[0084] The abnormal state Er is initialized to be off.

[0085] S143. Analyze the trend and estimated value of each data point in the relevant data of the test paper to be tested through a sliding window to determine whether the conditions for enabling the anomaly judgment are met.

[0086] In one embodiment, step S143 described above may include steps S1431 to S1433.

[0087] S1431. Use a data point from the relevant data of the test paper to be tested to establish a sliding window for calculating the trend of change, and use subsequent data points to establish a sliding window for anomaly judgment.

[0088] S1432. Calculate the average value, average increment, and estimated value of the two sliding windows.

[0089] In this embodiment, the average of the two sliding windows refers to the arithmetic mean of all data points in each specified set of data points (such as L window and R window).

[0090] The average increment of two sliding windows refers to the change in the data from the first to the last data point within the L-window, divided by the number of data points minus one (i.e., the average change per data point), used to reflect the direction and strength of the data trend.

[0091] The estimated value of two sliding windows refers to the prediction of the value of the next data point based on the data change trend within the L-window. It is usually estimated by adding the average increment to the last actual data point.

[0092] S1433. Based on the relationship between the average increment of the sliding window that calculates the trend of change and the current value and the estimated value, when the relationship meets the conditions and the decline rate exceeds the set threshold, the abnormal judgment mode is entered and relevant information is recorded; wherein, the abnormal judgment mode is to set the abnormal state to be enabled; the relevant information includes the left boundary value and the left boundary index.

[0093] For each data point i, construct two sliding windows:

[0094] L window: contains data points preceding the current point (e.g., the first 6);

[0095] R window: contains data points after the current point (e.g., the last 3).

[0096] Calculate the average value Laverage and average increment LIncrement of the L-window;

[0097] Calculate the average value (RAverage) of the R window;

[0098] The estimated value is calculated based on the average increment of the L-window.

[0099] If any of the following conditions are met:

[0100] The average increment of the L-window is positive and the current value is less than the previous value;

[0101] The average increment of window L is negative and the current value is less than the expected value;

[0102] Then verify whether the decline rate exceeds the set threshold (i.e., estimatedValue / RAverage-1≥DeclineThreshold). If it does, enter the anomaly judgment mode, set Er='On', and record the left boundary information.

[0103] S144. When the conditions for enabling anomaly detection are met, mark the current index as the starting point of the anomaly and record the left boundary information, activate the anomaly detection mode, and obtain the anomaly interval.

[0104] Once the anomaly detection mode is entered, the current index is used as the starting point of the anomaly interval, and the information of the left boundary is recorded.

[0105] S145. Continuously monitor the rate of decline, time span, lowest point position and corresponding minimum value of the abnormal interval to obtain monitoring results.

[0106] In this embodiment, the monitoring result refers to continuously tracking indicators such as the rate of decline, time span, lowest point position and corresponding minimum value of the identified abnormal interval, in order to evaluate the effectiveness and characteristics of the abnormal interval.

[0107] In anomaly detection mode, various indicators within the anomaly range are continuously monitored, including the rate of decline, time span, location of the lowest point, and its corresponding minimum value.

[0108] S146. Adjust the end position of the abnormal interval using the angle calculation method.

[0109] In one embodiment, step S146 described above may include steps S1461 to S1462.

[0110] S1461. Determine the position of the right boundary based on the monitoring results;

[0111] S1462. Use the angle calculation method to optimize the position of the right boundary to obtain the end position of the abnormal interval.

[0112] S147. Evaluate the validity of the abnormal interval based on its data characteristics. If the abnormal interval does not meet the standard, it is considered invalid.

[0113] In one embodiment, step S147 described above may include steps S1471 to S1472.

[0114] S1471. Verify the validity of the end position of the abnormal interval, including checking the descent concentration and calculating whether the descent angle meets a specific threshold.

[0115] S1472. If the end position of the abnormal interval is valid, add the end position of the abnormal interval to the set; otherwise, reset the relevant state variables.

[0116] S148. For cases where the sequence ends or the time span is not met, apply specified rules to process them in order to identify and confirm all valid abnormal intervals in order to obtain the abnormal region.

[0117] Specific rules are applied to handle abnormal intervals at the end of the sequence or those that do not meet the time span requirements, ensuring that all valid abnormal intervals are correctly identified and confirmed. This comprehensively covers all possible anomalies and improves the reliability of the detection results.

[0118] In this embodiment, if the user adds samples in the wrong order or adds too much sample, the test strip may exhibit repeated chromatography. In this case, the photoelectric signal data will be affected, showing a short-term 'decline-rebound' phenomenon on the curve. Based on this characteristic, the system can extract all abnormal areas and prompt the user to correct the sample addition behavior. When calculating the test results, these abnormal areas need to be ignored or corrected.

[0119] Setting parameters:

[0120] Abnormal interval set: Let Tes be the abnormal interval set on Ts, Bes be the abnormal interval set on Bs, and Ces be the abnormal interval set on Cs, where Ts, Bs, and Cs refer to the relevant data corresponding to the T line, B line, and C line, respectively.

[0121] Time Span (DTS): The time span between two data points is 2 seconds (i.e., one data point every two seconds).

[0122] Maximum and minimum time spans: The maximum time span for the abnormal interval is MaxInterval = 400 seconds, and the minimum time span is MinInterval = 50 seconds.

[0123] Decline Threshold: Decline Threshold = 0.0026 (same for T line and B line, independent for C line).

[0124] Maximum difference rate: The acceptable maximum difference rate is DiscrepancyRate=0.005 (T line and B line are the same, C line is independent).

[0125] Anomaly detection threshold: TDecision=15 (T line and B line are the same, C line is independent).

[0126] Boundary anomaly detection threshold: TBoundaryDecision=4 (T line and B line are the same, C line is independent).

[0127] Error detection status Er: Initial value is "Off".

[0128] The logic for identifying abnormal regions is as follows: Identify valid abnormal regions on Ts:

[0129] Traverse Ts (from index) At the beginning, at the end of each traversal, make Each iteration performs the following steps:

[0130] if Then jump to "Append i to" In the middle, calculate the time span of the current data. Calculate the current rate of decline. ,calculate corresponding The smallest element among the element values ​​is in Position index in (Right now The lowest point in the interval (if multiple points exist, the last one is selected), and the value of the lowest point is recorded. ".

[0131] exist Create a sliding window L of length 6 in front of it (i.e., L = [ ]), used to calculate the trend of curve changes, in After that, create a sliding window R of length 3 (i.e., R = [ If the remaining length is less than three, a window is created using the actual remaining length for anomaly detection. The average value of the window range L is calculated. average increment Calculate the estimated value Calculate the average value over the range of window R. Determine whether to enable anomaly detection: If the average increment of window L is positive (i.e., ... And the current value is less than the previous value (i.e.: ), or the average increment of the L-window is negative (i.e.: <0) and the current value is less than the estimated value (i.e.: If the current rate of decline exceeds the threshold, then verify whether the rate of decline exceeds the threshold: if the rate of decline exceeds the threshold (i.e.: Then it enters the exception detection mode, and lets ,Will Add to In the middle, record the left boundary value. Left boundary index Define the right boundary index Then proceed to the next iteration.

[0132] if If so, proceed to the next iteration; otherwise, continue with the subsequent steps. Append i to... In the middle, calculate the time span of the current data. Calculate the current rate of decline. ,calculate corresponding The smallest element among the element values ​​is in Position index in (Right now The lowest point in the interval (if multiple points exist, the last one is selected), and the value of the lowest point is recorded. .

[0133] Verify the validity of anomalies and the end position of calculation: Based on the judgment conditions of the current data's time span and index position, execute one of the following four processes: First: The current time span is greater than or equal to the maximum time span threshold (i.e., If the lowest point is located at the end of the interval (i.e.: If the interval is invalid, then the program will jump to the "Invalid interval, let" state. ='Close', roll i back to Location point (i.e.) ), clear "Proceed to the next traversal", otherwise reset the right boundary index. Optimize the right boundary index using an angle calculation method: Define the minimum descent angle. ; Traversal (from index) At the beginning, at the end of each traversal, make ), and execute the following logic: calculate the smoothing value of the current point. ;

[0134] Calculate the height difference between the left boundary value and the smoothed value of the current point. Let 'a' be the length of the right-angled leg 'a'. Calculate the distance from the current point to the right boundary on the time axis. Let b be the length of the right-angled side b.

[0135] Enhance the sensitivity of angle to changes in height difference (Where, "=" indicates assignment), reducing the sensitivity of angles to changes over time. (Where, "=" indicates assignment.) By adjusting these two sensitivities, an optimal balance can be found between the time span and the difference, resulting in a suitable ending position. Calculate the length of the hypotenuse c. Calculate the descent angle. .if Then let Right boundary index After the traversal ends, the position of the point with the smallest descent angle can be obtained. Based on this position, Cut off .

[0136] Second: The current set has reached its final boundary (i.e., If this condition is met, then in step "If the right boundary is not determined (i.e. The validity of the abnormal interval will not be further verified in the ")" section.

[0137] If the lowest point is within the last 10 data points of the set or the current time span is less than the minimum time span threshold (i.e. or Then, directly calculate the result: calculate the height difference between the left boundary and the lowest point of the abnormal interval. Let 'a' be the length of the right-angled leg 'a'. Calculate the distance on the time axis from the left boundary to the lowest point of the outlier interval. Let b be the length of the right-angled side b. Reduce the sensitivity of the angle to changes over time. (Where, "=" indicates assignment), to eliminate the magnified effect of the angle caused by the short distance.

[0138] Calculate the length of the hypotenuse Calculate the descent angle. .

[0139] If the descent angle matches the boundary descent angle, it is determined to be a valid anomaly region (i.e., ...). ), with the current point as the right boundary (i.e. ),Will Join In the middle, clear ,make ='Close', ends the iteration.

[0140] Otherwise, verify the recovery rate of the curve. If the recovery rate reaches 60% or more (i.e., If the right boundary index is optimized using the angle calculation method (consistent with the logic of optimizing the right boundary index using the angle calculation method in section 2), then the current point is used as the right boundary (i.e., ...). ).

[0141] Third: The current time span is within the threshold range (i.e.) );

[0142] If the decline and rebound have already been completed (i.e.) ) or recover to within an acceptable range of difference (i.e. If the current point is recorded as the right boundary (i.e., ...), then the current point is recorded as the right boundary. ).

[0143] Fourth: The current time span is less than the minimum time span threshold (i.e. )

[0144] Verify the validity of the current outlier interval: If the current value is greater than or equal to the left boundary value (i.e. )or( and If this is an abnormal interval, then the process is considered invalid, and the procedure is repeated: "Abnormal interval is invalid, let..." ='Close', roll i back to Location point (i.e.) ), clear Then proceed to the next iteration.

[0145] If the right boundary is not determined (i.e.) If the condition is true, proceed to the next iteration; otherwise, further verify the validity of the abnormal region.

[0146] Rank the decrease values ​​of each point in the decreasing interval of the abnormal region, and take the top N largest decrease values; when the interval length is ≥30, N=5; when 20≤interval length<30, N=4; when 10≤interval length<20, N=3; when 5≤interval length<10, N=2; when the interval length is <5, directly determine the abnormal interval as invalid, and jump to step "Abnormal interval is invalid, let..." ='Close', roll i back to Location point (i.e.) ), clear Then proceed to the next iteration.

[0147] Where: Interval length = If the first N decreasing values ​​contribute more than 65% of the decrease rate, the decrease is considered too concentrated and invalid, and the process jumps to step 2.0; otherwise, the decrease angle is calculated to see if it meets the threshold.

[0148] Calculate the height difference between the left boundary and the lowest point of the outlier interval. Let a be the length of the right-angled side a.

[0149] Calculate the distance on the time axis from the left boundary to the lowest point of the outlier interval. Let b be the length of the right-angled side b.

[0150] Calculate the length of the hypotenuse .

[0151] Calculate the descent angle .

[0152] If the descent angle is greater than the threshold (i.e.) If the exception is valid, proceed to step "Exception range valid, then..." Join In the middle, clear ,make ='Close', proceed to the next iteration.

[0153] The abnormal interval is invalid, let ='Close', roll i back to Location point (i.e.) ), clear Then proceed to the next iteration.

[0154] The abnormal range is valid, and will Join In the middle, clear ,make ='Close', proceed to the next iteration.

[0155] Furthermore, the recognition logic for Cs and Bs is the same as above, which will... Join and middle.

[0156] S150. Repair the abnormal area to restore the normal shape of the curve and obtain the corrected result.

[0157] In this embodiment, the correction result refers to processing the identified abnormal areas through a series of steps to restore them to their normal form, thereby ensuring that the final output data is accurate and reliable.

[0158] In one embodiment, step S150 described above may include steps S151 to S155.

[0159] S151. Set the single increment step size to an integer.

[0160] To preserve the variation characteristics of the original data, the increment step size must be an integer.

[0161] S152. Calculate the length of the abnormal region.

[0162] Record the length EL of the current abnormal region, i.e., EL = length(Tesk), where Tesk is an element in the set of abnormal regions.

[0163] S153. Determine the left boundary index and right boundary index of the abnormal region, and record the left boundary value and right boundary value of the abnormal region.

[0164] Determine the left boundary index LI (i.e., LI=Tesk,0) and the right boundary index RI (i.e., RI=Tesk,EL-1) of the current abnormal region.

[0165] Record the left boundary value LV (i.e., LV=Ts[LI]) and the right boundary value RV (i.e., RV=Ts[RI]).

[0166] S154. Calculate the integer increment based on the difference between the left boundary value and the right boundary value.

[0167] Calculate the difference between the left and right sides, DF: DF = LV - RV.

[0168] Calculate the required integer increment Increment based on the difference: Increment = Truncate(Abs(DF) / EL). If DF < 0, then Increment = -Increment; otherwise, Increment = Increment.

[0169] S155. When the abnormal region is located at the end of the curve or the integer increment meets the requirements, use LV to correct all values ​​in the abnormal region, including: initializing the left-side repair value and calculating the margin; when the margin is not 0, adjusting the values ​​in the abnormal interval using an arithmetic sequence method; if the margin is 0, setting the values ​​in a specific range to RV; if there is a margin and it is not 0, fine-tuning the values ​​at specific positions in the abnormal region.

[0170] Specifically, the repair value on the left is Rlv, and the initial value is Rlv=LV.

[0171] Calculate the margin: Remainder = Abs(DF) mod EL.

[0172] Adjust the values ​​in the abnormal range:

[0173] If the overall increment is not 0 (i.e., Increment ≠ 0), then let Rlv = Rlv + Increment, and use the following method to correct the values ​​in the abnormal interval:

[0174] For each point i within the abnormal interval, calculate its value as Ts[i] = RV + Increment * (i - LI), where LI ≤ i ≤ RI.

[0175] Other positions remain unchanged: Ts[i], others.

[0176] If the overall increment is 0 (i.e., Increment=0), then the value within the specific range is directly set to RV:

[0177] Ts[i]=RV, for LI+Remainder <i≤RI。

[0178] Other positions remain unchanged: Ts[i], others.

[0179] Fine-tune the values ​​at specific locations within the abnormal region:

[0180] If the margin is not 0 (i.e., Remainder ≠ 0), then the values ​​in the abnormal range will be further corrected:

[0181] For points within the range LI≤i≤LI+Remainder, adjust their values ​​as follows:

[0182] Ts[i] = Rlv - (i - LI), DF < 0;

[0183] Ts[i] = Rlv + (i - LI), DF > 0;

[0184] Other positions remain unchanged: Ts[i], others.

[0185] Suppose there is an anomalous region with a left boundary value LV of 100, a right boundary value RV of 90, and a length EL of 10. The calculated difference DF is 10, and the integer increment Increment is 1 (because Truncate(10 / 10) = 1). We can then repair it using the steps described above:

[0186] Initialize the left-side repair value Rlv to 100.

[0187] The calculated residual Remainder is 0 (because Abs(10)mod10=0).

[0188] Since the overall increment is not zero, an arithmetic sequence method is used for repair:

[0189] For each point i within the abnormal interval, calculate its value as Ts[i] = 90 + 1*(i - LI), where LI ≤ i ≤ RI.

[0190] If the margin is 0, no further fine-tuning is required.

[0191] This allows for the effective repair of abnormal areas, restoring them to their normal shape and ensuring the accuracy of subsequent analyses.

[0192] In this embodiment, the single incremental step size is set to be an integer to preserve the change characteristics of the original data; the following logic is executed to repair the curve data:

[0193] Traverse the set of abnormal intervals (from index) start):

[0194] Let the length of the current abnormal interval be... .

[0195] Let the left boundary index of the current abnormal region be . (Right now The right boundary index is (Right now ).

[0196] Let the left boundary value of the current abnormal region be... (Right now The right boundary value is (Right now ).

[0197] Calculate the difference between the left and right sides Calculate the required integer increment , ;

[0198] If the abnormal region is located at the end boundary of the curve (i.e.) ) or the difference is 0 (i.e. Then the values ​​in the outlier region are corrected using the left boundary value (i.e., ...). Otherwise, perform the following repair methods to correct the values ​​in the abnormal range:

[0199] Record the repair value on the left side. Calculate the margin .

[0200] If the overall increment is not 0 (i.e.) ), then let The abnormal range values ​​are repaired using a repair method:

[0201] (That is: to restore the values ​​in the abnormal range to a tolerance of 0.5%) (arithmetic sequence)

[0202] Otherwise, use repair method two to repair the values ​​in the abnormal range:

[0203] (That is: index the Ts set) Modify the value of all elements in the region up to index RI to RV).

[0204] If the margin is not 0 (i.e.) If so, further corrections will be made to the values ​​in the abnormal range: .

[0205] S160. Based on the correction result, determine the value window that does not intersect with all the abnormal regions to obtain the value position calculation result.

[0206] In this embodiment, the value location calculation result refers to a series of steps to ensure that the data points used to calculate the detection result do not contain any abnormal or repaired data, thereby improving the confidence and accuracy of the result.

[0207] In one embodiment, such as Figure 6 As shown, step S160 above may include steps S161 to S165.

[0208] S161. Record the length of the entire sequence of the correction result.

[0209] First, we need to know the length of the complete data sequence after anomaly identification and repair. This is to determine from which position to start traversing backward to find a suitable value window.

[0210] S162. Starting from the last element of the sequence, traverse backwards to find the value window.

[0211] Since the closer the value is to the end of the sequence, the more fully the test strip reacts and the higher the confidence level of the data, we choose to traverse backward from the end of the sequence to find the optimal value window.

[0212] S163. In each iteration, a set containing the indices of the most recent elements is created.

[0213] Here, "the most recent few elements" typically refers to a fixed-length window (e.g., 10 elements) that contains a subset of data points from the end of the sequence. Each iteration updates the position of this window and examines the data points within it.

[0214] S164. Check whether the set intersects with any of the abnormal regions.

[0215] For each created window, it needs to be checked whether it overlaps with any known anomalous areas. If there is any overlap, the window is not suitable for retrieving values.

[0216] S165. When a window that does not intersect with any of the aforementioned abnormal regions is found, the window is confirmed as a valid value window.

[0217] Once a window is found that does not overlap with any anomalous regions, it can be identified as a valid value window. This means that the data within this window is considered reliable enough to be used to calculate the final detection result.

[0218] Through this process, the system can effectively avoid using data points that may be affected by interference or errors, thereby improving the accuracy and reliability of the overall detection results. If such a window cannot be found, it may indicate that anomalies are too dense, rendering the detection results invalid.

[0219] In this embodiment, as Figure 5 As shown, after the abnormal region is extracted, the dynamic value positions of the three curves and the common value positions of the three curves are calculated based on the location distribution of the abnormal region. If no usable common value position of the three curves is found, it indicates that the abnormalities are too dense, the data confidence is too low, and the output detection result is invalid.

[0220] Record curve length is ;

[0221] Create a loop (from index) At the beginning, at the end of each loop, In each iteration of the loop, the following logic is executed:

[0222] Establish a set D of value window indices of length 10 (i.e., ), verify set D and abnormal interval set (i.e. If there is an intersection, the current loop ends and the next loop begins; otherwise, the loop ends with D as the value window.

[0223] If the number of loops is exhausted (i.e.) If no value window can be found, the detection result is deemed invalid.

[0224] Different sets of abnormal intervals are calculated based on the selection. This can calculate the dynamic value window of the tribometer and the same value window of the tribometer, such as: only calculating whether it is the same as the tribometer. If an intersection exists, the dynamic value position of the T-line can be calculated. Simultaneously, it is calculated whether it intersects with... If there is an intersection, the positions where the three curves have the same value can be calculated.

[0225] S170. Construct a model based on the abnormal region, the correction result, and the value location calculation result, and match it with historical cases to calculate the final detection result and its confidence level.

[0226] In this embodiment, the final detection result refers to the detection result calculated by analyzing and quantifying multi-dimensional features such as abnormal regions, correction results, and value locations, and matching these features with historical cases. This result includes not only specific numerical outputs but also a confidence score to indicate the reliability of the result.

[0227] Confidence level is an indicator that measures the reliability or accuracy of current detection results. It reflects the system's level of confidence in the obtained results and considers various factors such as the density of anomaly distribution, the consistency of anomaly region morphology, the degree of data repair, the rationality of the value selection location, and the correctness of the sample type. The higher the confidence level, the more reliable the result; conversely, a lower confidence level may require re-detection or further verification.

[0228] In one embodiment, such as Figure 7 As shown, step S170 above may include steps S171 to S174.

[0229] S171. Analyze and quantify the abnormal distribution, morphology, repair ratio, value location and sample type based on the abnormal area, the correction result and the value location calculation result to obtain multi-dimensional features.

[0230] In this embodiment, multi-dimensional features refer to:

[0231] Anomaly distribution characteristics: including the distribution of anomaly regions on the time axis (anomaly density, concentration, and overlap).

[0232] Anomaly morphological characteristics: These involve the descent angle of the anomaly region and its variance, used to assess the severity and consistency of the anomaly changes.

[0233] Repair ratio characteristics: describes the proportion of data points repaired to the total number of data points during the curve repair process, as well as the amount of numerical adjustment during the repair process.

[0234] Value location characteristics: Consider the confidence level of the value location (based on its distance from the end boundary of the data), location consistency and window stability.

[0235] Sample type characteristics: These reflect whether the sample added is correct and whether the amount added is standardized, directly affecting the overall confidence level of the test results.

[0236] S172. Construct a comprehensive model based on the multi-dimensional features to describe the overall features of the current detection situation.

[0237] In this embodiment, the comprehensive model refers to the integration of all multi-dimensional features obtained from S171 to form a mathematical or logical model that can comprehensively describe the current detection status. This model can be used for subsequent similarity comparison with historical cases to find the closest historical case as a reference to predict the final detection result.

[0238] S173. Find historical case models that meet the similarity requirements of the comprehensive model, and calculate the final detection result based on the algorithm and parameters.

[0239] This step involves using an algorithm to search a cloud-based library of historical feature models for historical case models that are most similar to the currently constructed comprehensive model. Once found, the final result of the detection is derived based on the calculation method and parameters corresponding to that historical case model.

[0240] S174. Based on preset rules and the confidence level of the final detection result, determine the content of high-confidence result, low-confidence suggestion, or invalid result prompt.

[0241] In this step, an overall confidence score is calculated based on pre-defined confidence increase / decrease conditions and all extracted features. The output content is then determined according to the score.

[0242] High confidence results: If the confidence level is higher than a certain threshold (e.g., 0.7), the detection result is output directly with a confidence score.

[0243] Low confidence results: If the confidence level is in the middle range (e.g., 0.5 to 0.7), the output will suggest that the user retest and provide the possible reasons for the low confidence level.

[0244] Invalid result prompt: When the confidence level is lower than the minimum acceptable threshold (e.g., 0.5), the system determines that the detection is invalid, does not output a specific value, and provides the reason for failure to guide the user to take the next step.

[0245] Through the above steps, the system can intelligently process complex photoelectric signal data, identify and correct anomalies, select appropriate value locations, and calculate reliable detection results based on rich historical data.

[0246] In this embodiment, a fusion feature model is constructed by comprehensively measuring the following multi-dimensional features: the distribution of abnormal regions on the three curves, the sample type determined by the algorithm, the span of fixed and dynamic value positions on the time axis, and the correction of abnormal regions. Subsequently, the model with the highest similarity is matched in the cloud feature model library, and the algorithm and parameters are calculated based on the result value corresponding to the model to calculate the detection result and output the confidence score.

[0247] Extracting anomaly distribution characteristics: The distribution of anomaly regions along the time axis reflects the stability of the detection process. By calculating anomaly density, concentration, and overlap, the frequency and range of anomalies can be quantified.

[0248] Anomaly density: defined as the ratio of the total length of the anomaly region to the total data length. High density may indicate a systemic problem in the detection process.

[0249] Anomaly concentration: This is assessed by calculating the average interval between anomaly regions; the smaller the interval, the more concentrated the anomalies.

[0250] Multi-curve overlap: reflects the synchronicity of the three curves. High overlap usually indicates a common problem in the sample addition process or the instrument.

[0251] Extracting morphological features of abnormal regions: The morphological features of abnormal regions include the descent angle and its variance.

[0252] Descent angle: Calculated using trigonometric functions, it reflects the severity of the anomalous change. The average descent angle provides a measure of the overall intensity of the anomaly.

[0253] Descent angle variance: Characterizes the consistency of anomaly morphology. Lower variance indicates relatively consistent anomaly types, while higher variance may indicate the presence of multiple different types of anomalies.

[0254] Extracting repair ratio features: Curve repair is a correction process for abnormal data, and repair features reflect the strength and scope of the repair.

[0255] Repair scope: Indicates the percentage of data points to be repaired out of the total data points.

[0256] Repair intensity: Quantifies the amount of numerical adjustment during the repair process. These features help assess the quality of the original data and the degree of intervention by the repair algorithm.

[0257] Extracting value location features: The choice of value location directly affects the final detection value. Location features include the confidence of the value location and the data stability on both sides of the value window.

[0258] Confidence level of the value location: calculated based on the distance from the end boundary of the data. The closer to the end boundary of the data (i.e., the later the value location is on the time axis, the more fully the test strip reacts and the higher the data confidence level), the higher the confidence level, and vice versa.

[0259] Position consistency and window stability: reflect the rationality of value selection.

[0260] Extracting sample type features:

[0261] Sample type characteristics reflect whether the user's sample addition is correct and whether the sample volume is standardized. Incorrect sample addition will fundamentally lead to deviations in the test results, reducing the overall confidence level of the test results.

[0262] Each model corresponds to a past detection record with clearly defined operation processes and curve characteristics (including its final validated detection result and the confidence level calculated at that time). After feature extraction, similarity comparison is performed with features within various models in the cloud-based model library. An algorithm searches for the "nearest neighbors" historical models in a multi-dimensional feature space. These matched historical models represent past cases most similar to the current detection situation. After matching the model with the highest similarity, the result value is calculated using the model's result value calculation algorithm and parameters. For example:

[0263] When the anomaly density is high and the data repair intensity is low, the result value is calculated by taking a fixed tail value from the repaired data.

[0264] When the anomaly density is low and the data repair intensity is high, the result value is calculated using dynamic value location.

[0265] When the anomaly density is extremely low or zero, the slope algorithm is used to calculate the result value.

[0266] The confidence level of the result is calculated based on the preset confidence level increase / decrease conditions and all extracted features.

[0267] The conditions for increasing or decreasing confidence level are as follows:

[0268] The higher the anomaly density, the lower the confidence level of the result;

[0269] The higher the variance among the outlier regions, the lower the confidence level of the result.

[0270] The closer the value is to the end of the data, the higher the confidence level of the result.

[0271] Incorrect sample type significantly reduces the confidence level of the results.

[0272] The system determines the output content based on the final overall confidence score:

[0273] High confidence results: When the confidence level is higher than the set reliability threshold (e.g., 0.7), the system directly outputs the detection result and attaches the confidence level, indicating that the result is reliable.

[0274] Low confidence results: When the confidence level is in the middle range (e.g., 0.5 to 0.7), the system will clearly indicate "low confidence level, retesting is recommended" while outputting the results, and will provide feedback on possible reasons (such as "secondary chromatography occurred", "sample suspected to be non-whole blood").

[0275] Invalid result: When the confidence level is lower than the minimum acceptable threshold (e.g., 0.5), the system determines that the test result is invalid, does not output a specific value, and prompts "Invalid test". At the same time, it provides the directional reason for the failure (such as "abnormal sample addition", "abnormal test strip position" or "signal data quality is too low") to guide the user to the next step.

[0276] Through the above steps, the system can intelligently process complex photoelectric signal data, identify and correct anomalies, select appropriate value locations, and calculate reliable detection results based on rich historical data. This process not only improves the accuracy of the detection results but also provides users with a clear operating guide to ensure that subsequent detection work is more accurate and efficient.

[0277] The aforementioned multi-algorithm integrated analysis method for detecting test strip anomalies acquires data from lines B and C of the test strip and compares the difference in the average values ​​at the ends of these two lines to determine whether line C anomalies exist, thus determining the validity of the detection results. For non-line C anomalies, it further analyzes changes in photoelectric signals caused by incorrect sample addition order or excessive sample addition, extracting and verifying abnormal intervals to identify abnormal regions. Next, the identified abnormal regions are repaired to restore the normal curve shape, yielding a corrected result. Based on this corrected result, a value window that does not intersect with any abnormal regions is determined to ensure the accuracy of the value location calculation. Finally, a model is constructed based on the abnormal regions, corrected results, and value locations and matched with historical cases to calculate the final detection result and its confidence level. This method not only effectively identifies various types of anomalies but also dynamically optimizes the value location to improve data confidence, significantly enhancing the accuracy and reliability of biomedical test strip detection results. It overcomes the shortcomings of existing technologies, such as low accuracy, high false alarm rate, and limited data repair capabilities, when dealing with issues like positional shifts and user operational errors during biomedical test strip detection.

[0278] Figure 2This is a schematic block diagram of a multi-algorithm integrated analysis system 300 for detecting abnormalities in test strips, provided in an embodiment of the present invention. Figure 2 As shown, corresponding to the above-described multi-algorithm comprehensive analysis method for detecting test strip anomalies, the present invention also provides a multi-algorithm comprehensive analysis system 300 for detecting test strip anomalies. This multi-algorithm comprehensive analysis system 300 includes a unit for executing the above-described multi-algorithm comprehensive analysis method for detecting test strip anomalies, and the system can be configured in a server. Specifically, please refer to... Figure 2 The multi-algorithm integrated analysis system 300 for detecting test strip anomalies includes an acquisition unit 301, a difference calculation unit 302, an invalidation unit 303, an analysis unit 304, a repair unit 305, a position calculation unit 306, and a multi-dimensional analysis unit 307.

[0279] The test strip is divided into several units: an acquisition unit 301, which acquires relevant data of lines B and C of the test strip to be tested; a difference calculation unit 302, which determines whether the test strip has a strip position offset by comparing the average difference at the end of lines B and C based on the relevant data of lines B and C; an invalidation unit 303, which determines the test result as invalid if the test strip has a strip position offset; an analysis unit 304, which analyzes the photoelectric signal changes caused by incorrect or excessive sample addition if the test strip does not have a strip position offset, extracts and verifies abnormal intervals to obtain abnormal regions; a repair unit 305, which repairs the abnormal regions to restore the normal shape of the curve to obtain a correction result; a position calculation unit 306, which determines the value window that does not intersect with all the abnormal regions based on the correction result to obtain the value position calculation result; and a multidimensional analysis unit 307, which constructs a model based on the abnormal regions, the correction result, and the value position calculation result and matches it with historical cases to calculate the final test result and its confidence level.

[0280] In one embodiment, the difference calculation unit 302 includes: an average value calculation subunit, used to calculate the average value of several data points at the end of lines B and C based on the relevant data of lines B and C; and an anomaly determination subunit, used to determine that the test strip has a test strip position offset when the absolute value of the difference between the data of lines B and C at the initial time and the average value of several data points at the end of lines B and C is not greater than a set difference threshold.

[0281] In one embodiment, the analysis unit 304 includes:

[0282] The system comprises the following sub-units: a data acquisition sub-unit for acquiring relevant data from the test strip; a parameter setting sub-unit for setting the abnormal interval set, time span, and threshold key parameters, and initializing the abnormal state to be off; an estimation sub-unit for analyzing the trend and estimated value of each data point in the relevant data of the test strip through a sliding window to determine whether the conditions for enabling abnormal judgment are met; a marking sub-unit for marking the current index as the abnormal starting point and recording the left boundary information when the conditions for enabling abnormal judgment are met, thereby activating the abnormal judgment mode to obtain the abnormal interval; a continuous monitoring sub-unit for continuously monitoring the decline rate, time span, lowest point position, and corresponding minimum value of the abnormal interval to obtain monitoring results; a position adjustment sub-unit for adjusting the end position of the abnormal interval using an angle calculation method; an evaluation sub-unit for evaluating the validity of the abnormal interval based on its data characteristics, and considering the abnormal interval as invalid if it does not meet the standard; and a processing sub-unit for applying specified rules to process cases at the end of the sequence or where the time span is not met, in order to identify and confirm all valid abnormal intervals to obtain the abnormal region.

[0283] In one embodiment, the prediction subunit includes:

[0284] The window creation module is used to create a sliding window for calculating the trend of change using a data point from the relevant data of the test paper to be tested, and to create a sliding window for anomaly judgment using subsequent data points; the ratio calculation module is used to calculate the average value, average increment, and estimated value of the two sliding windows; the recording module is used to enter the anomaly judgment mode and record relevant information based on the relationship between the average increment of the sliding window for calculating the trend of change and the current value and the estimated value, when the relationship meets the conditions and the decrease rate exceeds a set threshold; wherein, the anomaly judgment mode is to set the anomaly state to on; the relevant information includes the left boundary value and the left boundary index.

[0285] In one embodiment, the position adjustment subunit includes:

[0286] The location determination module is used to determine the position of the right boundary based on the monitoring results; the optimization module is used to optimize the position of the right boundary using an angle calculation method to obtain the end position of the abnormal interval.

[0287] In one embodiment, the evaluation subunit includes:

[0288] The validity verification module is used to verify the validity of the end position of the abnormal interval, including checking the descent concentration and calculating whether the descent angle meets a specific threshold; the addition module is used to add the end position of the abnormal interval to the set when the end position of the abnormal interval is valid, otherwise reset the relevant state variables.

[0289] In one embodiment, the correction unit includes:

[0290] The system includes a step size setting subunit for setting the single increment step size to an integer; a length calculation subunit for calculating the length of the abnormal region; an index determination subunit for determining the left and right boundary indices of the abnormal region and recording the left and right boundary values; an increment collection and distribution subunit for calculating the integer increment based on the difference between the left and right boundary values; and a value correction subunit for correcting all values ​​in the abnormal region using LV when the abnormal region is located at the end of the curve or the integer increment meets the requirements. This includes: initializing the left-side repair value and calculating the margin; adjusting the values ​​within the abnormal interval using an arithmetic sequence method when the margin is not 0; setting the values ​​within a specific range to RV if the margin is 0; and fine-tuning the values ​​at specific positions within the abnormal region if there is a margin and it is not 0.

[0291] In one embodiment, the position calculation unit 306 includes:

[0292] The sequence includes a length recording subunit for recording the length of the entire sequence of the corrected results; a traversal subunit for traversing backwards from the last element of the sequence to find a value window; a set building subunit for building a set containing the indices of the most recent elements in each iteration; an intersection check subunit for checking whether the set intersects with any of the abnormal regions; and a valid determination subunit for confirming a valid value window when a window with no intersection with any of the abnormal regions is found.

[0293] In one embodiment, the multidimensional analysis unit 307 includes:

[0294] The quantization subunit is used to analyze and quantify the abnormal distribution, morphology, repair ratio, value location, and sample type based on the abnormal region, the correction result, and the value location calculation result to obtain multi-dimensional features. The model construction subunit is used to construct a comprehensive model based on the multi-dimensional features to describe the overall features of the current detection situation. The approximate model finding subunit is used to find historical case models that meet the similarity requirements of the comprehensive model and calculate the final detection result according to the algorithm and parameters. The content determination subunit is used to determine the content of high-confidence results, low-confidence suggestions, or invalid results based on preset rules and the confidence level of the final detection result.

[0295] It should be noted that those skilled in the art can clearly understand that the specific implementation process of the multi-algorithm integrated analysis system 300 for detecting test strip anomalies and each unit can be referred to the corresponding description in the foregoing method embodiments. For the sake of convenience and brevity, it will not be repeated here.

[0296] The aforementioned multi-algorithm integrated analysis system 300 for detecting test strip anomalies can be implemented as a computer program, which can, for example... Figure 3 It runs on the computer device shown.

[0297] Please see Figure 3 , Figure 3 This is a schematic block diagram of a computer device provided in an embodiment of this application. The computer device 500 can be a server, wherein the server can be a standalone server or a server cluster composed of multiple servers.

[0298] See Figure 3 The computer device 500 includes a processor 502, a memory, and a network interface 505 connected via a system bus 501. The memory may include a non-volatile storage medium 503 and internal memory 504.

[0299] The non-volatile storage medium 503 may store an operating system 5031 and a computer program 5032. The computer program 5032 includes program instructions that, when executed, cause the processor 502 to perform a multi-algorithm integrated analysis method for detecting anomalies in test strips.

[0300] The processor 502 provides computing and control capabilities to support the operation of the entire computer device 500.

[0301] The internal memory 504 provides an environment for the operation of the computer program 5032 in the non-volatile storage medium 503. When the computer program 5032 is executed by the processor 502, the processor 502 can execute a multi-algorithm integrated analysis method for detecting anomalies in test strips.

[0302] This network interface 505 is used for network communication with other devices. Those skilled in the art will understand that... Figure 3 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device 500 to which the present application is applied. The specific computer device 500 may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0303] The processor 502 is used to run the computer program 5032 stored in the memory to implement all the steps of the multi-algorithm integrated analysis method for detecting test strip anomalies.

[0304] It should be understood that in the embodiments of this application, the processor 502 may be a central processing unit (CPU), or it may be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or any conventional processor.

[0305] It will be understood by those skilled in the art that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program includes program instructions and can be stored in a storage medium, which is a computer-readable storage medium. The program instructions are executed by at least one processor in the computer system to implement the process steps of the embodiments of the above methods.

[0306] Therefore, the present invention also provides a storage medium. This storage medium can be a computer-readable storage medium. The storage medium stores a computer program, wherein when executed by a processor, the computer program causes the processor to perform all the steps of the multi-algorithm integrated analysis method for detecting anomalies in test strips.

[0307] The storage medium can be any computer-readable storage medium capable of storing program code, such as a USB flash drive, portable hard drive, read-only memory (ROM), magnetic disk, or optical disk.

[0308] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.

[0309] In the embodiments provided by this invention, it should be understood that the disclosed systems and methods can be implemented in other ways. For example, the system embodiments described above are merely illustrative. For example, the division of each unit is only a logical functional division, and there may be other division methods in actual implementation. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed.

[0310] The steps in the method of this invention can be adjusted, merged, or reduced in order according to actual needs. The units in the system of this invention can be merged, divided, or reduced according to actual needs. Furthermore, the functional units in the various embodiments of this invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0311] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, a terminal, or a network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention.

[0312] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any person skilled in the art can easily conceive of various equivalent modifications or substitutions within the technical scope disclosed in the present invention, and these modifications or substitutions should all be covered within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A multi-algorithm integrated analysis method for detecting anomalies in test strips, characterized in that, include: Acquire relevant data of the B and C lines of the test strip to be tested; the relevant data of the B and C lines refers to the photoelectric signal intensity value or corresponding numerical representation collected from the test strip to be tested. These data are obtained by reading the changes in light intensity reflected or transmitted by different detection lines on the test strip using a specific instrument. The instrument collects the photoelectric signal timing data of the test strip during the step-by-step reaction process. Calculate the average signal strength at each time point and establish a time series; Based on the relevant data of lines B and C, the difference in the average value at the end of lines B and C is compared to determine whether there is a test strip position shift in the test strip to be tested; If the test strip is misaligned, the test result is deemed invalid. If the test strip to be tested does not have a test strip position offset, analyze the changes in photoelectric signal caused by incorrect sample addition sequence or excessive amount, extract and verify the abnormal interval to obtain the abnormal area; The abnormal areas are repaired to restore the normal shape of the curve, thus obtaining the corrected result; Based on the correction results, a value window that does not intersect with all the abnormal regions is determined to obtain the value location calculation result; A model is constructed based on the abnormal region, the correction result, and the value location calculation result, and matched with historical cases to calculate the final detection result and its confidence level. The step of determining whether the test strip has a positional shift by comparing the average difference at the ends of lines B and C based on the relevant data of lines B and C includes: Calculate the average value of several data points at the end of lines B and C based on the relevant data of lines B and C; When the absolute value of the difference between the data of lines B and C at the initial moment and the average value of several data points at the end of lines B and C is not greater than the set difference threshold, it is determined that the test strip has a test strip position offset. The repair of the abnormal region to restore the curve to its normal shape, thereby obtaining the corrected result, includes: Set the single increment step size to an integer; Calculate the length of the abnormal region; Determine the left and right boundary indices of the abnormal region, and record the left and right boundary values ​​of the abnormal region; The integer increment is calculated based on the difference between the left and right boundary values; When the abnormal region is located at the end of the curve or the integer increment meets the requirements, LV is used to correct all values ​​in the abnormal region, including: Initialize the left-side repair value and calculate the remainder; when the remainder is not 0, adjust the value in the abnormal interval using an arithmetic sequence method; if the remainder is 0, set the value in a specific range to RV; if there is a remainder and it is not 0, fine-tune the value at a specific position in the abnormal area. The step of determining the value window that does not intersect with all the abnormal regions based on the correction result, in order to obtain the value position calculation result, includes: Record the length of the entire sequence of the corrected results; Iterate backwards from the last element of the sequence to find the value window; Each iteration builds a set containing the indices of the most recent few elements; Check whether the set intersects with any of the abnormal regions; When a window that does not intersect with any of the aforementioned abnormal regions is found, the window is confirmed as a valid value window.

2. The multi-algorithm integrated analysis method for detecting test strip anomalies according to claim 1, characterized in that, The analysis examines changes in photoelectric signals caused by incorrect or excessive sample addition, extracts and verifies abnormal intervals to obtain abnormal regions, including: Obtain relevant data for the test strip to be tested; Set the abnormal interval set, time span, and threshold key parameters, and initialize the abnormal state to be off; By analyzing the trend and estimated value of each data point in the relevant data of the test paper under test through a sliding window, it is determined whether the conditions for enabling the anomaly detection are met. When the conditions for enabling anomaly detection are met, mark the current index as the start of anomaly and record the left boundary information, activate the anomaly detection mode, and obtain the anomaly interval; Continuously monitor the rate of decline, time span, lowest point location, and corresponding minimum value of the abnormal interval to obtain monitoring results; The end position of the abnormal interval is adjusted using an angle calculation method; The validity of the data is evaluated based on the data characteristics of the abnormal intervals, and the abnormal intervals are considered invalid if they do not meet the criteria. For cases at the end of a sequence or where the time span is not met, specified rules are applied to identify and confirm all valid outlier intervals, thus obtaining the outlier region.

3. The multi-algorithm integrated analysis method for detecting test strip anomalies according to claim 2, characterized in that, The step of analyzing the trend and estimated value of each data point in the relevant data of the test strip under test through a sliding window to determine whether the conditions for enabling anomaly detection are met includes: A sliding window for calculating the trend of change is established using a certain data point from the relevant data of the test paper to be tested, and a sliding window for anomaly judgment is established using the subsequent data points. Calculate the average value, average increment, and estimated value of the two sliding windows; Based on the relationship between the average increment of the sliding window that calculates the trend and the current value and the estimated value, when the relationship meets the conditions and the decline rate exceeds the set threshold, the system enters the anomaly detection mode and records relevant information. The anomaly detection mode is to set the anomaly state to "on". The relevant information includes the left boundary value and the left boundary index.

4. The multi-algorithm integrated analysis method for detecting test strip anomalies according to claim 2, characterized in that, The method of adjusting the end position of the abnormal interval using angle calculation includes: The location of the right boundary is determined based on the monitoring results; The position of the right boundary is optimized using an angle calculation method to obtain the end position of the abnormal interval.

5. The multi-algorithm integrated analysis method for detecting test strip anomalies according to claim 2, characterized in that, The step of evaluating the validity of data based on the characteristics of the abnormal intervals, whereby the abnormal intervals are considered invalid if they do not meet the criteria, includes: The validity of the end position of the abnormal interval is verified, including checking the descent concentration and calculating whether the descent angle meets a specific threshold. If the end position of the abnormal interval is valid, add the end position of the abnormal interval to the set; otherwise, reset the relevant state variables.

6. The multi-algorithm integrated analysis method for detecting test strip anomalies according to claim 1, characterized in that, The step of constructing a model based on the abnormal region, the correction result, and the value location calculation result, and matching it with historical cases to calculate the final detection result and its confidence level includes: Based on the abnormal region, the correction result, and the value location calculation result, the abnormal distribution, morphology, repair ratio, value location, and sample type are analyzed and quantified to obtain multi-dimensional features; A comprehensive model is constructed based on the aforementioned multi-dimensional features to describe the overall characteristics of the current detection situation; Find historical case models that meet the similarity requirements with the comprehensive model, and calculate the final detection result based on the algorithm and parameters; Based on preset rules and the confidence level of the final detection result, the content of high-confidence results, low-confidence suggestions, or invalid result prompts is determined.

7. A multi-algorithm integrated analysis system for detecting abnormalities in test strips, characterized in that, The system uses the multi-algorithm integrated analysis method for detecting test strip anomalies as described in any one of claims 1 to 6, including: The acquisition unit is used to acquire relevant data of the B line and C line of the test paper to be tested; The difference calculation unit is used to determine whether the test strip has a positional shift by comparing the average difference at the end of the B line and the C line based on the relevant data of the B line and the C line. The invalidation unit is used to determine that the test result is invalid if the test strip under test has a positional offset. The analysis unit is used to analyze the changes in photoelectric signals caused by incorrect sample addition sequence or excessive sample addition if the test strip to be tested does not have a test strip position offset, and to extract and verify the abnormal interval to obtain the abnormal area; The repair unit is used to repair the abnormal area and restore the curve to its normal shape in order to obtain the correction result; The location calculation unit is used to determine the value window that does not intersect with all the abnormal regions based on the correction result, so as to obtain the value location calculation result; The multidimensional analysis unit is used to construct a model based on the abnormal region, the correction result, and the value location calculation result, and match it with historical cases to calculate the final detection result and its confidence level.

Images