Tear film break-up time automatic label generation and version mapping export system

Abstract

The application discloses a tear film break-up time automatic label generation and version mapping export system, and particularly relates to the technical field of medical data processing and ophthalmic examination data management, which comprises the following steps: collecting a polarization sequence and a phase sequence corresponding to the same dry eye examination, reading quality control results, equipment numbers, calibration identifiers, session numbers and source version identifiers corresponding to the same dry eye examination, performing merging according to the same session number, and outputting an examination record; storing a version table, wherein the version table at least comprises break-up time generation methods, heat map generation methods, confidence level generation methods, release writing rules and field corresponding relationships corresponding to each version; generating break-up time labels, heat map labels and confidence level labels through the examination record and binding digital thread identifiers, and then performing qualification review on the three types of labels according to export version items, performing field rewriting when the review is passed, and performing re-generation and then exporting a standard data package when the review is not passed.

Description

Technical Field

[0001] This invention relates to the field of medical data processing and ophthalmic examination data management technology, and more specifically, to a system for automatically generating and exporting tear film breakup time labels and version mappings. Background Technology

[0002] In the process of organizing and migrating tear film breakup time data, the current processing focus is on ensuring that breakup time labels, breakup heatmap labels, and confidence labels formed at different stages can continue to be included in a unified data output structure. In practice, the label results are first generated according to the rules corresponding to the generation. Then, before output, the existing label results are organized into data items of the target version based on the field correspondence, compatible templates, or version conversion relationships, combined with edge computing nodes. Taking the case of an ophthalmology clinic continuously accumulating dry eye examination records and regularly importing historical data into a research training database as an example, the system cannot change the existing acquisition links and existing storage results, nor can it re-verify a large number of historical samples one by one. At the same time, it also requires polarization sequences, phase sequences, quality control results and session records to directly participate in subsequent processing on the basis of existing retention, and to complete the pre-processing in an edge computing environment. However, in this usage scenario, the following phenomenon can be repeatedly verified in practice: although the time-break labels, heatmap labels, and confidence labels in some historical samples can be transcribed according to the data structure of the target version, they are no longer suitable as the same set of published results under the judgment criteria of the target version. This manifests as the three types of labels that could originally be output not being recognized after being transferred to the new version, which should be downgraded, have their coverage rewritten, or be regenerated. The root cause is that the existing processing only completed the migration of the label expression form, without verifying whether the set of labels still has the basis to continue to be valid and output together under the target version before migration. The technical problem this application aims to solve is: how to avoid directly transcribing and outputting tear film breakup time labels, heatmap labels, and confidence labels that are only valid in the original generated version but are no longer suitable for use as results in the same group in the target version during the automatic generation and version mapping export of tear film breakup time labels based on edge computing. Summary of the Invention

[0003] To overcome the aforementioned deficiencies in the prior art, embodiments of the present invention provide an automatic tear film breakup time label generation and version mapping export system. This system generates breakup time labels, heatmap labels, and confidence labels from inspection records and binds them to digital thread identifiers. Then, it performs a qualification review on the three types of labels based on the exported version items. If the review is successful, it performs field rewriting; if the review fails, it performs regeneration and exports a standard data package, thereby solving the problems mentioned in the background art.

[0004] To achieve the above objectives, the present invention provides the following technical solution: an automatic tear film breakup time tag generation and version mapping export system, comprising: The acquisition device is used to acquire the polarization sequence and phase sequence corresponding to the same dry eye examination, and read the quality control results, equipment number, calibration identifier, session number and source version identifier corresponding to the same dry eye examination. It performs merging according to the same session number and outputs the examination record. Version storage device is used to store version table. The version table includes at least the method of generating the break time, the method of generating the heat map, the method of generating the confidence level, the release writing rules and the field correspondence for each version. It performs retrieval according to the source version identifier and the exported version identifier in the inspection record and outputs the source version item and the exported version item. The tag generation module performs fracturing time extraction, heatmap generation, and confidence level generation on the inspection record based on the source version item, and writes the fracturing time, heatmap, and confidence level obtained in the same generation process into the same tag group, and outputs the source tag group; The thread binding module is used to perform corresponding bindings on the inspection record, source version item, and source tag group, generate a numeric thread identifier that runs through the inspection record, source version item, and source tag group, and output the thread binding result.

[0005] In a preferred embodiment, it further includes: The qualification re-verification module is used to call the exported version item, perform export review on the source tag group based on the thread binding result, and output the mapping license when all three checks are true, and output the regeneration instruction when any one check is false.

[0006] In a preferred embodiment, it further includes: The regeneration module is used to call the exported version item to re-execute the rupture time extraction, heatmap generation and confidence generation on the inspection record after receiving the regeneration instruction, and write the regenerated rupture time, heatmap and confidence into the same exported label group, and output the regenerated label group.

[0007] In a preferred embodiment, it further includes: The mapping export module is used to perform field rewriting on the source tag group based on the field correspondence in the exported version item after receiving the mapping permission, and write the rewritten exported tag group with the digital thread identifier, device number, calibration identifier and session number, and output a standard data packet; it is also used to write the regenerated tag group with the digital thread identifier, device number, calibration identifier and session number after receiving the regenerated tag group, and output a standard data packet.

[0008] In a preferred embodiment, the acquisition device further includes reading the acquisition time, truncating overlapping portions, performing merging, and writing to an inspection record: Read the acquisition time corresponding to each frame in the polarization sequence and the acquisition time corresponding to each frame in the phase sequence, perform a comparison between the acquisition time period of the polarization sequence and the acquisition time period of the phase sequence, and extract the data corresponding to the overlapping part when the session number is consistent, and output the polarization sequence segment and the phase sequence segment. While outputting polarization sequence fragments and phase sequence fragments, the system reads the write times corresponding to the quality control results, device number, calibration identifier, and source version identifier. It then performs a comparison between the write times and the overlapping portions. The system merges the quality control results, device number, calibration identifier, session number, and source version identifier with the polarization sequence fragments and phase sequence fragments according to a fixed field order and writes them into the overlapping portions, and outputs the inspection record.

[0009] In a preferred embodiment, the version storage device further includes storing a version table, reading source version identifiers and exported version identifiers, performing a retrieval, and outputting source version items and exported version items. Write the version identifier, the break time generation method identifier, the heat map generation method identifier, the confidence level generation method identifier, the release writing rule identifier, and the field correspondence table number into the same version row. Write each version row into the version table according to the version identifier. In the field correspondence table, write the source field name, the exported field name, and the writing position according to the exported version identifier. Output the version table. After outputting the version table, read the source version identifier and exported version identifier from the inspection record. Perform a line-by-line comparison of the source version identifier in the version table to locate the source version row. Perform a line-by-line comparison of the exported version identifier in the version table to locate the exported version row. Read the fracture moment generation method identifier, heatmap generation method identifier, and confidence generation method identifier from the source version row and output the source version item. Read the release write rule identifier from the exported version row and combine it with the source field name, exported field name, and write position that are consistent with the exported version identifier in the field correspondence table to output the exported version item.

[0010] In a preferred embodiment, the tag generation module further includes performing fracture moment extraction, heatmap generation, confidence level generation, and writing to the source tag group: The polarization sequence, phase sequence, quality control results, and source version item in the inspection record are used as input. The same frame groups are formed by pairing them according to the acquisition time. For each same frame group, the difference position between the current polarization frame and the previous polarization frame, and the difference position between the current phase frame and the previous phase frame are calculated. The intersection calculation is performed with the reserved position in the quality control result. The frames are sorted from front to back according to the acquisition time. The first acquisition time that has a non-empty intersection position set and has a corresponding position with the intersection position set of the next same frame group is taken as the candidate rupture time. For the intersection position corresponding to the candidate rupture time, the same frame group after the candidate rupture time is checked according to the order of acquisition time. The same position that falls into the polarization difference position, phase difference position and retention position is written into the heat map area, and the common position set of the intersection position corresponding to the candidate rupture time and the check position of the next same frame group is used as the target heat map area. Based on the candidate rupture time, target heatmap area, quality control results, and source version, perform sequential verification on each position in the target heatmap area. Write a first-level confidence level when the candidate rupture time, the next co-occurrence frame group, and the next co-occurrence frame group are all written to the reserved position; write a second-level confidence level when the candidate rupture time and the next co-occurrence frame group are written to the reserved position but the next co-occurrence frame group is not written to the reserved position; and write a third-level confidence level when the candidate rupture time is written to the reserved position but the next co-occurrence frame group is not written to the reserved position. Write the candidate rupture time, target heatmap area, and confidence level to the same label group and output the source label group.

[0011] In a preferred embodiment, the thread binding module further includes extracting fields, writing binding strings, generating numeric thread identifiers, and outputting thread binding results. Perform field extraction on the session number, device number, and calibration identifier in the inspection record, the source version identifier in the source version item, and the rupture time, heatmap, and confidence level in the source tag group. Write these fields into the binding string in the fixed field order of session number, device number, calibration identifier, source version identifier, rupture time, heatmap, and confidence level, and output the binding string. Write the field position and field length of each field in the binding string in sequence, and concatenate the field length and field content in the order of field position to generate a numeric thread identifier. Then write the numeric thread identifier into the inspection record, source version item and source tag group respectively, and output the thread binding result.

[0012] In a preferred embodiment, the qualification re-examination module further includes creating a peer review table and outputting a review field group: The digital thread identifier is split into a time bit, a heatmap bit, and a confidence bit. The rupture time, heatmap, and confidence level in the source tag group are written into the time bit, heatmap bit, and confidence bit, respectively, to generate a peer review table. Perform position extraction on the time writing rules, heatmap writing rules and confidence writing rules in the exported version item, compare the position in the same position verification table with the position in the exported version item item by item, write the set of fields with the same position into the verification field group, and output the verification field group.

[0013] In a preferred embodiment, the qualification re-examination module further includes performing peer comparison and outputting the qualification re-examination results: The rupture time, heatmap, and confidence level in the verification field group are sequentially written into the time writing rule, heatmap writing rule, and confidence level writing rule to obtain a positive result. The positive result is then written back into the verification field group in its original position to obtain a reverse result. Perform a peer comparison on the positive result, negative result, and verification field group. Output mapping permission when the three items of rupture time, heatmap, and confidence are consistent. If any item is inconsistent, write the inconsistent item into the regeneration instruction and output the qualification re-examination result.

[0014] In a preferred embodiment, the regeneration module further includes forming a regeneration frame group, performing intersection calculation, performing peer verification, and outputting a regeneration tag group: The regeneration command and the inspection record are paired according to the acquisition time to form a regeneration frame group. The polarization difference position and phase difference position are calculated for each regeneration frame group according to the exported version item, and the intersection calculation is performed with the retained position in the quality control result. The acquisition time that is not empty in the intersection position set and has the same position intersection position with the next regeneration frame group is taken as the regeneration break time according to the acquisition time order. Based on the intersection position corresponding to the regeneration break time, perform a frame-by-frame co-location check on subsequent regeneration frame groups. Write the co-location positions that simultaneously fall into the polarization difference position, phase difference position, and retention position into the regeneration heatmap. Write the retention write results in the regeneration break time, the next regeneration frame group, and the next regeneration frame group into the regeneration confidence score. Then write the regeneration break time, the regeneration heatmap, and the regeneration confidence score into the same exported label group and output the regeneration label group.

[0015] In a preferred embodiment, the mapping export module further includes reading the source field name, the export field name, and the write position, performing field substitution, corresponding writing, and outputting a standard data packet. After receiving the mapping permission, read the source field name, export field name and write position in the export version item, locate the break time, heatmap and confidence in the source tag group according to the source field name, replace the field name according to the export field name and rearrange the field order according to the write position to generate the export tag group. Perform corresponding write operations on the exported tag group or regenerated tag group, write the digital thread identifier, device number, calibration identifier, and session number into the header area in a fixed order, write the exported tag group or regenerated tag group into the data area in the field order, and then perform splicing with the header area first and the data area last to output a standard data packet.

[0016] The technical effects and advantages of this invention are as follows: 1. This solution first verifies the basis for the co-occurrence of the fracture time label, heatmap label and confidence label according to the exported version item, and then decides whether to map or regenerate, thereby relatively suppressing the direct transcribing and output of triple labels that only exist under the original generated version. 2. Write the inspection record, source version item, and source tag group into the same numeric thread identifier so that qualification re-examination, regeneration, and mapping export all retrieve data along the same source chain, thereby relatively improving the field mismatch and back-pointing confusion in the version migration process; 3. Extract the break time, heatmap and confidence level in sequence according to the same frame group, intersection position and same position review, and perform pre-processing in combination with edge computing to make the three types of labels originate from the same generation process, thereby relatively reducing the probability of temporal misalignment and spatial mismatch between labels; 4. When the qualification re-examination fails, the exported version item is called to recalculate the rupture time, heatmap and confidence, instead of using the old label results that failed the review. This helps to make the exported results adapt to the writing rules of the target version. 5. Write the digital thread identifier, device number, calibration identifier, and session number into the packet header area, and write the exported tag group or regenerated tag group into the data area. Combine this with edge computing to complete the standard data packet output, thereby improving the version traceability capability of the standard data packet and the efficiency of subsequent training. Attached Figure Description

[0017] Figure 1 This is a schematic diagram of the system module structure of the present invention. Detailed Implementation

[0018] 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 embodiments of the present invention, and not all embodiments. 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.

[0019] Refer to the instruction manual appendix Figure 1 The tear film breakup time automatic tag generation and version mapping export system of the present invention includes: The acquisition device is used to acquire the polarization sequence and phase sequence corresponding to the same dry eye examination, and read the quality control results, equipment number, calibration identifier, session number and source version identifier corresponding to the same dry eye examination. It performs merging according to the same session number and outputs the examination record. This implementation focuses on the merging and writing of data corresponding to the same dry eye examination by the acquisition device. The purpose is to first extract data from the same session and the same time period from the polarization sequence and the phase sequence, and then write the quality control results, device number, calibration identifier and source version identifier that fall within the same time period into the inspection record, so that subsequent version retrieval, tag generation and thread binding are all based on the same input object. In the process, the common data range of the two types of sequences is first determined, then auxiliary fields are filtered in according to the writing time, and finally the inspection record is formed according to the predetermined field order. After the acquisition device reads the polarization sequence and phase sequence, the acquisition times of each frame are extracted and arranged chronologically. Then, the later value between the acquisition times of the first frame of the polarization sequence and the first frame of the phase sequence is taken as the overlap start time, and the earlier value between the acquisition times of the last frame of the polarization sequence and the last frame of the phase sequence is taken as the overlap end time. Subsequently, the session numbers corresponding to the two types of sequences are verified. Only when the session numbers are the same and the overlap start time is not later than the overlap end time, frames whose acquisition times fall between the overlap start time and the overlap end time are extracted from the polarization sequence to form a polarization sequence segment, and frames whose acquisition times fall between the overlap start time and the overlap end time are extracted from the phase sequence to form a phase sequence segment. At the same time, the overlap start time and overlap end time are written into the segment record for subsequent reading. If the session numbers are different, or the overlap start time is later than the overlap end time, or there are no frames between the overlap start time and the overlap end time in either sequence, an empty segment mark is written and the generation of this inspection record ends. After obtaining the polarization sequence fragments and phase sequence fragments, the acquisition device continues to read the write times corresponding to the quality control results, device number, calibration identifier, and source version identifier, and compares each write time with the overlap start time and overlap end time. Quality control results, device numbers, calibration identifiers, and source version identifiers whose write times fall between the overlap start time and overlap end time and whose session numbers are the same as the corresponding session numbers of the polarization sequence fragments and phase sequence fragments are all included in the merging object. Then, they are written to the inspection record according to the pre-defined field order in the system configuration table. The system configuration table sequentially writes the corresponding field positions of the session number, device number, calibration identifier, source version identifier, polarization sequence fragment, phase sequence fragment, and quality control result. During merging and writing, the fields are written from front to back according to their positions. If any auxiliary field's write time is missing, the session establishment time is read first and then added. If it still cannot be obtained, a missing marker is written and saved along with the inspection record. If the session number corresponding to any auxiliary field is different from the session number in the fragment record, the corresponding field is not written and a mismatch marker is written. In practical applications: After completing a dry eye examination, the outpatient data acquisition terminal first calculates the common data range from the polarization sequence and phase sequence. For example, 10:02:15 is written as the overlap start time and 10:02:27 is written as the overlap end time. Then, polarization sequence segments and phase sequence segments within this time period are extracted. Subsequently, the quality control results, device number, calibration identifier, and source version identifier are read and written to the corresponding times. Fields falling between 10:02:15 and 10:02:27 with the same session number are written to the examination record according to the field order in the configuration table. After this processing, the source version identifier can be directly read when retrieving the version table, the polarization sequence segment, phase sequence segment, and quality control results can be directly read when generating tags, and the session number, device number, and calibration identifier can also be directly read when binding threads.

[0020] Version storage device is used to store version table. The version table includes at least the method of generating the break time, the method of generating the heat map, the method of generating the confidence level, the release writing rules and the field correspondence for each version. It performs retrieval according to the source version identifier and the exported version identifier in the inspection record and outputs the source version item and the exported version item. This implementation focuses on the establishment and retrieval of version tables and field correspondence tables in a version storage device. The aim is to separate the generation method information corresponding to the source version identifier and the write information corresponding to the exported version identifier from the same storage structure, allowing tag generation, qualification re-verification, and mapping export to read their respective required fields. The process first writes the generation method identifier, release write rule identifier, and field correspondence table number corresponding to each version into the version table. Then, based on the source version identifier and exported version identifier in the inspection record, a row-by-row retrieval is performed to retrieve the source version item and exported version item from the version table and field correspondence table, respectively. This implementation process includes the following steps: First, write the version information into row records that can be directly located during subsequent retrieval. The input includes the version identifier, rupture time generation method identifier, heatmap generation method identifier, confidence level generation method identifier, publication write rule identifier, field mapping table number, source field name, exported field name, and write position. The processing action is to write the version identifier, rupture time generation method identifier, heatmap generation method identifier, confidence level generation method identifier, publication write rule identifier, and field mapping table number into the same version row, one for each version identifier. Then, write the version information from front to back into the version table, and add corresponding field information. In the relation table, the exported version identifier, source field name, exported field name, and write position are written to the same corresponding relation row. Among them, the record pointed to by the published write rule identifier must at least write the time field name, heatmap field name, confidence field name, and their respective write positions. The output is a version table and a field correspondence table. The version table is used for version retrieval, and the field correspondence table is used for exported field rewriting. The exception handling is that when multiple records of the same version identifier appear, the first row is taken as the valid version row according to the writing order, and the subsequently written version rows are marked as duplicate version rows. When the field correspondence table number is missing, the writing of the corresponding version row is stopped and a missing mark is written. After obtaining the version table, the source version identifier and exported version identifier in the inspection record are converted into source version item and exported version item, respectively; the input quantities are the source version identifier, exported version identifier, version table and field correspondence table in the inspection record; the processing action is to first read the source version identifier, then compare the version identifiers line by line according to the writing order of the version table, locate the first consistent version line as the source version line, and read the fracture moment generation method identifier, heat map generation method identifier and confidence level generation method identifier from the source version line and write them into the source version item; Subsequently, the exported version identifier is read, and then the version identifiers are compared row by row according to the writing order of the version table. The first consistent version line is located as the exported version line, and the release write rule identifier is read from the exported version line. Then, the source field name, exported field name, and write position are read from the field correspondence table using the exported version identifier as the search key, and then merged and written into the exported version item. The output consists of the source version item and the exported version item. The source version item is used for tag generation and reading, and the exported version item is used for qualification re-examination and mapping export. The exception handling is as follows: if the source version identifier does not locate the source version line, a source version missing mark is written and the output of the source version item is stopped; if the exported version identifier does not locate the exported version line, an exported version missing mark is written and the output of the exported version item is stopped; if no corresponding relationship line matching the exported version identifier is found in the field correspondence table, an exported correspondence missing mark is written and subsequent field replacement is prohibited. In practical applications: The version storage device first writes the fracture time generation method identifier, heatmap generation method identifier, confidence generation method identifier, release write rule identifier, and field correspondence table number corresponding to version V1 into the version table. Then, it writes the information corresponding to version V2 into the next version row, and at the same time, it writes the source field name, export field name, and write position corresponding to V2 into the field correspondence table. When the source version identifier in the inspection record is V1 and the export version identifier is V2, the version storage device first retrieves the three generation method identifiers corresponding to V1 from the version table to form the source version item. Then, it retrieves the release write rule identifier, source field name, export field name, and write position corresponding to V2 from the version table and the field correspondence table to form the export version item. Subsequently, the tag generation reads the source version item and executes the source tag generation. The qualification re-examination and mapping export reads the export version item and executes the export review and field replacement.

[0021] The tag generation module performs fracturing time extraction, heatmap generation, and confidence level generation on the inspection record based on the source version item, and writes the fracturing time, heatmap, and confidence level obtained in the same generation process into the same tag group, and outputs the source tag group; This implementation focuses on the tag generation module's processes for extracting rupture moments, generating heatmaps, and generating confidence scores from inspection records. The aim is to convert polarization sequences, phase sequences, and quality control results into a source tag group corresponding to the same generation process, ensuring that subsequent thread binding, qualification re-examination, and mapping export are all based on the same set of three-way results. The process first forms a frame group based on the acquisition time and determines candidate rupture moments. Then, it retrieves the corresponding positions based on the candidate rupture moments to form the target heatmap area. Finally, it determines the confidence score based on the retention and writing results of the target heatmap area in subsequent frame groups and writes it into the source tag group. This implementation process includes the following steps: First, candidate breakage moments are located from the inspection records. This is to identify the starting moment when polarization change, phase change, and quality control retention occur simultaneously and continue into the next phase. The inputs are the polarization sequence, phase sequence, quality control results, and source version from the inspection records. The processing involves first sorting the polarization sequence in ascending order of acquisition time, then sorting the phase sequence in ascending order of acquisition time. A polarization frame and a phase frame with the same acquisition time are taken to form a co-equal frame group. If there are no co-equal times, no co-equal frame group is formed. Then, for each co-equal frame group, the position where the difference between the pixel values ​​of the current polarization frame and the previous polarization frame at the same row and column number is non-zero is calculated and written to the polarization difference position. The position where the difference between the phase values ​​of the current phase frame and the previous phase frame at the same row and column number is non-zero is calculated and written to the phase difference position. Finally, the position where the retention mark is written from the same acquisition time is read from the quality control results and written to the retention position. Then, the intersection of the polarization difference position, phase difference position, and reserved position is calculated to obtain the intersection position set. The intersection position set is then searched from front to back according to the acquisition time. The first acquisition time that is not empty in the intersection position set and has the same row number and column number as the intersection position set of the next same frame group is written as the candidate break time. At the same time, the corresponding intersection position set is written into the candidate record for subsequent reading. The output is the candidate break time and the candidate record. The exception handling is that if there is no same frame group, the candidate break time is not detected, or the candidate record is empty, a non-detection mark is written and the generation of the source tag group is stopped. Then, the starting intersection positions corresponding to the candidate breakage time are checked backwards, retaining the set of positions that still maintain co-location. The input quantities are candidate breakage time, candidate record, polarization sequence, phase sequence, and quality control results. The processing action is to first read the set of intersection positions corresponding to the candidate breakage time from the candidate record, and then read the polarization difference position, phase difference position, and retained position from the same frame group after the candidate breakage time in the order of acquisition time. For each position in the intersection position set corresponding to the candidate breakage time, a co-location check is performed. All positions in the subsequent same frame group are checked. The positions that fall within the polarization difference position, phase difference position, and reserved position are written into the re-examination position set. Then, the common position calculation is performed on the intersection position set corresponding to the candidate rupture time and the re-examination position set corresponding to the next co-occurrence frame group to obtain the target heatmap area, and it is written into the heatmap area record in the order of row number and column number for subsequent reading. The output is the target heatmap area and the heatmap area record. The exception handling is that if the next co-occurrence frame group does not exist, the intersection position set corresponding to the candidate rupture time is directly written as the target heatmap area. If it is empty, an empty heatmap area mark is written and confidence generation is stopped. Finally, based on the retention and writing results of the target heatmap area in the candidate rupture time, the next co-occurrence frame group, and the next co-occurrence frame group, the rupture time and heatmap area are mapped to the graded confidence results; the input quantities are candidate rupture time, target heatmap area, heatmap area record, quality control result, and source version item; the processing action is to perform sequential verification on each position in the target heatmap area by row number and column number, first read the retention position in the co-occurrence frame group corresponding to the candidate rupture time, then read the retention position in the next co-occurrence frame group, and then read the retention position in the next co-occurrence frame group; When the same location is written to a reserved position at the candidate rupture time, the next co-occurrence frame group, and the subsequent co-occurrence frame group, the first-level identifier in the source version item is written to the confidence field. When the same location is written to a reserved position at the candidate rupture time and the next co-occurrence frame group, but not at the subsequent co-occurrence frame group, the second-level identifier in the source version item is written to the confidence field. When the same location is written to a reserved position only at the candidate rupture time and not at the next co-occurrence frame group, the third-level identifier in the source version item is written to the confidence field. When there are multiple locations in the target heatmap area, the first-level identifier, second-level identifier, and third-level identifier are written as one, two, and three in confidence level order. Then, the smallest value is taken as the overall confidence level. Subsequently, the candidate rupture time, target heatmap area, overall confidence level, and current generation time are written to the same tag group in a fixed field order to form the source tag group. The output is the source tag group. The exception handling is when the target heatmap area is empty or the overall confidence level is not generated, a non-generated confidence level mark is written and the publication of the source tag group is stopped. In practical applications: After the outpatient terminal completes the examination record generation, the tag generation module first forms a same-time frame group from the polarization sequence and phase sequence according to the same acquisition time. Then, it calculates the intersection set of polarization difference position, phase difference position, and reserved position from each same-time frame group. The first acquisition time that can form a corresponding position with the next same-time frame group is taken from the front and written as the candidate rupture time. Subsequently, the intersection set corresponding to the candidate rupture time is checked against the next same-time frame group, and the common position set is written as the target heat map area. Finally, the reserved writing results in the candidate rupture time, the next same-time frame group, and the next next same-time frame group are read one by one for the positions in the target heat map area to generate a first-level confidence level, a second-level confidence level, or a third-level confidence level. The candidate rupture time, the target heat map area, and the confidence level are written into the source tag group. After this processing, the rupture time, heat map area, and confidence level in the source tag group come from the same generation process. When the thread is bound later, the three items can be retrieved directly, and the review can be performed directly by field during qualification re-examination.

[0022] The thread binding module is used to perform corresponding bindings on the inspection record, source version item, and source tag group, generate a numeric thread identifier that runs through the inspection record, source version item, and source tag group, and output the thread binding result. This implementation focuses on the associated writing of inspection records, source version items, and source tag groups by the thread binding module. The aim is to push the data source, version source, and tag results corresponding to the same inspection into the same numerical thread identifier, so that qualification re-examination, regeneration, and mapping export can all retrieve the original fields along the same identifier. During processing, the fields involved in the binding are first extracted from the three types of objects, then written into the binding string in a fixed field order. Subsequently, the field position, field length, and field content are concatenated into a numerical thread identifier, which is then written back to the inspection record, source version item, and source tag group respectively. This implementation process includes the following steps: First, the source field, version source field, and tag field are arranged into the same write sequence to provide a unique input for subsequent digital thread identifier encoding. The input includes the session number, device number, and calibration identifier from the inspection record, the source version identifier from the source version item, and the rupture time, heatmap, and confidence level from the source tag group. The processing action is to first read the session number, device number, and calibration identifier from the inspection record, then read the source version identifier from the source version item, and then read the rupture time, heatmap, and confidence level from the source tag group. The heatmap is expanded into a position string in the order of row number first and column number last. Then, the session number, device number, calibration identifier, source version identifier, rupture time, heatmap, and confidence level are written in the order of the fields pre-written in the system configuration table to form a binding string. The field order is also written to the binding record for subsequent reading. The output is the binding string and the binding record. The exception handling is that if any field is missing, a missing flag is written first and then the binding string is written. If the heatmap is empty, an empty heatmap flag is written. If the field order is missing, thread binding is stopped and a binding failure flag is written. After obtaining the binding string, a numeric thread identifier is generated. The binding string is converted into a fixed encoding structure, and the same encoding result is distributed to the inspection record, the source version item, and the source tag group. The input is the binding string and the binding record. The processing action is to read each field in the binding string from front to back according to the field order in the binding record. First, the field position is written to each field, then the field length is written, where the field length is the number of characters of the corresponding field content. Then the field content is written, and the field position, field length, and field content are concatenated from front to back according to the field position to form the numeric thread identifier. Then, the numeric thread identifier is written to the thread identifier field in the inspection record, the thread identifier field in the source version item, and the thread identifier field in the source tag group to form the thread binding result. The output is the numeric thread identifier and the thread binding result. The exception handling is to stop the generation of the numeric thread identifier when the binding string is empty. If the numeric thread identifier fails to be written back to any object, a write-back failure flag is written and the numeric thread identifiers in the remaining objects are retained for subsequent qualification re-examination to read first according to the objects that have been written back. In practical applications: After a dry eye inspection completes the generation of the source tag group, the thread binding module first extracts the session number, device number, and calibration identifier from the inspection record, extracts the source version identifier from the source version item, and then extracts the rupture time, heatmap, and confidence level from the source tag group. The heatmap is then expanded into a position string and written into a binding string according to a fixed field order. Subsequently, the binding string is encoded into a digital thread identifier according to the field position, field length, and field content. The same digital thread identifier is then written into the inspection record, the source version item, and the source tag group. After this processing, when reading the digital thread identifier for qualification re-examination, the time bit, heatmap bit, and confidence bit can be directly extracted. When reading the digital thread identifier for mapping export, the exported tag group can also be matched with the original inspection source.

[0023] The qualification re-verification module is used to call the exported version item, perform export review on the source tag group based on the thread binding result, and output the mapping license when all three checks are true, and output the regeneration instruction when any one check is false. This implementation focuses on the export review of the source tag group by the qualification re-examination module. The purpose is to first re-attach the break time, heatmap, and confidence level in the source tag group to their corresponding positions in the digital thread identifier. Then, the time writing rules, heatmap writing rules, and confidence level writing rules in the exported version item are introduced into the review process to determine whether the source tag group can still be written according to its original positional relationship under the exported version item. The processing involves first extracting the time bit, heatmap bit, and confidence bit, then forming a review field group, subsequently generating positive and negative results, and finally outputting a mapping permission or regeneration instruction based on the same-position comparison results. This implementation process includes the following steps: First, rewrite the three items from the source tag group back to the corresponding positions of the numeric thread identifier to ensure that subsequent reviews do not deviate from the original binding relationship. The inputs are the numeric thread identifier in the thread binding result and the rupture time, heatmap, and confidence level from the source tag group. The processing actions are as follows: first, split the numeric thread identifier from front to back according to the field position; read the field position corresponding to the rupture time as the time position; read the field position corresponding to the heatmap as the heatmap position; read the field position corresponding to the confidence level as the confidence level; then, write the rupture time from the source tag group into the time position; expand the heatmap in the order of row number first and column number last and write it into the heatmap position; write the confidence level into the confidence level to form a peer review table; and write the time position, heatmap position, and confidence level into the review record for subsequent reading. The outputs are the peer review table and the review record. The exception handling is as follows: if the numeric thread identifier cannot be split into three positions, write a position missing flag and directly generate a regeneration instruction; if any field in the source tag group is missing, write a field missing flag and stop the mapping permission determination. Then, the positional relationships in the peer review table are re-aligned with the write positions in the exported version items to filter out the set of fields to enter the exported review. The inputs are the peer review table, review records, and the time write rules, heatmap write rules, and confidence write rules in the exported version items. The processing actions are as follows: first, extract the time write position from the time write rules, extract the heatmap write position from the heatmap write rules, and extract the confidence write position from the confidence write rules. Then, compare the time position with the time write position in the peer review table item by item, compare the heatmap position with the heatmap write position item by item, and compare the confidence position with the confidence write position item by item. Those with the same position are written into the review field group, and those with different positions are written into an inconsistent record. The inconsistent record is written into the review log for subsequent reading. The outputs are the review field group and the review log. The exception handling is as follows: if any write rule is missing a write position, the write rule is marked as missing and a regeneration instruction is generated directly. If none of the three positions are written into the review field group, an empty review field group is marked and subsequent mapping permission determination is stopped. Next, the review field group is written once in the export direction according to the exported version item, and then once in the retrieve direction according to the original position, so that the export direction and retrieve direction are mutually verified; the input quantities are the time writing rules, heatmap writing rules, and confidence writing rules in the review field group and the exported version item; the processing action is to first write the rupture time in the review field group into the corresponding field of the time writing rule to obtain the time positive result, write the heatmap into the corresponding field of the heatmap writing rule to obtain the heatmap positive result, write the confidence into the corresponding field of the confidence writing rule to obtain the confidence positive result, and then write the time positive result back into the review field group according to the time position to obtain... The time-reverse result is generated by writing back the heatmap forward result to the verification field group according to the heatmap position, and writing back the confidence forward result to the verification field group according to the confidence position, resulting in the confidence reverse result. The time-forward result, heatmap forward result, confidence forward result, time-reverse result, heatmap reverse result, and confidence reverse result are written to the result table for subsequent reading. The output consists of forward result, reverse result, and result table. The exception handling is as follows: if any field is not found in the writing rule, a failure flag is written and forward result generation is stopped; if any reverse result is not written back in the original position, a write-back failure flag is written and the regeneration instruction is initiated. Finally, a peer comparison is used to determine whether the source tag group can still maintain the original field relationship according to the exported version item; the input quantities are positive results, negative results, verification field group, and verification log; the processing actions are to perform peer comparisons between the time-based positive results, time-based negative results, and the rupture time in the verification field group; to perform peer comparisons between the heatmap positive results, heatmap negative results, and the confidence level in the verification field group; if all three comparisons are the same, the digital thread is identified and exported. Version item identifier and mapping permission flag are written to the qualification re-examination result table, and mapping permission is output; if any item is different, the inconsistent field name, original position, written position and field content are written to the regeneration instruction table, the regeneration instruction is output, and the mapping rejection flag is written to the qualification re-examination result table; the output is either mapping permission or regeneration instruction and qualification re-examination result table; if the exception handling is that the review log already has a rule missing flag, position missing flag, write failure flag or writeback failure flag, the mapping permission output is not entered, and the regeneration instruction is generated directly; In practical applications: The qualification re-examination module first splits the digital thread identifier into time bits, heatmap bits, and confidence bits, and then writes the rupture time, heatmap, and confidence level from the source tag group back to the corresponding positions to form a corresponding position review table; then it reads the time writing rules, heatmap writing rules, and confidence level writing rules from the exported version items, extracts the writing positions, and filters out the review field groups; then it completes a write operation according to the exported version items to obtain a positive result, and then completes a write-back operation according to the original positions to obtain a reverse result; finally, it compares the positive result, the reverse result, and the review field groups item by item. If the three items are the same, it outputs a mapping permission; if any item is different, it writes the inconsistent field name, original position, writing position, and field content into the regeneration instruction; after this processing, there is a direct basis for determining whether the source tag group has entered the field rewriting path or the regeneration path, and subsequent mapping export and regeneration can continue to be processed along the qualification re-examination result table.

[0024] The regeneration module is used to call the exported version item to re-execute the rupture time extraction, heatmap generation and confidence generation on the inspection record after receiving the regeneration instruction, and write the regenerated rupture time, heatmap and confidence into the same exported label group, and output the regenerated label group. This implementation focuses on the regeneration of label results after the regeneration module fails the qualification re-examination. The aim is to relocate the inspection records based on the inconsistencies recorded in the regeneration instruction, and then recalculate the rupture time, heatmap, and confidence level according to the generation method corresponding to the exported version item. This ensures that the exported label group no longer uses the source label results that failed the review. The process involves first retrieving the corresponding inspection records based on the digital thread identifier and forming a regeneration frame group. Then, based on the intersection position corresponding to the regenerated rupture time, a corresponding review is performed. Finally, the regenerated rupture time, regenerated heatmap, and regenerated confidence level are written into the same exported label group. This implementation process includes the following steps: First, a regenerated frame group is formed and the regeneration failure time is determined. Its purpose is to pull the fields that failed the qualification re-examination back to the original inspection data and start the recalculation according to the exported version item. The input quantities are regeneration command, inspection record, exported version item and quality control result. The processing action is as follows: first, read the digital thread identifier and inconsistent field name from the regeneration command, then locate the corresponding session number, polarization sequence segment and phase sequence segment in the inspection record with the digital thread identifier, then arrange the polarization sequence segment and phase sequence segment in ascending order according to the acquisition time, and take a polarization frame and a phase frame with the same acquisition time to form a regenerated frame group. Then, for each regenerated frame group, calculate the position where the result of subtracting the pixel value of the current polarization frame from the previous polarization frame at the same row number and column number is not zero, and write it to the polarization difference position. Calculate the position where the result of subtracting the phase value of the current phase frame from the previous phase frame at the same row number and column number is not zero, and write it to the phase difference position. Finally, read the position of the reserved mark written at the same acquisition time from the quality control result and write it to the reserved position. Then, the intersection of the polarization difference position, phase difference position, and retained position is calculated to obtain the intersection position set. The set is then searched from front to back according to the acquisition time. The first acquisition time that is not empty in the intersection position set and has the same row and column number as the intersection position set of the next regenerated frame group is written as the regeneration break time. Simultaneously, the corresponding intersection position set is written into the regeneration record for subsequent reading. The output consists of the regenerated frame group, the regeneration break time, and the regeneration record. Anomaly handling is implemented when the digital thread identifier fails to locate the check record, the regenerated frame group is not formed, or the regeneration break time is not detected. In such cases, a regeneration failure flag is written, and the export of the tag group generation is stopped. Then, the starting intersection position corresponding to the regenerated rupture time is checked backward to form a new three-part label result; the input quantities are the regenerated rupture time, regeneration record, regeneration frame group, quality control result, and exported version item; the processing action is to first read the intersection position set corresponding to the regenerated rupture time from the regeneration record, and then read the polarization difference position, phase difference position, and reserved position from the next regeneration frame group and the next regeneration frame group in the order of acquisition time. For each position in the intersection position set, a corresponding position check is performed. Any position that falls into the polarization difference position, phase difference position, and reserved position in the next regeneration frame group is written into the check position set; then, the intersection position set and the check position set are calculated for common position to obtain the regenerated heatmap, and written into the heatmap record in the order of row number and column number; Next, for each position in the regenerated heatmap, the reserved positions in the regeneration break time, the subsequent regeneration frame group, and the next subsequent regeneration frame group are read. A level 1 confidence score is written when the same position is written as a reserved position in all three; a level 2 confidence score is written when the same position is written as a reserved position in the regeneration break time and the subsequent regeneration frame group but not in the next subsequent regeneration frame group; and a level 3 confidence score is written when the same position is written as a reserved position only in the regeneration break time and not in the subsequent regeneration frame group. When the regenerated heatmap contains... When there are multiple locations, first write the first-level confidence, second-level confidence, and third-level confidence as "I", "II", and "III" respectively. Then, take the smallest value and write it as the regeneration confidence. Finally, write the regeneration failure time, regeneration heatmap, regeneration confidence, and current generation time into the same exported label group in a fixed field order to form a regeneration label group. The output is the regeneration heatmap, regeneration confidence, and regeneration label group. The exception handling is to write an empty heatmap flag when the regeneration heatmap is empty, write a no-generated confidence flag when the regeneration confidence is not generated, and stop the regeneration label group from being published. In practical applications: After the qualification re-examination module outputs the regeneration instruction, the regeneration module first reads the digital thread identifier from the regeneration instruction, then retrieves the polarization sequence fragment and phase sequence fragment from the corresponding inspection record, forms a regeneration frame group according to the acquisition time, and retrieves the first acquisition time that can form a corresponding connection with the next regeneration frame group from front to back, and writes it as the regeneration break time; then, starting from the set of intersection positions corresponding to the regeneration break time, it performs a corresponding review on the next regeneration frame group and the next regeneration frame group after that, forming a regeneration heatmap, and writes the results based on the three sets of reserved positions to form the regeneration confidence. Finally, the regeneration break time, regeneration heatmap, and regeneration confidence are written into the same export label group; after this processing, the source label results that fail the qualification re-examination no longer directly enter the export path, but are regenerated by the generation method corresponding to the export version item before entering the mapping export processing.

[0025] The mapping export module is used to perform field rewriting on the source tag group based on the field correspondence in the exported version item after receiving the mapping permission, and write the rewritten exported tag group with the digital thread identifier, device number, calibration identifier and session number, and output a standard data packet; it is also used to write the regenerated tag group with the digital thread identifier, device number, calibration identifier and session number after receiving the regenerated tag group, and output a standard data packet. This implementation focuses on the mapping export module performing field replacement and data packet writing on the source tag group or regenerated tag group. The aim is to write the source tag results that have passed the qualification re-verification or the regenerated tag results output by the regeneration module into data objects with the same structure, so that subsequent training reading, version tracking, and result verification directly target the standard data packet. During processing, an export tag group is first generated based on the source field name, export field name, and writing position in the export version item. Then, the digital thread identifier, device number, calibration identifier, and session number are written into the packet header area. The export tag group or regenerated tag group is written into the data area, and finally, the standard data packet is output according to a fixed concatenation order. This implementation process includes the following steps: First, rewrite the field content in the source tag group to the target version field according to the exported version item. The inputs are mapping permission, exported version item, and source tag group. The processing action is as follows: after receiving the mapping permission, first read the source field name, exported field name, and write position from the exported version item, and then locate the fracture time field, heatmap field, and confidence field in the source tag group according to the source field name. The heatmap field is written as a position string in the order of row number first and column number last. Then, replace the original field name with the located fracture time, heatmap, and confidence according to the exported field name, and then rearrange the field order from front to back according to the write position to form the exported tag group. The exported field name and write position are written to the exported record for subsequent reading. The outputs are the exported tag group and the exported record. The exception handling is as follows: if the field content of any source field name is not located, write a field missing flag and stop the generation of exported tag group. If any write position is missing, write a position missing flag and prohibit entry into standard data packet splicing. Subsequently, the thread source field and tag field are written to the header area and data area respectively, and then a single output object is formed in a fixed order; the input quantities are the exported tag group or regenerated tag group, digital thread identifier, device number, calibration identifier, and session number; the processing actions are as follows: first, the header area is established, and the digital thread identifier, device number, calibration identifier, and session number are written to the header area in the order of the fields pre-written in the system configuration table. The field names and field positions are written sequentially in the system configuration table, and the fields are written from front to back according to their positions; then, the data area is established, and if the exported tag group exists, the rupture time, heatmap, and confidence level in the exported tag group are written to the data area in the order of the fields. If the output tag group does not exist but the regenerated tag group does, the break time, heatmap, and confidence score from the regenerated tag group are written into the data area in field order. Then, the packet header is written first and the data area is written last to perform concatenation. The number of fields in the packet header, the number of fields in the data area, and the total length after concatenation are written into the packet record to form a standard data packet. The output consists of a standard data packet and a packet record. Exception handling is as follows: if any field in the packet header is missing, a packet header missing flag is written and concatenation is stopped; if the data area is empty, an empty data area flag is written and standard data packet output is stopped; if the total length after concatenation is not equal to the sum of the field lengths in the packet record, a concatenation failure flag is written and standard data packet distribution is prohibited. In practical applications: After the qualification re-examination module outputs the mapping permission, the mapping export module first reads the source field name, export field name, and write position from the export version item. Then, it locates the break time, heatmap, and confidence level in the source tag group, rewrites these three items as target version fields, and forms an export tag group according to the write position. If the qualification re-examination module has not output the mapping permission but the regeneration module has already output the regeneration tag group, the mapping export module directly reads the regeneration tag group and proceeds with subsequent writing. Subsequently, the mapping export module writes the digital thread identifier, device number, calibration identifier, and session number into the packet header area, writes the export tag group or regeneration tag group into the data area, and then forms a standard data packet in the order of packet header area first and data area last, and writes the number of fields and total length into the packet record. After this processing, the source tag path and the regeneration tag path finally output a standard data packet with the same structure. When the training side reads it later, it only needs to split it according to the packet header area and data area to retrieve the thread source field and tag field.

[0026] Further explanation is needed regarding the data structure of the standard data packet, which includes a header area, a tag area, a version mapping area, a flight pin area, a training tag area, and a manual review area. The header area contains the digital thread identifier, device number, calibration identifier, session number, source version identifier, and exported version identifier. The tag area contains the break time, heatmap, confidence level, tag generation time, and tag source flag, with the tag source flag taking the value of either source mapping or exported regeneration. The version mapping area contains the source field name, exported field name, write position, field length, and field content summary, which is generated by concatenating the field contents in the order of their write positions. The flight pin area contains the check record pointer, source version item pointer, exported version item pointer, source tag group pointer, regenerated tag group pointer, and qualification re-examination result pointer. If the regeneration path has not been entered, the regenerated tag group pointer is marked with a null pointer. The training tag area contains the training sample number, training break time, training heatmap area, training confidence level, and training inclusion flag. The manual review area contains the review entry number, review status, reviewer identifier, review time, review conclusion, and review write-back field. The training label area is generated jointly by the label area and the qualification re-examination result. When the qualification re-examination result is a mapping permission, the training break time, training heatmap area, and training confidence level are read from the exported label group, and the training inclusion mark is written as "included". When the qualification re-examination result is a regeneration instruction and the regeneration label group has been generated, the training break time, training heatmap area, and training confidence level are read from the regeneration label group, and the regeneration break time, regeneration heatmap, and regeneration confidence level are written as "regeneration". When the regeneration label group has not been generated, the label area contains undetected labels, empty heatmap area labels, or no confidence level labels are generated, the training inclusion mark is written as "pending review", and the corresponding digital thread identifier is written to the review entry number of the manual review area, so that the manual review end can retrieve the inspection record, polarization sequence fragment, phase sequence fragment, quality control result, and qualification re-examination result according to the digital thread identifier. The manual review interface is only used for registering the status and writing back the conclusion of the items to be reviewed in the standard data packet. It does not participate in the calculation of the automatic generation rules for the rupture time, heat map, and confidence level. After reading the review entry number, the review end displays the field content in the packet header area, tag area, version mapping area, and navigation pin area, and allows the reviewer to write the review status as unprocessed, confirmed, returned for regeneration, or removed from training. When the review conclusion is confirmed, the training inclusion flag is changed from pending review to included. When the review conclusion is returned for regeneration, the regeneration module is triggered again according to the digital thread identifier. When the review conclusion is removed from training, the training inclusion flag is written as not included. Through the above structure, the exported objects of the standard data packet are limited to data export tags, version mapping relationships, source tracing fields, training tag fields, and manual review interface fields, which is different from the data tracing scheme based on MRF graph structure for lesion area segmentation or diagnostic inference.

[0027] Working principle: This solution first extracts time-overlapping polarization sequence segments and phase sequence segments from the same dry eye inspection, and writes them into an inspection record along with the quality control results, equipment number, calibration identifier, session number, and source version identifier. Then, based on the source version identifier and the exported version identifier, the source version item and the exported version item are extracted respectively. The source version item is then combined with edge computing to generate the break time, heatmap, and confidence level of the inspection record, forming a source tag group. On this basis, the inspection record, source version item, and source tag group are bound to the same digital thread identifier. Then, the source tag group is re-evaluated based on the exported version item to determine whether this set of three tags can continue to be used under the exported version. If it can be used, the fields are rewritten and exported. If it cannot be used, the tags are regenerated according to the exported version item and then exported. Finally, a standard data packet with thread source information is output. For example, when accumulating dry eye examination data and periodically organizing training data in an ophthalmology clinic, the system first identifies the time period covered by both polarization and phase sequences from a single examination, and writes the data within this time period, along with the quality control results, into an examination record. Then, in an edge computing environment, it generates the break-off time, heatmap, and confidence level according to the old version rules. It then checks whether this set of three results can still be directly written into the new fields under the new version rules. If it still holds true, the field names and positions are changed to meet the new version requirements before outputting. If it no longer holds true, the system returns to the original examination record, recalculates the new break-off time, heatmap, and confidence level according to the new version rules, and then packages and outputs them together with the digital thread identifier, device number, calibration identifier, and session number. This way, the organized data retains the source relationship while adapting to the target version.

[0028] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A system for automatically generating and exporting tear film breakup time tags and version mappings, characterized in that, include: The acquisition device is used to acquire the polarization sequence and phase sequence corresponding to the same dry eye examination, and read the quality control results, equipment number, calibration identifier, session number and source version identifier corresponding to the same dry eye examination. It performs merging according to the same session number and outputs the examination record. Version storage device is used to store version table. The version table includes at least the method of generating the break time, the method of generating the heat map, the method of generating the confidence level, the release writing rules and the field correspondence for each version. It performs retrieval according to the source version identifier and the exported version identifier in the inspection record and outputs the source version item and the exported version item. The tag generation module performs fracturing time extraction, heatmap generation, and confidence level generation on the inspection record based on the source version item, and writes the fracturing time, heatmap, and confidence level obtained in the same generation process into the same tag group, and outputs the source tag group; The thread binding module is used to perform corresponding bindings on the inspection record, source version item, and source tag group, generate a numeric thread identifier that runs through the inspection record, source version item, and source tag group, and output the thread binding result. The qualification re-verification module is used to call the exported version item, perform export review on the source tag group based on the thread binding result, and output the mapping license when all three checks are true, and output the regeneration instruction when any one check is false. The qualification re-examination module also includes creating a peer review table and outputting the review field group: The digital thread identifier is split into a time bit, a heatmap bit, and a confidence bit. The rupture time, heatmap, and confidence level in the source tag group are written into the time bit, heatmap bit, and confidence bit, respectively, to generate a peer review table. Perform position extraction on the time writing rules, heatmap writing rules and confidence writing rules in the exported version item, compare the position in the same position verification table with the position in the exported version item item by item, write the set of fields with the same position into the verification field group, and output the verification field group.

2. The automatic tear film breakup time tag generation and version mapping export system according to claim 1, characterized in that: Also includes: The regeneration module is used to call the exported version item to re-execute the rupture time extraction, heatmap generation and confidence generation on the inspection record after receiving the regeneration instruction, and write the regenerated rupture time, heatmap and confidence into the same exported label group, and output the regenerated label group.

3. The automatic tear film breakup time tag generation and version mapping export system according to claim 2, characterized in that: Also includes: The mapping export module is used to perform field rewriting on the source tag group based on the field correspondence in the exported version item after receiving the mapping permission, and write the rewritten exported tag group with the digital thread identifier, device number, calibration identifier and session number, and output a standard data packet; it is also used to write the regenerated tag group with the digital thread identifier, device number, calibration identifier and session number after receiving the regenerated tag group, and output a standard data packet.

4. The automatic tear film breakup time tag generation and version mapping export system according to claim 3, characterized in that: The acquisition device also includes functions such as reading the acquisition time, extracting overlapping portions, performing merging, and writing to the inspection record. Read the acquisition time corresponding to each frame in the polarization sequence and the acquisition time corresponding to each frame in the phase sequence, perform a comparison between the acquisition time period of the polarization sequence and the acquisition time period of the phase sequence, and extract the data corresponding to the overlapping part when the session number is consistent, and output the polarization sequence segment and the phase sequence segment. While outputting polarization sequence fragments and phase sequence fragments, the system reads the write times corresponding to the quality control results, device number, calibration identifier, and source version identifier. It then performs a comparison between the write times and the overlapping portions. The system merges the quality control results, device number, calibration identifier, session number, and source version identifier with the polarization sequence fragments and phase sequence fragments according to a fixed field order and writes them into the overlapping portions, and outputs the inspection record.

5. The automatic tear film breakup time tag generation and version mapping export system according to claim 4, characterized in that: The version storage device also includes a version table, the ability to read source version identifiers and exported version identifiers, and the ability to perform retrieval and output source version items and exported version items. Write the version identifier, the break time generation method identifier, the heat map generation method identifier, the confidence level generation method identifier, the release writing rule identifier, and the field correspondence table number into the same version row. Write each version row into the version table according to the version identifier. In the field correspondence table, write the source field name, the exported field name, and the writing position according to the exported version identifier. Output the version table. After outputting the version table, read the source version identifier and exported version identifier from the inspection record. Perform a line-by-line comparison of the source version identifier in the version table to locate the source version row. Perform a line-by-line comparison of the exported version identifier in the version table to locate the exported version row. Read the fracture moment generation method identifier, heatmap generation method identifier, and confidence generation method identifier from the source version row and output the source version item. Read the release write rule identifier from the exported version row and combine it with the source field name, exported field name, and write position that are consistent with the exported version identifier in the field correspondence table to output the exported version item.

6. The automatic tear film breakup time tag generation and version mapping export system according to claim 5, characterized in that: The tag generation module also includes extracting the fracture time, generating a heatmap, generating confidence scores, and writing them to the source tag group. The polarization sequence, phase sequence, quality control results, and source version item in the inspection record are used as input. The same frame groups are formed by pairing them according to the acquisition time. For each same frame group, the difference position between the current polarization frame and the previous polarization frame, and the difference position between the current phase frame and the previous phase frame are calculated. The intersection calculation is performed with the reserved position in the quality control result. The frames are sorted from front to back according to the acquisition time. The first acquisition time that has a non-empty intersection position set and has a corresponding position with the intersection position set of the next same frame group is taken as the candidate rupture time. For the intersection position corresponding to the candidate rupture time, the same frame group after the candidate rupture time is checked according to the acquisition time order. The same position that falls into the polarization difference position, phase difference position and retention position is written into the heat map area. The common position set of the intersection position corresponding to the candidate rupture time and the check position of the next same frame group is used as the target heat map area. Based on the candidate rupture time, target heatmap area, quality control results, and source version, perform sequential verification on each position in the target heatmap area. Write a first-level confidence level when the candidate rupture time, the next co-occurrence frame group, and the next co-occurrence frame group are all written to the reserved position; write a second-level confidence level when the candidate rupture time and the next co-occurrence frame group are written to the reserved position but the next co-occurrence frame group is not written to the reserved position; and write a third-level confidence level when the candidate rupture time is written to the reserved position but the next co-occurrence frame group is not written to the reserved position. Write the candidate rupture time, target heatmap area, and confidence level to the same label group and output the source label group.

7. The automatic tear film breakup time tag generation and version mapping export system according to claim 6, characterized in that: The thread binding module also includes extracting fields, writing binding strings, generating numeric thread identifiers, and outputting thread binding results. Perform field extraction on the session number, device number, and calibration identifier in the inspection record, the source version identifier in the source version item, and the rupture time, heat map, and confidence level in the source tag group. Write these fields into the binding string in the fixed field order of session number, device number, calibration identifier, source version identifier, rupture time, heat map, and confidence level, and output the binding string. Write the field position and field length of each field in the binding string in sequence, and concatenate the field length and field content in the order of field position to generate a numeric thread identifier. Then write the numeric thread identifier into the inspection record, source version item and source tag group respectively, and output the thread binding result.

8. The automatic tear film breakup time tag generation and version mapping export system according to claim 7, characterized in that: The qualification re-examination module also includes performing peer comparison and outputting the qualification re-examination results: The rupture time, heatmap, and confidence level in the verification field group are sequentially written into the time writing rule, heatmap writing rule, and confidence level writing rule to obtain a positive result. The positive result is then written back into the verification field group in its original position to obtain a reverse result. Perform a peer comparison on the positive result, negative result, and verification field group. Output mapping permission when the three items of rupture time, heatmap, and confidence are consistent. If any item is inconsistent, write the inconsistent item into the regeneration instruction and output the qualification re-examination result.

9. The automatic tear film breakup time tag generation and version mapping export system according to claim 8, characterized in that: The regeneration module also includes forming a regeneration frame group, performing intersection calculation, performing peer verification, and outputting the regeneration tag group: The regeneration command and the inspection record are paired according to the acquisition time to form a regeneration frame group. The polarization difference position and phase difference position are calculated for each regeneration frame group according to the exported version item, and the intersection calculation is performed with the retained position in the quality control result. The acquisition time that is not empty in the intersection position set and has the same position intersection position with the next regeneration frame group is taken as the regeneration break time according to the acquisition time order. Based on the intersection position corresponding to the regeneration break time, perform a frame-by-frame co-location check on subsequent regeneration frame groups. Write the co-location positions that simultaneously fall into the polarization difference position, phase difference position, and retention position into the regeneration heatmap. Write the retention write results in the regeneration break time, the next regeneration frame group, and the next regeneration frame group into the regeneration confidence score. Then write the regeneration break time, the regeneration heatmap, and the regeneration confidence score into the same exported label group and output the regeneration label group.

10. The automatic tear film breakup time tag generation and version mapping export system according to claim 9, characterized in that: The mapping export module also includes reading the source field name, the exported field name, and the write position, performing field substitution, corresponding writing, and outputting a standard data packet: After receiving the mapping permission, read the source field name, export field name and write position in the export version item, locate the break time, heatmap and confidence in the source tag group according to the source field name, replace the field name according to the export field name and rearrange the field order according to the write position to generate the export tag group. Perform corresponding write operations on the exported tag group or regenerated tag group, write the digital thread identifier, device number, calibration identifier, and session number into the header area in a fixed order, write the exported tag group or regenerated tag group into the data area in the field order, and then perform splicing with the header area first and the data area last to output a standard data packet.

Images