A lacrimal passage flushing navigation method, system and terminal based on image recognition

Abstract

The application belongs to the technical field of lacrimal passage flushing, and particularly discloses a lacrimal passage flushing navigation method, system and terminal based on image recognition. The method comprises the following steps: generating a lacrimal passage structure diagram from an eye region image, and identifying an occlusion region on the diagram; constructing one or more candidate dredging paths for the occlusion region; establishing a functional relationship model between the injection amount and the pressure change value of the candidate dredging path; and evaluating the candidate dredging path based on the functional relationship model to output a recommended scheme. The application can realize the identification and positioning of the lacrimal passage structure and the occlusion region, quantify the relationship between the injection amount and the pressure change value of the flushing liquid, and evaluate and select the candidate dredging path based on the functional relationship model, thereby improving the success rate and safety of lacrimal passage flushing.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of lacrimal passage flushing, and particularly relates to a lacrimal passage flushing navigation method, system and terminal based on image recognition. BACKGROUND

[0002] In ophthalmic clinical practice, lacrimal passage flushing is used to assess the patency of the lacrimal passage system by injecting liquid into the lacrimal passage system, and can also relieve the symptoms of patients such as epiphora and eye discomfort by flushing and removing the blockage or inflammatory secretions in the pipeline, thereby assisting in the diagnosis of lacrimal passage-related diseases.

[0003] In the path planning stage before the lacrimal passage flushing of the prior art, the lacrimal passage structure is mainly manually interpreted and divided by the experience of doctors. Due to the highly complex and individualized anatomical characteristics of the lacrimal passage system, it contains numerous branch structures with different shapes and staggered distribution. Especially in the area where the intersection is not obvious or the shape is atypical, it is easy to miss some branches.

[0004] The path recognition method of the prior art is usually based on qualitative judgment of images, which is difficult to evaluate the patency, direction and applicability of each potential branch path, so that the path selection is blind and the real blocked area or main drainage channel is easily missed, so that the subsequent flushing operation cannot hit the target and the treatment effect is poor.

[0005] Based on the above, how to realize objective and accurate automatic identification and evaluation of complex branch pipelines, so as to plan a safe and efficient flushing path, has become a technical problem to be solved in the current ophthalmic diagnosis and treatment technical field. SUMMARY

[0006] The purpose of the present application is to provide a lacrimal passage flushing navigation method, system and terminal based on image recognition to solve the problems of branch structure misjudgment or misjudgment and high risk of lacrimal passage flushing in the prior art.

[0007] The technical solutions adopted by the present application are as follows:

[0008] A lacrimal passage flushing navigation method based on image recognition, when an occluded area is identified on a lacrimal passage structure map generated from an eye region image, the following steps are performed:

[0009] Constructing one or more candidate unblocking paths for the occluded area;

[0010] Establishing a functional relationship model between the injection volume and the pressure change value of the candidate unblocking path;

[0011] Applying the functional relationship model to score each candidate unblocking path, and outputting the candidate unblocking path with a score higher than a preset threshold as a recommended solution.

[0012] The recommendation scheme output includes: if the score values of all the candidate unblocking paths are higher than the preset threshold, the anatomically normal path is selected preferentially; if the score values of part of the candidate unblocking paths are higher than the preset threshold, the candidate unblocking path with the highest score value is selected from the candidate unblocking paths with the score values higher than the preset threshold; and if the score values of all the candidate unblocking paths are not higher than the preset threshold, the candidate unblocking path with the highest score value is selected from all the candidate unblocking paths.

[0013] Preferably, the generation of the lacrimal passage structure diagram includes calibrating the eye region image to obtain structures including the lacrimal punctum, lacrimal canaliculus, lacrimal sac and nasolacrimal duct; and the identification of the obstruction region includes locating the injection path of the flushing liquid on the lacrimal passage structure diagram and extracting the key restricted sites on the injection path.

[0014] Preferably, the identification of the obstruction region further includes: converting the lacrimal passage structure diagram into a gray-scale image; finding a region with a lower brightness than a preset brightness threshold and a continuous distribution in the gray-scale image; and judging the obstruction region by comparing the offset degree of the region and a standard shape.

[0015] Preferably, the function relationship model between the injection volume and the pressure change value of the candidate unblocking path is established by: injecting the flushing liquid into the communication node set for each candidate unblocking path, and extracting the pressure change value corresponding to different injection volumes to form a data pair; and establishing the function relationship model with the injection volume as a first parameter and the pressure change value as a second parameter.

[0016] Preferably, in the process of establishing the function relationship model, when the pressure change value is higher than a preset upper limit threshold of pressure measurement or lower than a preset lower limit threshold of pressure measurement, recalibration or adjustment of the calculation period is performed.

[0017] An image recognition-based lacrimal passage flushing navigation system includes the following modules:

[0018] A lacrimal passage structure analysis module configured to generate a lacrimal passage structure diagram from an acquired eye region image and identify an obstruction region on the lacrimal passage structure diagram;

[0019] An unblocking path evaluation module configured to: in response to the identification of the obstruction region by the lacrimal passage structure analysis module, construct one or more candidate unblocking paths, establish a function relationship model between the injection volume and the pressure change value of the candidate unblocking path, and evaluate the candidate unblocking path based on the function relationship model to output a recommendation scheme.

[0020] Preferably, the unblocking path evaluation module, when establishing the function relationship model, is further configured to monitor the pressure change value and trigger recalibration when the pressure change value is higher than a preset upper limit threshold of pressure measurement or lower than a preset lower limit threshold of pressure measurement.

[0021] Preferably, the dredging path evaluation module scores each candidate dredging path by applying a functional relationship model, and outputs the candidate dredging path with a score value higher than a preset threshold as a recommended scheme.

[0022] Preferably, the functional relationship model between the injection amount and the pressure change value of the candidate dredging path is established by injecting flushing liquid to the communication node set for each candidate dredging path, and extracting the pressure change value corresponding to different injection amounts to form a data pair, and establishing the functional relationship model with the injection amount as a first parameter and the pressure change value as a second parameter.

[0023] An image recognition-based lacrimal passage flushing navigation terminal, comprising a processor and a memory electrically connected to the processor; the memory has computer program instructions, and when the instructions are executed by the processor, the terminal executes an image recognition-based lacrimal passage flushing navigation method as described above.

[0024] Advantages

[0025] The image recognition-based lacrimal passage flushing navigation method provided by the application generates a lacrimal passage structure diagram from an eye region image, identifies an occlusion region on the lacrimal passage structure diagram, determines an injection path and key restricted positions on the injection path, constructs candidate dredging paths for the occlusion region, sets a starting point, an ending point and intermediate nodes for the candidate dredging paths, records key turning points in the candidate dredging paths as communication nodes, realizes identification and positioning of the lacrimal passage structure and the occlusion region, and provides an anatomical basis for dredging path planning, thereby overcoming the problems of inaccurate positioning and strong subjectivity of traditional methods.

[0026] The application injects flushing liquid to the communication nodes, extracts the pressure change values corresponding to different injection amounts to form a data pair, establishes a functional relationship model between the first parameter and the second parameter with the injection amount as the first parameter and the pressure change value as the second parameter, quantifies the relationship between the flushing liquid injection amount and the pressure change value, objectively evaluates the patency degree and occlusion condition of the lacrimal passage, overcomes the defect that the prior art cannot accurately quantify the occlusion degree of the lacrimal passage, and provides a reliable physiological basis for effective evaluation of the dredging path.

[0027] The application applies a function relationship model to score each candidate dredging path, and outputs the candidate dredging path with a score value higher than a preset threshold as a recommended scheme; and calculates the maximum allowable injection amount and the tolerance threshold of each candidate dredging path in the flushing process according to the calculation result of the function relationship model, selects the dredging path, avoids damage to the patient due to flushing failure, improves the success rate and safety of the lacrimal duct flushing, and improves the intelligentization and automation level of the navigation system, ensures the stability and safety of the lacrimal duct flushing process. BRIEF DESCRIPTION OF DRAWINGS

[0028] Fig. 1 is a method flowchart of the application;

[0029] Fig. 2 is a recommended scheme output flowchart based on the score value of the application. DETAILED DESCRIPTION

[0030] The technical solutions of the patent will be further described in detail in combination with specific embodiments. The following embodiments are used to illustrate the application, but cannot be used to limit the protection scope of the application. The conditions in the embodiments can be further adjusted according to specific conditions, and simple improvements of the method of the application under the concept of the application belong to the protection scope of the application.

[0031] Example 1

[0032] Please refer to Figs. 1-2 The embodiment provides a lacrimal duct flushing navigation method based on image recognition, and specifically comprises the following steps:

[0033] An image of an eye region of a patient to be tested is acquired, and the image of the eye region of the patient to be tested is collected by an infrared camera or a high-definition camera to ensure the definition and details of the image. The image of the eye region is processed in blocks. The position of a lacrimal point is determined by recognizing and geometrically positioning specific physiological feature points such as a canthus and a pupil edge in the image. A neighboring region containing main anatomical structures of a lacrimal passage is demarcated with the lacrimal point as the center. In the neighboring region, color information of the image is analyzed, and a color region with a color depth greater than a preset color depth threshold value is extracted. In this embodiment, the preset color depth threshold value is a color depth limit value for distinguishing lacrimal passage tissues from surrounding non-lacrimal passage tissues such as skin or conjunctiva, which is obtained according to a large amount of clinical data statistical analysis to ensure the accuracy of recognition. The lacrimal passage tissues are pink or light red mucosa. The shape, size, continuity and other geometric information of the color region are compared with standard parameters stored in a database. In this embodiment, the standard parameters refer to a reference data set established based on the diameter range of the lacrimal canaliculus, the morphological characteristics of the lacrimal sac, the running path of the nasolacrimal duct and other normal lacrimal passage anatomical structures.

[0034] Specifically, if the geometric information of the color region deviates from the standard parameters within a preset tolerance range, the region is regarded as a key structural part such as the lacrimal canaliculus, the lacrimal sac and the nasolacrimal duct in the lacrimal passage. The preset tolerance range is the maximum acceptable deviation range between the measured or calculated value and the standard value when data comparison or judgment is performed. The key structural part is image calibrated and distinguished by different colors or lines to generate a complete lacrimal passage structure diagram for next step processing. The lacrimal passage structure diagram is an image in which the key structural parts such as the lacrimal canaliculus, the lacrimal sac and the nasolacrimal duct in the lacrimal passage are distinguished by different colors or lines through image recognition and calibration. The lacrimal passage structure diagram is used to visualize the lacrimal passage anatomical structure.

[0035] Further, in the calibrated lacrimal passage structure diagram, the injection path is marked by recognizing the initial flow trajectory of the flushing liquid in the lacrimal passage. The injection path can be used for subsequent analysis and navigation. The injection path is further scanned, the lacrimal passage structure diagram is converted into a gray-scale image, the display effect of the brightness difference in the image is enhanced, and the structural details inside the lacrimal passage are presented. In the gray-scale image, a region with a brightness lower than a preset brightness threshold value and a continuous distribution is found through pixel brightness value analysis. The preset brightness threshold value is a brightness limit value set according to the gray-scale characteristics of normal lacrimal passage tissues. In this embodiment, the average gray-scale value of healthy lacrimal passage mucosa is preferred. The preset brightness threshold value is used to preliminarily screen out regions that may have abnormalities.

[0036] Further, the region with a low luminance and continuous distribution is subjected to shape analysis, and the contour thereof is compared with a standard shape such as a circular or elliptical cross section of a healthy lacrimal duct lumen, and the deviation degree is calculated. If the deviation degree is not greater than a preset tolerance range, for example, the region presents abnormal features such as a tumor-shaped expansion, a lumen stenosis, or an irregular shape, the region is determined as an obstruction region. The obstruction region is a region that causes the flushing liquid to be unable to pass smoothly, and the obstruction region is divided into a key limited part that constitutes a main obstacle to the flow of the flushing liquid and is highlighted for subsequent evaluation of an alternative path.

[0037] Further, according to the position and range of the obstruction region, a branch adjacent to the obstruction region in the lacrimal duct structure diagram is automatically searched, and a lacrimal duct branch path for bypassing the obstruction region is constructed as a candidate dredging path. The candidate dredging path can bypass the obstruction region, and each candidate dredging path is provided with a clear starting point, an ending point, and an intermediate node. In the construction of the candidate dredging path, the ending point of the obstruction region is reversely extended to an adjacent lacrimal duct intersection, and the intersection is taken as a starting point to continue to search for a branch direction, so as to construct a plurality of possible dredging paths. In the plurality of candidate dredging paths, the communication nodes are key turning points in the candidate dredging paths, and the position selection principle is that the branch is an intersection or a branch starting point of an adjacent main structure. These communication nodes are potential positions where the flushing liquid can be safely injected and effectively dredged to ensure that the flushing liquid can pass smoothly and cause minimal damage to the surrounding tissue.

[0038] Further, a functional relationship model between the injection amount and the pressure change value of the candidate dredging path is established, the obtained communication nodes are sequentially injected with flushing liquids such as physiological saline or cooling liquid, and the injection amount of each injection is accurately recorded. The pressure change value of the instantaneous pressure in the lacrimal duct during the injection process is synchronously collected, each pair of injection amount and pressure change value is saved as a data pair, and after the data collection is completed, the injection amount is taken as a first parameter, and the pressure change value is taken as a second parameter. The data pairs are formed in the analysis software and are used to construct a functional relationship model describing the quantitative relationship between the two parameters. All data pairs are input into a modeling module, and a functional relationship model between the first parameter and the second parameter is established by linear fitting or polynomial regression statistical methods.

[0039] The functional relationship model between the first parameter and the second parameter is a calculation model describing the correlation between the injection amount of the flushing liquid and the pressure change in the lacrimal duct, and can be used to predict the influence on the intraocular pressure under different injection speeds and volumes, and to judge whether there is a risk in the dredging process.

[0040] The functional relationship between the first parameter and the second parameter established by linear fitting is:

[0041]

[0042] The function relationship between the first parameter and the second parameter is established by polynomial regression, and is as follows:

[0043]

[0044] In the above formula, V represents the injection amount, which means the volume of the flushing liquid injected each time, and the unit is mL; ΔP represents the pressure change value, which means the change amount of the instantaneous pressure in the tear duct during the injection of the flushing liquid, and the unit is mmHg; a, b, c, d, and e represent fitting coefficients, which are parameters learned from the data by a statistical method of least squares and are used to describe the mathematical relationship between the injection amount and the pressure change. Any of the above function relationship formulas can be used to predict the influence on the intraocular pressure under different injection speeds and volumes.

[0045] The function relationship model between the first parameter and the second parameter can describe the correlation between the injection amount of the flushing liquid and the pressure change in the tear duct, and can be used to predict the influence on the intraocular pressure under different injection speeds and volumes, and to judge whether there is a risk in the dredging process. In the process of establishing the function relationship model, if the pressure change value is higher than the preset upper limit threshold of the pressure measurement or lower than the preset lower limit threshold of the pressure measurement, the injection will be immediately paused and the abnormal state will be recorded, and the current node will be marked as an unusable point that is not suitable for continuing the injection. Then, the next available node is selected for injection testing, and the calculation period is recalibrated or adjusted. In this embodiment, the preset upper limit threshold and the preset lower limit threshold of the pressure measurement are respectively the safe upper limit and the safe lower limit of the pressure change value in the tear duct, which are set according to clinical experience and the physiological tolerance range of the patient. Exceeding the preset upper limit threshold of the pressure measurement may cause damage to the tear duct tissue, and being lower than the preset lower limit threshold of the pressure measurement may indicate an abnormal state. The calculation period refers to the time interval required for re-evaluation and adjustment of the injection strategy under abnormal conditions. Recalibrating or adjusting the calculation period is to ensure the accuracy and safety of data acquisition under abnormal conditions.

[0046] Further, the function relationship model between the first parameter and the second parameter is applied to score each of the candidate dredging paths. According to the calculation result of the function relationship model, the flow limit value of each of the candidate dredging paths is calculated. The flow limit value is the maximum flushing liquid flow that the path can withstand without causing damage to the tear duct. Thus, the maximum allowed injection amount of the tear duct per time or per unit time under the premise of not causing damage to the tear duct is calculated, and the tolerance threshold of the patient is evaluated. The tolerance threshold is the maximum flushing liquid injection amount or pressure that the tear duct can withstand, which is comprehensively considered in terms of the physiological response of the patient and the elasticity of the tear duct tissue.

[0047] The score value of each candidate dredging path is compared with a preset threshold value, the preset threshold value is a safety operation standard set according to clinical experience, and is used for evaluating the score value of the dredging path. A set of preset sorting rules are used to score each candidate dredging path. The preset sorting rules are predetermined criteria for comparing and sorting the score values of the dredging paths to determine the recommended scheme.

[0048] Specifically, the preset sorting rules of the recommended scheme are as follows: if the score values of all candidate dredging paths are higher than the preset threshold value, an anatomically normal path that generally has better physiological adaptability is preferentially selected as the recommended scheme; if the score values of part of the candidate dredging paths are higher than the preset threshold value, the candidate dredging path with the highest score value is selected as the recommended scheme from the candidate dredging paths with the score values higher than the preset threshold value, so as to ensure that the optimal safety path is selected; if the score values of all candidate dredging paths are not higher than the preset threshold value, the candidate dredging path with the highest score value is selected from all candidate dredging paths as the recommended scheme, and a risk prompt is given, and the doctor is suggested to operate carefully or consider other treatment schemes. The lacrimal duct flushing treatment plan output by the recommended scheme and the corresponding operation suggestion is taken as the lacrimal duct flushing scheme.

[0049] Embodiment Two

[0050] The embodiment provides a lacrimal duct flushing navigation system based on image recognition, which automatically generates a lacrimal duct structure diagram from an eye region image, identifies an obstruction region, quantitatively evaluates a plurality of candidate dredging paths based on a fluid dynamics model, and outputs a scientific and reliable recommended scheme for an operator, thereby improving the success rate and safety of lacrimal duct flushing surgery.

[0051] The system can be an independent lacrimal duct flushing navigation terminal based on image recognition, or can be integrated into an existing ophthalmic diagnostic device. The lacrimal duct flushing navigation terminal includes a processor and a memory electrically connected to the processor. The memory stores computer program instructions. When the instructions are executed by the processor, the functions of the following logical modules of the system are realized.

[0052] The lacrimal duct structure analysis module is responsible for processing the input eye region images, reconstructing the lacrimal duct structure, and identifying obstruction areas. This module acquires one or more eye region images, calibrates them to identify and delineate key structural components of the lacrimal duct, such as the lacrimal punctum, lacrimal canaliculi, lacrimal sac, and nasolacrimal duct, thereby generating a visualized lacrimal duct structure map. After generating the lacrimal duct structure map, it identifies obstruction areas. The lacrimal duct structure analysis module locates the irrigation fluid injection path on the lacrimal duct structure map. By analyzing the geometric morphology or signal features of this path, it extracts areas with significantly narrowed diameters or abnormal structures as key restricted areas. The lacrimal duct structure map is converted into a grayscale image. In the grayscale image, it searches for areas with brightness below a preset brightness threshold and continuous spatial distribution. By comparing the shape of the identified area with a standard shape, its offset is calculated. When the offset exceeds a preset tolerance range, the area is confirmed as an obstruction area, and its location and extent information are transmitted to the unblocking path evaluation module.

[0053] The unblocking path evaluation module is activated after receiving information about the obstructed area identified by the lacrimal duct structure analysis module. It constructs and evaluates recommended solutions. In response to the identification of the obstructed area, it constructs one or more candidate unblocking paths around the obstructed area and establishes a functional relationship model between the injection volume and pressure change value for each candidate unblocking path. This module sets communication nodes at key locations in each path and injects flushing fluid into these communication nodes through simulation or actual guidance, while simultaneously extracting the pressure change values ​​corresponding to different injection volumes, forming a series of data pairs. The unblocking path evaluation module establishes a functional relationship model with the injection volume as the first parameter and the pressure change value as the second parameter. During the establishment of the functional relationship model, the unblocking path evaluation module also performs a monitoring task: continuously monitoring the collected pressure change values. When the pressure change value is found to be higher than the preset upper limit threshold or lower than the preset lower limit threshold, it immediately recalibrates or adjusts the calculation cycle to ensure the validity of the data on which the model is based.

[0054] After the functional relationship model of the dredging path evaluation module is established, it is applied to comprehensively score each candidate dredging path. The calculated score is compared with a preset threshold, and all candidate dredging paths with scores higher than the preset threshold are selected as alternatives. The final recommended solution is determined and output according to the following rules: If the scores of all candidate dredging paths are higher than the preset threshold, the system will prioritize the anatomically normal path; if only some candidate dredging paths have scores higher than the preset threshold, the path with the highest score among these paths will be selected as the recommended solution; if the scores of all candidate dredging paths are not higher than the preset threshold, the system will still select the path with the highest score among all paths, and may include risk warnings for the operator's decision-making reference.

[0055] Through the cooperative work of the above tear duct structure analysis module and the dredging path evaluation module, the system of the embodiment can convert the complex tear duct flushing process into a navigation task based on image analysis and model calculation, realize accurate diagnosis of the blocked area, provide quantitative decision basis for the selection of the dredging path, improve the scientificity of the operation planning and the precision of the operation, and can be used to handle complex clinical scenarios.

[0056] The above only describes the preferred embodiments of the present application, and it should be noted that those skilled in the art can make several improvements and refinements without departing from the principles of the present application, and these improvements and refinements should also be considered within the protection scope of the present application.

Claims

1. An image recognition based lacrimal passage irrigation navigation method, characterized by, When the occlusion area is identified on the lacrimal passage structure map generated from the eye region image, the following steps are performed: One or more candidate unblocking paths for the occlusion area are constructed; A functional relationship model between the injection amount and the pressure change value of the candidate unblocking path is established; The function relationship model is applied to score each candidate unblocking path, and the candidate unblocking path with a score higher than a preset threshold is output as a recommended solution; Wherein, the recommended solution output includes: if the scores of all candidate unblocking paths are higher than the preset threshold, the anatomically normal path is preferentially selected; if the scores of some candidate unblocking paths are higher than the preset threshold, the candidate unblocking path with the highest score is selected from the candidate unblocking paths with a score higher than the preset threshold; if the scores of all candidate unblocking paths are not higher than the preset threshold, the candidate unblocking path with the highest score is selected from all candidate unblocking paths; Wherein, the generation of the lacrimal passage structure map includes calibrating the eye region image to obtain structures including the lacrimal punctum, lacrimal canaliculus, lacrimal sac and nasolacrimal duct; the identification of the occlusion area includes locating the injection path of the flushing liquid on the lacrimal passage structure map and extracting the key restricted sites on the injection path; Wherein, the identification of the occlusion area further includes: converting the lacrimal passage structure map into a grayscale image; finding a region with a lower brightness than a preset brightness threshold and a continuous distribution in the grayscale image; determining the occlusion area by comparing the offset degree of the region and the standard shape.

2. The image recognition based lacrimal passage irrigation navigation method according to claim 1, characterized in that, Establishing a functional relationship model between the injection amount and the pressure change value of the candidate unblocking path includes: injecting flushing liquid into the communication node set for each candidate unblocking path, and extracting the pressure change value corresponding to different injection amounts to form a data pair; establishing a functional relationship model with the injection amount as the first parameter and the pressure change value as the second parameter.

3. The image recognition based lacrimal passage irrigation navigation method according to claim 2, characterized in that, During the establishment of the functional relationship model, when the pressure change value is higher than a preset upper limit threshold or lower than a preset lower limit threshold, recalibration or adjustment of the calculation period is performed.

4. An image recognition based lacrimal passage irrigation navigation system, characterized by, Comprise the following modules: A lacrimal passage structure analysis module configured to generate a lacrimal passage structure map from an acquired eye region image and identify an occlusion area on the lacrimal passage structure map; A unblocking path evaluation module configured to, in response to the identification of the occlusion area by the lacrimal passage structure analysis module, construct one or more candidate unblocking paths, establish a functional relationship model between the injection amount and the pressure change value of the candidate unblocking path, and evaluate the candidate unblocking path based on the functional relationship model to output a recommended solution; Wherein, the generation of the lacrimal passage structure map includes calibrating the eye region image to obtain structures including the lacrimal punctum, lacrimal canaliculus, lacrimal sac and nasolacrimal duct; the identification of the occlusion area includes locating the injection path of the flushing liquid on the lacrimal passage structure map and extracting the key restricted sites on the injection path; Wherein, the identification of the occlusion area further includes: converting the lacrimal passage structure map into a grayscale image; finding a region with a lower brightness than a preset brightness threshold and a continuous distribution in the grayscale image; determining the occlusion area by comparing the offset degree of the region and the standard shape.

5. The image recognition based lacrimal passageway flushing navigation system according to claim 4, wherein, The unblocking path evaluation module is also used to monitor the pressure change value when establishing the functional relationship model, and trigger recalibration when the pressure change value is higher than a preset upper limit threshold or lower than a preset lower limit threshold.

6. The image recognition based lacrimal passageway flushing navigation system according to claim 4, wherein, The dredging path evaluation module scores each candidate dredging path by applying a function relationship model, and outputs the candidate dredging path with a score higher than a preset threshold as a recommended scheme.

7. The image recognition based lacrimal passageway flushing navigation system of claim 4, wherein, The function relationship model between the injection volume and the pressure change value of the candidate dredging path is established by injecting flushing liquid into the communication node set for each candidate dredging path and extracting the pressure change value corresponding to different injection volumes to form a data pair, and establishing the function relationship model with the injection volume as a first parameter and the pressure change value as a second parameter.

8. An image recognition-based lacrimal passage irrigation navigation terminal, characterized by, The terminal comprises a processor and a memory electrically connected to the processor; the memory has computer program instructions, and when the instructions are executed by the processor, the terminal executes the image recognition-based lacrimal passage irrigation navigation method according to any one of claims 1 to 3.

Images