Interpretation Method for Gynecological Dry Chemical Test Results

Abstract

The present invention discloses a method for interpreting the results of gynecological dry chemistry tests. Reaction colors for different test results are selected in a two-dimensional color space; a two-dimensional partition map interpretation model is constructed; the coordinate values of the RGB values of the reaction colors for different test results in the two-dimensional color space are substituted into the two-dimensional partition map interpretation model, and the test results are obtained according to the interpretation area where the coordinate values are located. The advantages of the present invention are that based on the principle most favorable for the segmentation of sample experimental identification points, the reaction colors used to represent different test results are autonomously selected, which can improve the accuracy of the interpretation area division and the scalability of the interpretation results of the interpretation model. Based on the principle that the coincidence rate of the consistency between the manual interpretation results and the machine-readable results meets the requirements and the number of specific identification points of the test results in adjacent interpretation areas is the same, the interpretation experience of professional qualification interpreters is quantified into the interpretation method to achieve the optimal division of the interpretation area, overcoming the problems existing in the existing manual interpretation.

Description

Technical Field

[0001] The present invention relates to the field of interpretation research of test results, and in particular to a method for interpreting the results of dry chemical tests in gynecology. Background Art

[0002] At present, the interpretation of the results of dry chemical tests in gynecology mainly relies on manual comparison with a standard colorimetric card to complete the interpretation of test results. However, the method of manual interpretation has the defects of strong subjectivity, being easily affected by personal color discrimination ability and external lighting conditions, long interpretation time and low efficiency. At the same time, long-term manual interpretation is prone to visual fatigue, and it is difficult to accurately determine the color dividing line, which is very likely to cause deviation in the interpretation results.

[0003] In a similar invention patent, the patent with the closest inventive concept, "Method for Interpreting Temperature of Thermochromic Paint" (Authorization Announcement Number: CN 109087311B), proposes a method for interpreting the temperature of a plate according to its color at high temperature, which has the following defects: 1. The temperature interpretation method mentioned in this patent uses a histogram to construct an interpretation model, which is not conducive to manual and simple intuitive completion of the pre-classification task, nor can it integrate manual interpretation experience into the interpretation model. 2. The color of the temperature and the temperature level are strictly consistent and cannot be artificially set. For example, steel at high temperature, extremely high temperature, and ultra-high temperature will show specific red, white, and cyan colors, but these colors are relatively difficult to simply distinguish in a two-dimensional color space, which will affect the interpretation accuracy of the automatic interpretation model for completing the classification task. Summary of the Invention

[0004] The purpose of the present invention is to provide a method for interpreting the results of dry chemical tests in gynecology.

[0005] To achieve the above purpose, the present invention adopts the following technical solutions:

[0006] The method for interpreting the results of dry chemical tests in gynecology according to the present invention includes the following steps:

[0007] S1, select the reaction colors of different test results in a two-dimensional color space;

[0008] S2, construct a two-dimensional partition diagram interpretation model;

[0009] S3, substitute the coordinate values of the RGB values of the reaction colors of different test results in the two-dimensional color space into the two-dimensional partition diagram interpretation model, and obtain the test results according to the interpretation area where the coordinate values are located. Based on the principle that is most conducive to the segmentation of sample experimental identification points, the reaction colors used to represent different test results are independently selected, which can improve the accuracy of the division of the interpretation area, thereby improving the interpretation accuracy of the two-dimensional partition diagram interpretation model. At the same time, the independent selection of the reaction colors can improve the scalability of the two-dimensional partition diagram interpretation model.

[0010] Further, in step S1, the different test results include negative, weak positive, and positive.

[0011] Further, in step S1, the reaction colors of the selected different test results need to have a specific trend, aggregation, and discreteness in the two-dimensional color space.

[0012] Further, step S2 includes the following steps:

[0013] S2.1, conduct N0 sampling experiments;

[0014] S2.2, obtain the manual interpretation results of the N0 sampling experiments;

[0015] S2.3, obtain the coordinate values of the RGB values of the reaction colors of the N0 sampling experiments in the two-dimensional color space;

[0016] S2.4, plot the coordinate values in step S2.3 into the two-dimensional color space;

[0017] S2.5, adjust the partition lines drawn in the two-dimensional color space to determine the machine interpretation regions of different test results;

[0018] S2.6, determine the machine interpretation results of the N0 sampling experiments according to the machine interpretation regions where the coordinate values are located;

[0019] S2.7, calculate the consistency compliance rate between the manual interpretation results and the machine interpretation results, and the number of specific identification points in adjacent machine interpretation regions;

[0020] S2.8, if the consistency compliance rate is lower than the threshold or the number of specific identification points in adjacent two machine interpretation regions is different, then execute steps S2.5 to S2.7 until the consistency compliance rate meets the threshold requirement and the number of specific identification points in adjacent two machine interpretation regions is the same;

[0021] S2.9, repeat the N1 sampling experiments, and execute steps S2.2, S2.3, S2.4, S2.6, and S2.7; if the consistency compliance rate meets the threshold requirement and the number of specific identification points in adjacent two machine interpretation regions is the same, end the construction process of the two-dimensional partition map interpretation model;

[0022] S2.10, if the consistency compliance rate is lower than the threshold or the number of specific identification points in adjacent two machine interpretation regions is different, then merge the N0 and N1 sampling experiments into N0 times, and re-execute steps S2.2 to S2.9.

[0023] Further, the manual interpretation results are jointly determined by multiple qualified professionals.

[0024] Further, when obtaining the RGB values of the reaction colors in the sampling experiment, Gaussian blur and mean filtering need to be performed before calculating the RGB values.

[0025] Further, the specific identification points refer to the test results where the manual interpretation results are inconsistent with the machine interpretation results.

[0026] The advantages of the present invention are as follows:

[0027] 1. Based on the principle most favorable for segmenting the sample experiment identification points, the reaction colors used to represent different test results are autonomously selected, which can improve the accuracy of the division of the interpretation area, thereby improving the interpretation accuracy of the two-dimensional partition map interpretation model. At the same time, the autonomous selection of the reaction colors can improve the scalability of the two-dimensional partition map interpretation model.

[0028] 2. Based on the principle that the number of specific identification points for different test results is the same, and using the dividing line to manually divide the interpretation area, the interpretation experience of professional interpreters can be quantified into the interpretation method, making the probabilities of misinterpreting negative as weakly positive, weakly positive as negative, weakly positive as positive, or positive as weakly positive the same for interpreters.

[0029] 3. By adopting the <iteration - statistics - iteration> method, based on the principle that the coincidence rate of the consistency between the manual interpretation results and the machine interpretation results meets the requirements and the number of specific identification points for the test results in adjacent interpretation areas is the same, the optimal division of the interpretation area is achieved.

[0030] 4. After the partition map is delimited, it can be used to quickly, accurately, and continuously interpret the test results of subsequent sample tests for 24 hours, overcoming the problems existing in the existing manual interpretation. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] Figure 1 is a flowchart of the method of the present invention.

[0032] Figure 2 is a schematic diagram of selecting the reaction colors of different test results in the two-dimensional color space of the method of the present invention.

[0033] Figure 3 is a schematic diagram of an eight-well card for collecting sample image data of the method of the present invention.

[0034] Figure 4 is a schematic diagram of the two-dimensional partition map interpretation model of the method of the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0035] The technical solutions in the embodiments of the present invention will be clearly and completely described below. Apparently, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0036] As Figure 1 , the method for interpreting the results of gynecological dry chemistry tests according to the present invention includes the following steps:

[0037] S1. Select the reaction colors of different test results in the two-dimensional color space; where different test results include but are not limited to negative, weakly positive, and positive.

[0038] As Figure 2 shown, the two-dimensional color space is selected as the CIE - Yxy color space, and the color of the test result being negative is the position marked by serial number 1, the weakly positive test result is the position marked by serial number 2, and the positive test result is the position marked by serial number 3, such that the change of the identification colors of different test results has a specific trend and high discreteness along the trend, serving as the basis for delimiting the interpretation boundary line in the two-dimensional coordinate system later.

[0039] S2. Construct a two-dimensional partition map interpretation model, which is divided into the following steps:

[0040] S2.1. Conduct N0 sampling experiments; collect the sample image data of the sampling experiments using an eight-hole card as Figure 3 shown.

[0041] S2.2. Obtain the manual interpretation results of the N0 sampling experiments; the manual interpretation results are determined by the joint interpretation of multiple qualified professionals.

[0042] S2.3. Obtain the coordinate values of the RGB values of the reaction colors of the N0 sampling experiments in the two-dimensional color space; use an image acquisition device such as a camera to collect the reaction color images of the sampling experiments under the condition that the device, environment, and lighting conditions are completely the same, and then after Gaussian blurring, calculate the RGB values of the reaction colors of the sampling experiments using the mean filter algorithm.

[0043] Taking Figure 3 the seventh hole of the eight-hole card shown as an example, first use the Opencv Hough circle detection method to locate the position of the seventh hole, then extract the RGB values of the reaction color of the seventh hole, and convert them to obtain the two-dimensional coordinate values (x, y) in the CIE - Yxy color space. The calculation formula is as follows:

[0044]

[0045] Among them, n11, n12, n12, n21, n22, n23, n31, n32, n33 are determined constants, and R, G, B are the values of the R, G, B channels of the image.

[0046] S2.4, Plot the coordinate values in step S2.3 into a two-dimensional color space; when plotting, different coordinate values can be marked with colors or shapes. As Figure 4 shown, in the two-dimensional color space, negative identification points are marked with blue diamonds, weakly positive identification points are marked with red squares, and positive identification points are marked with yellow triangles.

[0047] S2.5, Adjust the partition line drawn in the two-dimensional color space to determine the machine interpretation regions for different test results;

[0048] The partition line is represented by the formula y = kx + b. By changing the values of k and b, the slope of the partition line and its position in the vertical direction in the two-dimensional coordinate graph can be achieved, thereby adjusting the machine interpretation regions for different test results. As Figure 4 shown, partition line 4 and partition line 5 divide the entire two-dimensional color space into 3 regions, namely the negative region 6, the weakly positive region 7, and the positive region 8. By adjusting the partition line, the ranges of the 3 regions can be adjusted. The entire process of constructing the two-dimensional partition graph interpretation model is the process of determining the values of k and b in the partition line y = kx + b.

[0049] S2.6, Determine the machine interpretation results of the N0 - time sampling experiment according to the machine interpretation region where the coordinate values are located;

[0050] S2.7, Calculate the consistency compliance rate between the manual interpretation result and the machine interpretation result, and the number of specific identification points within adjacent machine interpretation regions;

[0051] The consistency compliance rate includes the perfect compliance rate and the qualitative compliance rate. From the perspective of diagnosis, the qualitative compliance rate needs to be calculated. In the qualitative compliance rate, only the negative and positive are qualitatively determined, and the negative and positive are mutually exclusive relationships. That is, the negative in the test result is qualitatively determined as negative, and the weakly positive, positive, and strongly positive are all qualitatively determined as positive.

[0052] From the perspective of the instrument, the perfect compliance rate needs to be calculated. That is, the complete consistency between the manual interpretation result and the machine interpretation result. For example, if the manual interpretation result is negative and the machine interpretation result is also negative; if the manual interpretation result is weakly positive and the machine interpretation result is also weakly positive, the two are completely consistent.

[0053] The calculation of the consistency compliance rate is as follows:

[0054] Assume that in the manual interpretation results and the machine interpretation results, negative, weakly positive, and positive are represented by 1, 2, and 3 respectively. For the experiment on 10 samples, 10 sample image data of the sampling experiment are collected. First, manually interpret the results of the 10-sample experiment, and the manual interpretation results are 1, 1, 1, 1, 1, 1, 1, 1, 3, 2; then calculate the coordinate values of the 10 sample image data in the two-dimensional color space. According to the machine interpretation regions where the coordinate values are located, determine the machine interpretation results of the 10-sample experiment as 1, 1, 1, 1, 1, 1, 1, 1, 2, 2; among them, 9 of the manual interpretation results and the machine interpretation results are exactly the same, so the perfect match rate in the consistency compliance rate is 9 / 10 = 0.9. The number of test results where the machine interpretation result is different from the manual interpretation result is 1, that is, the number of specific identification points is 1. The manual interpretation result of this specific identification point is weakly positive, and the machine interpretation result is strongly positive. According to the standard of the qualitative compliance rate, the manual interpretation result and the machine interpretation result of this specific identification point are consistent, both are positive, so the qualitative compliance rate is: 10 / 10 = 100%

[0055] S2.8, if the consistency compliance rate is lower than the threshold or the number of specific identification points in adjacent two machine interpretation regions is different, then execute steps S2.5 to S2.7 until the consistency compliance rate meets the threshold requirement and the number of specific identification points in adjacent two machine interpretation regions is the same;

[0056] That is, it is necessary to continuously judge whether the consistency compliance rate is lower than the threshold and whether the number of specific identification points in adjacent two machine interpretation regions is the same. If it is judged that the consistency compliance rate is equal to or higher than the threshold and the number of specific identification points in adjacent two machine interpretation regions is the same, then execute step S2.9, otherwise repeat to execute steps S2.5 to S2.7 until the consistency compliance rate meets the requirements and the number of specific identification points in adjacent two machine interpretation regions is the same;

[0057] S2.9, repeat the sampling experiment N1 times, and execute steps S2.2, S2.3, S2.4, S2.6, and S2.7; if the consistency compliance rate meets the threshold requirement and the number of specific identification points in adjacent two machine interpretation regions is the same, complete the construction process of the two-dimensional partition map model.

[0058] This step is after N0 sampling tests, the initially constructed two-dimensional partition map interpretation model meets the requirements, and then N1 sampling tests are carried out again, where the value of N1 is much larger than N0, to verify the effectiveness of the initially constructed two-dimensional partition map interpretation model. If the consistency compliance rate of the machine interpretation results for the N1 sampling tests in this time meets the threshold requirement and the number of specific identification points in adjacent two machine interpretation regions is the same, complete the construction process of the two-dimensional partition map model.

[0059] S2.10, if the consistency compliance rate in step S2.9 is lower than the threshold or the number of specific identification points in two adjacent machine interpretation regions is different, then combine the N0 - th and N1 - th sampling experiments into the N0 - th sampling, and re - execute steps S2.2 to S2.9;

[0060] That is to say, after initially constructing the two - dimensional partition map interpretation model, a verification process with the number of sampling experiments N1 much larger than the number of sampling experiments N0 must be carried out again until the consistency compliance rate meets the threshold requirement and the number of specific identification points in two adjacent machine interpretation regions is the same, then the construction process of the two - dimensional partition map model ends.

[0061] S3, substitute the coordinate values of the RGB values of the reaction colors of different test results in the two - dimensional color space into the two - dimensional partition map interpretation model, and obtain the test results according to the interpretation region where the coordinate values are located.

[0062] As Figure 4 shown, the dividing line 4 between negative and weak - positive in the constructed two - dimensional partition map interpretation model is: Y = 2X + 3 (Equation 1); the dividing line 5 between weak - positive and positive is Y = 2X + 1 (Equation 2).

[0063] In the sample experiment, the (x, y) values of the 7th well in the CIE - Yxy color space are obtained, and the x values are respectively substituted into the above Equation 1 and Equation 2 to calculate y1 and y2. If y > y1, the test result of the 7th well is determined to be negative; if y < y2, the test result of the 7th well is determined to be positive; if y1 > y > y2, the test result of the 7th well is determined to be weak - positive.

[0064] In the gynecological dry - chemistry sampling test, the reaction color is not only related to the presence of the detected target substance, but also related to the concentration of the detected target substance. That is, the reaction color is a continuous change process. Therefore, near the dividing line, there will be partial overlap of the identification points between negative and weak - positive, and between weak - positive and positive. This is determined by the characteristics of the dry - chemistry detection method and cannot be completely avoided.

[0065] When the present invention divides the machine interpretation region, based on the principle of the same number of specific identification points for different test results and using the dividing line to divide the interpretation region, it can realize quantifying the interpretation experience of professional - qualified interpreters into the interpretation method, making the probabilities of interpreters misjudging negative as weak - positive, weak - positive as negative, weak - positive as positive, or positive as weak - positive the same, thus well solving this problem.

Claims

1. A method for interpreting the results of dry chemical tests in gynecology, characterized in that, Including the following steps: S1. Select the response colors of different test results in a two-dimensional color space; the two-dimensional color space is selected as the CIE—Yxy color space; the selected response colors of different test results should have specific trends, aggregation and discreteness along the trend in the two-dimensional color space; S2. Construct a two-dimensional partition map interpretation model; Step S2 includes the following steps: S2.

1. Conduct N0 sampling experiments; S2.

2. Obtain the manual interpretation results of the N0 sampling experiments; S2.

3. Obtain the coordinate values of the RGB values of the response colors of the N0 sampling experiments in the two-dimensional color space; S2.

4. Plot the coordinate values in step S2.3 into the two-dimensional color space; S2.

5. Adjust the partition lines drawn in the two-dimensional color space to determine the machine interpretation regions of different test results; S2.

6. Determine the machine interpretation results of the N0 sampling experiments according to the machine interpretation regions where the coordinate values are located; S2.

7. Calculate the consistency compliance rate between the manual interpretation results and the machine interpretation results, and the number of specific identification points in adjacent machine interpretation regions; S2.

8. If the consistency compliance rate is lower than the threshold or the number of specific identification points in adjacent two machine interpretation regions is different, then execute steps S2.5 to S2.7 until the consistency compliance rate meets the threshold requirement and the number of specific identification points in adjacent two machine interpretation regions is the same; the specific identification points refer to the test results where the manual interpretation results and the machine interpretation results are inconsistent in adjacent machine interpretation regions; S2.

9. Repeat the N1 sampling experiments, and execute steps S2.2, S2.3, S2.4, S2.6, and S2.7; if the consistency compliance rate meets the threshold requirement and the number of specific identification points in adjacent two machine interpretation regions is the same, end the construction process of the two-dimensional partition map interpretation model; S2.

10. If the consistency compliance rate is lower than the threshold or the number of specific identification points in adjacent two machine interpretation regions is different, then merge the N0 and N1 sampling experiments into N0 times, and re-execute steps S2.2 to S2.9; S3. Substitute the coordinate values of the RGB values of the response colors of different test results in the two-dimensional color space into the two-dimensional partition map interpretation model, and obtain the test results according to the interpretation regions where the coordinate values are located. In step S1, the different test results include negative, weakly positive, and positive.

2. The method for interpreting the results of dry chemical gynecological tests according to claim 1, characterized in that: The manual interpretation results are jointly determined by multiple qualified professionals.

3. The method for interpreting the results of dry chemistry tests for gynecology according to claim 1, characterized in that: When obtaining the RGB values of the response colors of the sampling experiments, Gaussian blur and mean filtering need to be performed before calculating the RGB values.

4. The method for interpreting the results of dry chemical tests for gynecology according to claim 1, characterized in that: ​

Images