Fluorescence image-based autoimmune encephalitis antibody titer identification method and system

By using an image selection model based on fluorescence images and adjusting exposure parameters, the problem of low accuracy and efficiency in encephalitis titer detection using immunofluorescence method was solved, achieving efficient and accurate titer identification.

CN119006765APending Publication Date: 2024-11-22WEST CHINA HOSPITAL SICHUAN UNIV
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202411138219.8
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-19
Publication Date
2024-11-22

AI Technical Summary

Technical Problem

Existing immunofluorescence methods suffer from low accuracy and efficiency in encephalitis titer detection due to image complexity, especially difficulties in identification caused by overlapping exposed images, unclear fluorescence features, and differences in exposure levels.

Method used

An image selection model based on fluorescence images is adopted. By adjusting exposure parameters and identifying the optimality, the preferred fluorescence image is intelligently determined. The noise, stripes and background brightness fluctuation features are identified. A two-dimensional coordinate system is constructed to draw the exposure optimality curve, find the maximum optimality data, and determine the optimal exposure parameters to obtain accurate titer results.

Benefits of technology

It improves the efficiency and accuracy of encephalitis titer identification, enabling rapid and accurate determination of encephalitis titer results.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119006765A_ABST
    Figure CN119006765A_ABST
Patent Text Reader

Abstract

The invention provides an autoimmune encephalitis antibody titer recognition method and system based on a fluorescence image, belongs to the field of encephalitis titer auxiliary diagnosis, and is used for solving the problem that the efficiency and accuracy of encephalitis titer recognition are poor in the prior art. The initial image of the detection sample and the constructed image selection model are utilized to efficiently and intelligently determine the preferable fluorescence image beneficial to accurate identification of the encephalitis titer, and then the preferable fluorescence image is utilized to determine titer result data, so that the efficiency and accuracy of encephalitis titer identification are improved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present application relates to the field of encephalitis titer-assisted diagnosis, and in particular to a method and system for identifying autoimmune encephalitis antibody titer based on fluorescence images. Background Art

[0002] Immunofluorescence is a typical method for detecting encephalitis titers. When using immunofluorescence to detect encephalitis titers, the dilution multiple of the test sample is generally estimated based on the doctor's experience, and then the test sample is used for antigen-antibody binding, and finally the immunofluorescence method is used to complete the fluorescence effect in the monitoring sample of antigen-antibody binding.

[0003] In one specific method, the fluorescence effect is specifically manifested on the cell membrane. When actually identifying the titer of encephalitis, it is necessary to analyze the number of fluorescent features (cell membrane) contained in the immunofluorescence image in order to determine the number of antigen antibodies and then determine the titer result of the test sample.

[0004] Since there are many complex situations in the immunofluorescence images of encephalitis titers, such as overlapping exposure images of different fluorescence features, incomplete antigen-antibody binding resulting in the fluorescence features not presenting a clear cell membrane shape, and obvious differences in the observation effects of fluorescence images under different exposure levels, the current algorithms for identifying encephalitis titers based on immunofluorescence images have low accuracy and efficiency. Summary of the invention

[0005] The present application provides a method and system for identifying autoimmune encephalitis antibody titers based on fluorescence images, which is beneficial to improving the efficiency and accuracy of encephalitis titer identification.

[0006] In a first aspect, the present application provides a method for identifying the titer of autoimmune encephalitis antibodies based on fluorescence images. The method comprises: Acquire an initial fluorescence image of the detection sample; Inputting the initial fluorescence image into a pre-constructed image selection model to obtain a preferred fluorescence image; The preferred fluorescent image is identified to obtain the titer result data.

[0007] By adopting the above technical solution, the preferred fluorescent image can be intelligently determined based on the initial fluorescent image and the pre-constructed image selection model, and the titer result data of the encephalitis titer obtained by using the preferred fluorescent image is beneficial to improving the efficiency and accuracy of encephalitis titer identification.

[0008] Furthermore, the image selection model includes: Get the input image; Performing a plurality of groups of exposure parameter adjustments on the input image to obtain a plurality of groups of exposure adjustment images, wherein the exposure parameters of a group of the exposure adjustment images are symmetrical with respect to the exposure parameters of the input image; Substituting the input image and the exposure-adjusted image into a pre-built preference recognition model to obtain preference data of the input image and the exposure-adjusted image; A two-dimensional coordinate system is constructed with the exposure parameter as the horizontal axis and the optimization degree data as the vertical axis, and an exposure optimization degree curve is drawn; identifying maximum preference data in the exposure preference curve; Determining whether the exposure parameter corresponding to the maximum preference data is an exposure parameter endpoint value; If yes, re-input the exposure-adjusted image corresponding to the maximum preference data as the input image; If not, a plurality of preferred exposure parameters are determined according to the exposure parameters corresponding to the maximum preference data, and a preferred fluorescent image is determined as output relative to each preferred exposure parameter.

[0009] Further, the step of adjusting the input image by several groups of exposure parameters to obtain several groups of exposure adjusted images includes: Determine multiple groups of exposure adjustment parameters based on the pre-acquired unit exposure parameters, where the exposure adjustment parameters are integer multiples of the unit exposure parameters, and a group of exposure adjustment parameters has two parameters that are opposite to each other; An exposure adjustment image is determined according to each exposure adjustment parameter.

[0010] Furthermore, the preference identification model includes: Identify noise features, stripe features, background brightness fluctuation features, and fluorescence features in the image to be identified; Analyze and determine the feature quantity, maximum feature area and cumulative feature area of ​​noise features and stripe features, as well as the maximum background brightness and minimum background brightness of background brightness fluctuation features, and the cumulative feature area of ​​fluorescence features; The preference data of the image to be identified is determined based on the feature quantity, maximum feature area and cumulative feature area of ​​the noise feature and stripe feature, the highest background brightness and the lowest background brightness of the background brightness fluctuation feature, and the cumulative feature area of ​​the fluorescence feature.

[0011] Furthermore, the determining of the preference data of the image to be identified according to the feature quantity and feature pixel area of ​​the noise feature and the stripe feature, the maximum background brightness and the minimum background brightness of the background brightness fluctuation feature, and the feature pixel area of ​​the fluorescence feature includes: ; ; ; ; In the formula, z is the preferred data, m is the number of noise features, is the characteristic pixel area of ​​all noise features, is the maximum characteristic area of ​​the noise feature, n is the number of stripe features, is the characteristic pixel area of ​​all fringe features, is the maximum characteristic area of ​​the fringe feature, is the maximum background brightness, is the minimum background brightness, is the characteristic pixel area of ​​all fluorescent features, is the preset area value, All are preset constants greater than zero.

[0012] Further, the determining of a plurality of preferred exposure parameters according to the exposure parameters corresponding to the maximum preference data includes: Determining a preset number of exposure parameter adjustment amounts based on a pre-acquired exposure parameter constant, each of the exposure parameter adjustment amounts being a preset integer multiple of the exposure parameter constant; A preferred exposure parameter is determined according to the exposure parameter corresponding to the maximum preference data and each exposure parameter adjustment amount, and the preferred exposure parameter is equal to the exposure parameter of the maximum preference data minus the exposure parameter adjustment amount.

[0013] Further, the identifying the preferred fluorescent image and obtaining the titer result data comprises: Identify each of the preferred fluorescence images and obtain an image fluorescence result respectively; The maximum value, minimum value, median value, average value or interquartile range of all image fluorescence results are determined as the titer result data.

[0014] In the second aspect, the present application provides a system for identifying the titer of autoimmune encephalitis antibodies based on fluorescence images. The system includes a fluorescence image acquisition device and a titer result determination device, wherein the fluorescence image acquisition device is used to acquire the initial fluorescence image, and the titer result determination device is used to execute any one of the methods described in the first aspect above.

[0015] In summary, this application at least has the following beneficial effects: Provided is a method and system for identifying the titer of autoimmune encephalitis antibodies based on fluorescence images, which can intelligently determine the preferred fluorescence image by using the initial fluorescence image and a pre-constructed image selection model, thereby being able to efficiently and accurately determine the titer result data of the encephalitis titer.

[0016] It should be understood that the contents described in the Summary of the Invention are not intended to limit the key or important features of the embodiments of the present application, nor are they intended to limit the scope of the present application. Other features of the present application will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] The above and other features, advantages and aspects of the embodiments of the present application will become more apparent with reference to the following detailed description in conjunction with the accompanying drawings. In the accompanying drawings, the same or similar reference numerals represent the same or similar elements, wherein: Figure 1 A block diagram of a fluorescence image-based autoimmune encephalitis antibody titer identification system in an embodiment of the present application is shown; Figure 2 A flow chart of a method for identifying autoimmune encephalitis antibody titer based on fluorescence images in an embodiment of the present application is shown. DETAILED DESCRIPTION

[0018] In order to make the purpose, technical solution and advantages of the embodiments of the present application clearer, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0019] In addition, the term "and / or" in this article is only a description of the association relationship between the associated objects, indicating that there can be three relationships. For example, A and / or B can represent: A exists alone, A and B exist at the same time, and B exists alone. In addition, the character " / " in this article generally indicates that the associated objects before and after are in an "or" relationship.

[0020] The present application provides a method and system for identifying autoimmune encephalitis antibody titer based on fluorescence images, which can use an image selection model to quickly optimize the initial fluorescence image and determine the preferred fluorescence image, thereby facilitating the efficient and accurate determination of the titer result data of the encephalitis titer.

[0021] In a first aspect, the present application provides an autoimmune encephalitis antibody titer identification system based on fluorescence images.

[0022] Figure 1 A block diagram of a fluorescence image-based autoimmune encephalitis antibody titer identification system in an embodiment of the present application is shown.

[0023] Reference Figure 1 The system includes a fluorescence image acquisition device 110 and a titer result determination device 120.

[0024] The fluorescent image acquisition device 110 is used to acquire the initial fluorescent image. The fluorescent image acquisition device 110 is used to acquire the fluorescent image of the detection sample bound by the antigen and antibody, and the obtained image is the initial fluorescent image. The fluorescent image acquisition device 110 itself can set an exposure rate. Here, the exposure rate of the fluorescent image acquisition device 110 can be manually configured based on data experience according to the attribute information (such as category, concentration, etc.) of the detection sample, and the initial fluorescent image is acquired at this exposure rate.

[0025] The titer result determination device 120 is communicatively connected to the fluorescence image acquisition device 110 to acquire the initial fluorescence image. The titer result determination device 120 is used to determine the titer result data of the test sample according to the initial fluorescence image. This process is implemented based on the intelligent algorithm configured in the titer result determination device 120. The intelligent algorithm is introduced based on the disclosure of the second aspect of the present application.

[0026] In a second aspect, the present application provides a method for identifying the titer of autoimmune encephalitis antibodies based on fluorescence images. The method can be Figure 1 The titer result determination device 120 is executed.

[0027] Figure 2 A flow chart of a method for identifying autoimmune encephalitis antibody titer based on fluorescence images in an embodiment of the present application is shown.

[0028] Reference Figure 2 , the method specifically comprises the following steps: S210: Acquire an initial fluorescence image of the detection sample.

[0029] The titer result determination device 120 is communicatively connected to the fluorescence image acquisition device 110 , so as to acquire the initial fluorescence image collected by the fluorescence image acquisition device 110 and determine the actual exposure parameters of the initial fluorescence image.

[0030] S220: Inputting the initial fluorescence image into a pre-constructed image selection model to obtain a preferred fluorescence image.

[0031] The method of this step is mainly implemented by an image selection model. The image selection model is used to automatically optimize the exposure adjustment of the initial fluorescent image, in order to find an exposure image that can more accurately determine the encephalitis titer with higher efficiency.

[0032] In the method of this step, the image selection model includes: obtaining an input image; performing several groups of exposure parameter adjustments on the input image to obtain several groups of exposure adjustment images, wherein the exposure parameters of one group of the exposure adjustment images are symmetrical about the exposure parameters of the input image; substituting the input image and the exposure adjustment image into a pre-constructed preference recognition model to obtain the preference data of the input image and the exposure adjustment image; constructing a two-dimensional coordinate system with the exposure parameter as the horizontal axis and the preference data as the vertical axis, and drawing an exposure preference curve; identifying the maximum preference data in the exposure preference curve; determining whether the exposure parameter corresponding to the maximum preference data is an exposure parameter endpoint value; if so, re-inputting the exposure adjustment image corresponding to the maximum preference data as the input image; if not, determining several preferred exposure parameters according to the exposure parameter corresponding to the maximum preference data, and determining a preferred fluorescent image as output for each preferred exposure parameter.

[0033] Specifically, the process of obtaining the exposure adjustment image utilizes a preset image exposure adjustment model. The image exposure adjustment model can be an adjustment model pre-trained for immunofluorescence images. For example, a large number of immunofluorescence images of test samples at different exposures are used as training samples, so that the changes in immunofluorescence images when the exposure changes can be determined. The image exposure adjustment model can also be configured as an existing model (such as PS software, etc.). In short, the exposure adjustment model can adjust the exposure parameters of the input image to obtain an immunofluorescence image with the expected adjusted exposure parameters, or it can be explained that the image exposure adjustment model is essentially a fitting model of exposure and image effect. Under a known exposure and immunofluorescence image (initial image effect), another target exposure is input to determine the expected image effect of the immunofluorescence image at the target exposure. The expected image effect can be considered to be the same as the image effect of the immunofluorescence image actually collected at the target exposure. Based on this principle, various software processing can be performed on the actual exposure of the initial fluorescence image to obtain the expected fluorescence image at any target exposure.

[0034] Based on the foregoing, in a specific image selection model, the step of performing several sets of exposure parameter adjustments on the input image to obtain several sets of exposure adjusted images specifically includes: determining multiple sets of exposure adjustment parameters based on pre-acquired unit exposure parameters, the exposure adjustment parameters are integer multiples of the unit exposure parameters, and a set of exposure adjustment parameters has two and are mutually reciprocal; and determining one of the exposure adjustment images according to each exposure adjustment parameter.

[0035] In the initial state, the input image is the initial fluorescence image. The actual exposure of the initial fluorescence image is known. Suppose the actual exposure of the initial fluorescence image is , the unit exposure parameter is , multiple groups of exposure adjustment parameters can be determined, and a group of two exposure adjustment parameters are and , a is a positive integer. In the embodiment of the present application, a is 1, 2, or 3, that is, the exposure adjustment parameter is , these exposure adjustment parameters are substituted into the image exposure adjustment model to adjust the initial fluorescence image, and an exposure adjustment image can be determined relative to each exposure adjustment parameter.

[0036] The preference recognition model is used to determine the preference data of the image to be recognized. In the method of this step, the preference recognition model includes: identifying noise features, stripe features, background brightness fluctuation features and fluorescence features in the image to be recognized; analyzing and determining the feature quantity, maximum feature area and cumulative feature area of ​​the noise features and stripe features, as well as the maximum background brightness and minimum background brightness of the background brightness fluctuation feature, and the cumulative feature area of ​​the fluorescence feature; determining the preference data of the image to be recognized based on the feature quantity, maximum feature area and cumulative feature area of ​​the noise features and stripe features, as well as the maximum background brightness and minimum background brightness of the background brightness fluctuation feature, and the cumulative feature area of ​​the fluorescence feature.

[0037] The image to be identified is an immunofluorescence image of encephalitis titer, and its fluorescence feature (target feature) is a partial or complete cell membrane feature presented after the antigen-antibody binds, showing an obvious complete or partial cell membrane state, and the other parts are all background, which will theoretically be significantly darker than the fluorescence feature. Abnormal features will inevitably appear in the image to be identified, such as noise features, background brightness fluctuation features and stripe features. The noise feature is an obvious noise point in the image to be identified, such as random points or spots that are obviously abnormally too bright or too dark in the image. The stripe feature may be caused by an abnormal failure or influence of the fluorescence image acquisition device 110, resulting in obvious abnormal stripe features in the image to be identified. The background brightness fluctuation feature is generally caused by the instability of the light source of the fluorescence image acquisition device 110, scattering in the optical system or uneven detection of the sample, etc., which is manifested as obvious differences in the background brightness of different parts of the image to be identified.

[0038] Based on pre-training or existing image feature recognition algorithms, the fluorescence features, noise features, background brightness fluctuation features and stripe features in the image to be identified can be identified. Pre-training can be performed by using artificially annotated immunofluorescence images containing fluorescence features, noise features, background brightness fluctuation features and stripe features as training samples. Of course, existing intelligent image feature recognition algorithms can also be used directly. The image feature recognition algorithm can identify the characteristic pixel area of ​​each fluorescence feature, noise feature or stripe feature. It is generally believed that the same type of identification features with continuous characteristic pixels (for example, all noise features or all stripe features) is an identification feature. The characteristic pixel area is represented by the number of pixels surrounded by the minimum closed curve that encloses all characteristic pixels of the identification feature. Similarly, based on pre-training, the image feature recognition algorithm can identify the background part of the image to be identified, so that the brightness of each pixel in the background part can be determined, that is, the highest background brightness and the lowest background brightness in the background brightness fluctuation feature can be determined.

[0039] Based on the foregoing, the preferred data of the image to be identified is determined according to the feature quantity and feature pixel area of ​​the noise feature and the stripe feature, the highest background brightness and the lowest background brightness of the background brightness fluctuation feature, and the feature pixel area of ​​the fluorescence feature, including: ; ; ; ; In the formula, z is the preferred data, m is the number of noise features, is the characteristic pixel area of ​​all noise features, is the maximum characteristic area of ​​the noise feature, n is the number of stripe features, is the characteristic pixel area of ​​all fringe features, is the maximum characteristic area of ​​the fringe feature, is the maximum background brightness, is the minimum background brightness, is the characteristic pixel area of ​​all fluorescent features, is the preset area value, All are preset constants greater than zero.

[0040] In this way, the preference data of each image to be recognized can be determined. The preference data of each image to be recognized can be determined by substituting the input image and the exposure adjusted image into the preference recognition model one by one.

[0041] In the image selection model, after determining the exposure parameters and preference data of the input image and each exposure adjustment image, a two-dimensional coordinate system can be constructed with the exposure parameters as the horizontal axis and the preference data as the vertical axis, and all coordinate points can be fitted with a continuous smooth curve to obtain an exposure preference curve.

[0042] Consider that there must be an exposure parameter that is most conducive to the identification of encephalitis titer, that is, there is only one exposure parameter that makes the preference data of the corresponding image the largest, and the preference data of the images corresponding to other exposure parameters should be lower than the optimal exposure parameter, and the farther away from the optimal exposure parameter, the lower the preference data will be in theory. In other words, the exposure preference curve has only one maximum value, which is also the maximum value. The image corresponding to the maximum exposure parameter is theoretically the image that is most conducive to the identification of encephalitis titer.

[0043] In this way, the optimal exposure parameters can be found by using the circular logic. After determining multiple adjacent exposure parameters (i.e., the exposure parameters of the input image and the exposure parameters of the corresponding multiple groups of exposure adjustment images), the change trend of the preference data corresponding to these multiple exposure parameters can be determined. If the preference data under these multiple exposure parameters change monotonically, it means that the maximum data does not exist between these multiple exposure parameters. At this time, the image selection model is continued to be executed in the direction of the maximum preference data within these multiple exposure parameters, that is, the optimal exposure parameters are continued to be found. On the contrary, if the preference data under these multiple exposure parameters change non-monotonically, it means that the maximum data exists between multiple exposure parameters, indicating that the optimal exposure parameters have been found. The fluorescence image under the optimal exposure parameters can be used as the optimal preferred fluorescence image, and a number of preferred fluorescence images are correspondingly determined as output.

[0044] In one example, the preferred fluorescent image may be determined as only one fluorescent image under optimal exposure parameters.

[0045] In another example, determining a number of preferred exposure parameters based on the exposure parameters corresponding to the maximum preference data includes: determining a preset number of exposure parameter adjustments based on a pre-acquired exposure parameter constant, each of the exposure parameter adjustment amounts being a preset integer multiple of the exposure parameter constant; determining a preferred exposure parameter based on the exposure parameter corresponding to the maximum preference data and each exposure parameter adjustment amount, the preferred exposure parameter being equal to the exposure parameter of the maximum preference data minus the exposure parameter adjustment amount.

[0046] Specifically, the exposure parameters of the optimal fluorescent image are assumed to be , the unit exposure parameter is , multiple exposure parameter constants can be determined. The exposure parameter constants are , b is an integer, then the corresponding preferred exposure parameter is In this example, b can be -2, -1, 0, 1, or 2. Substituting the preferred exposure parameters into the aforementioned image exposure adjustment model, the corresponding preferred fluorescent image can be determined.

[0047] In this way, all preferred fluorescence images can be determined.

[0048] In the method of this step, in order to determine the optimal exposure parameters more quickly, all previous exposure parameters and preference data can be retained during the cyclic optimization process of the image selection model for use in determining the exposure preference curve next time, so as to determine the optimal exposure parameters more quickly and accurately.

[0049] S230: Identify the preferred fluorescent image and obtain the titer result data.

[0050] The method of this step includes: identifying each of the preferred fluorescence images to obtain an image fluorescence result; determining the maximum value, minimum value, median value, mean value or interquartile range of all image fluorescence results as the titer result data. The method of this step can also use pre-trained or existing image feature recognition algorithms to identify and determine titer result data, which will not be repeated.

[0051] According to each preferred fluorescent image, an image fluorescence result can be identified and obtained, and the titer result data can be calculated according to all the image fluorescence results. In the embodiment of the present application, the titer result data is the average value of all the fluorescence result data.

[0052] In summary, this method can realize the identification of encephalitis titer based on the initial fluorescence image, and intelligently process the initial fluorescence image in a reasonable manner, which is conducive to more accurate determination of encephalitis titer and efficient search for image effects that can accurately determine the encephalitis titer, thereby realizing efficient and accurate identification of encephalitis titer.

[0053] It should be understood that the method is essentially a machine learning model, and the parameters in the machine learning model can be trained based on the accuracy of the artificial composite titer result data and the relationship between the exposure parameters. The trainable parameters include unit exposure parameters, the number of exposure adjustment parameter groups, , exposure parameter constants, the number of exposure parameter adjustments and the specific multiples of the exposure parameter constants used, the method of calculating the titer result data using the image fluorescence results, etc. During long-term use, the parameters of the machine learning model are trained, so that the method can determine the titer result data more efficiently and accurately.

[0054] It should be noted that, for the aforementioned method embodiments, for the sake of simplicity of description, they are all expressed as a series of action combinations, but those skilled in the art should be aware that the present application is not limited by the described order of actions, because according to the embodiments of the present application, certain steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also be aware that the embodiments described in the specification are all optional embodiments, and the actions and modules involved are not necessarily required by the present application.

[0055] In summary, this application at least has the following beneficial effects: Provided is a method and system for identifying the titer of autoimmune encephalitis antibodies based on fluorescence images, which can intelligently determine the preferred fluorescence image by using the initial fluorescence image and a pre-constructed image selection model, thereby being able to efficiently and accurately determine the titer result data of the encephalitis titer.

[0056] The above description is only a preferred embodiment of the present application and an explanation of the technical principles used. Those skilled in the art should understand that the scope of disclosure involved in the present application is not limited to the technical solution formed by a specific combination of the above technical features, but should also cover other technical solutions formed by any combination of the above technical features or their equivalent features without departing from the aforementioned disclosed concept. For example, the above features are replaced with the technical features with similar functions disclosed in this application (but not limited to) by each other to form a technical solution.

Claims

1. A method for identifying autoimmune encephalitis antibody titer based on fluorescence images, characterized in that: include: Acquire an initial fluorescence image of the detection sample; Inputting the initial fluorescence image into a pre-constructed image selection model to obtain a preferred fluorescence image; The preferred fluorescent image is identified to obtain the titer result data.

2. The method according to claim 1, characterized in that The image selection model includes: Get the input image; Performing a plurality of groups of exposure parameter adjustments on the input image to obtain a plurality of groups of exposure adjustment images, wherein the exposure parameters of a group of the exposure adjustment images are symmetrical with respect to the exposure parameters of the input image; Substituting the input image and the exposure-adjusted image into a pre-built preference recognition model to obtain preference data of the input image and the exposure-adjusted image; A two-dimensional coordinate system is constructed with the exposure parameter as the horizontal axis and the optimization degree data as the vertical axis, and an exposure optimization degree curve is drawn; identifying maximum preference data in the exposure preference curve; Determining whether the exposure parameter corresponding to the maximum preference data is an exposure parameter endpoint value; If yes, re-input the exposure-adjusted image corresponding to the maximum preference data as the input image; If not, a plurality of preferred exposure parameters are determined according to the exposure parameters corresponding to the maximum preferred degree data, and a preferred fluorescent image is determined as output relative to each preferred exposure parameter.

3. The method according to claim 2, characterized in that The step of adjusting the input image by several groups of exposure parameters to obtain several groups of exposure adjusted images comprises: Determine multiple groups of exposure adjustment parameters based on the pre-acquired unit exposure parameters, where the exposure adjustment parameters are integer multiples of the unit exposure parameters, and a group of exposure adjustment parameters has two parameters that are opposite to each other; An exposure adjustment image is determined according to each exposure adjustment parameter.

4. The method according to claim 3, characterized in that The preference recognition model includes: Identify noise features, stripe features, background brightness fluctuation features, and fluorescence features in the image to be identified; Analyze and determine the feature quantity, maximum feature area and cumulative feature area of ​​noise features and stripe features, as well as the maximum background brightness and minimum background brightness of background brightness fluctuation features, and the cumulative feature area of ​​fluorescence features; The preference data of the image to be identified is determined based on the feature quantity, maximum feature area and cumulative feature area of ​​the noise feature and stripe feature, the highest background brightness and the lowest background brightness of the background brightness fluctuation feature, and the cumulative feature area of ​​the fluorescence feature.

5. The method according to claim 4, characterized in that The step of determining the preference data of the image to be identified according to the feature quantity and feature pixel area of ​​the noise feature and the stripe feature, the highest background brightness and the lowest background brightness of the background brightness fluctuation feature, and the feature pixel area of ​​the fluorescence feature comprises: ; ; ; ; In the formula, z is the preferred data, m is the number of noise features, is the characteristic pixel area of ​​all noise features, is the maximum characteristic area of ​​the noise feature, n is the number of stripe features, is the characteristic pixel area of ​​all fringe features, is the maximum characteristic area of ​​the fringe feature, is the maximum background brightness, is the minimum background brightness, is the characteristic pixel area of ​​all fluorescent features, is the preset area value, All are preset constants greater than zero.

6. The method according to claim 2, characterized in that Determining a plurality of preferred exposure parameters according to the exposure parameters corresponding to the maximum preference data comprises: Determining a preset number of exposure parameter adjustment amounts based on a pre-acquired exposure parameter constant, each of the exposure parameter adjustment amounts being a preset integer multiple of the exposure parameter constant; A preferred exposure parameter is determined according to the exposure parameter corresponding to the maximum preference data and each exposure parameter adjustment amount, and the preferred exposure parameter is equal to the exposure parameter of the maximum preference data minus the exposure parameter adjustment amount.

7. The method according to claim 6, characterized in that The identifying the preferred fluorescent image and obtaining the titer result data comprises: Identify each of the preferred fluorescence images and obtain an image fluorescence result respectively; The maximum value, minimum value, median value, average value or interquartile range of all image fluorescence results are determined as the titer result data.

8. A fluorescence image-based autoimmune encephalitis antibody titer recognition system, characterized in that: It comprises a fluorescence image acquisition device (110) and a titer result determination device (120), wherein the fluorescence image acquisition device (110) is used to acquire the initial fluorescence image, and the titer result determination device (120) is used to execute the method according to any one of claims 1 to 7.