A pre-cataract surgery visual function expectation assessment method

Abstract

The application discloses a pre-cataract surgery visual function expectation evaluation method, collects preoperative basic examination data, subjective expectation expression data and a preset visual task scene library, and screens a candidate visual task scene set; a scene acceptance result set is obtained through patient acceptance interaction discrimination; subjective expectation expression data are combined and analyzed to generate a visual function requirement parameter set; a visual task tolerance threshold is further quantified and structured and coded into an expectation feature vector, and a preoperative visual function expectation evaluation result is output, which is used for preoperative lens matching, preoperative communication prompting or postoperative satisfaction risk prediction.

Description

Technical Field

[0001] This invention relates to the field of medical and health information processing and medical test data analysis technology, and in particular to a method for assessing preoperative visual function expectations in cataract surgery based on preoperative examination data, subjective expectation expression data, and visual task scene interaction discrimination data. Background Technology

[0002] While current preoperative visual function assessment techniques for cataract surgery have gradually expanded from simply focusing on visual acuity, refractive status, and ocular biometric parameters to considering patients' subjective visual needs and postoperative satisfaction, in practice, existing methods for obtaining "patient expectations" still mainly remain at the level of oral examinations, questionnaires, or simple needs categorization. Patient expectations are often summarized as vague descriptions such as "clearer distance vision," "good near vision," and "reduced need for glasses." This approach superficially completes the expectation collection, but in reality, it assumes that patients can accurately identify and express their true visual needs, ignoring the fact that patients often do not describe themselves based on specific visual function indicators before cataract surgery, but rather express their needs through vague life experiences, habitual references, and subjective feelings. Therefore, their verbal expectations often do not correspond to the key visual tasks that truly determine postoperative satisfaction.

[0003] Looking further, in specific scenarios such as reading small print, recognizing road signs at night, identifying object boundaries in low-light environments, maintaining visual comfort in highly reflective scenes, and visual recovery speed when switching between indoor and outdoor environments, patients are not truly concerned with "seeing clearly" in an abstract sense, but rather with their tolerance threshold for the consequences of failure in different visual tasks. However, existing assessment processes often fail to extract these contextualized, task-oriented, and threshold-based implicit needs from natural language expressions. As a result, the system input only obtains coarse-grained demand labels after language compression, rather than the true demand entity that can directly support preoperative matching and postoperative expectation management. Based on this, regardless of the parameter analysis, plan recommendation, or satisfaction prediction methods used subsequently, they are all actually built on distorted demand input, thus forming a continuous error chain of "inaccurate input—assessment deviation—matching defocus—decreased satisfaction."

[0004] Therefore, the most critical, essential, and easily overlooked problem in current technology is that existing preoperative visual function expectation assessments for cataract surgery lack the ability to reduce patients' natural language expectations to specific visual task tolerance thresholds. This causes the true expectations to be blurred and distorted during the acquisition stage, ultimately failing to provide accurate and structured demand basis for subsequent assessment and decision-making. Summary of the Invention

[0005] To overcome the aforementioned deficiencies of the prior art, embodiments of the present invention provide a method for assessing preoperative visual function expectations in cataract surgery. This method collects preoperative basic examination data, subjective expectation expression data, and visual task scenario interaction discrimination results from cataract patients. It then fuses, analyzes, quantifies, and structures the patient's visual need types and tolerance for failure consequences in different visual task scenarios, generating an assessment result that represents the patient's true preoperative visual function expectations. This addresses the problem mentioned in the background art that existing preoperative visual function expectation assessments for cataract surgery cannot accurately restore the patient's natural language expectations to specific visual task tolerance thresholds, causing the true expectations to be blurred and distorted during the data collection stage, thus failing to provide accurate and structured demand basis for subsequent assessment and decision-making.

[0006] To achieve the above objectives, the present invention provides the following technical solution: a method for assessing preoperative visual function expectations in cataract surgery, comprising the following steps: S1. Collect preoperative basic examination data, subjective expectation expression data and preset visual task scene library of cataract patients, and select a set of candidate visual task scenes that match the patient's current visual ability from the preset visual task scene library based on the preoperative basic examination data. S2. Output each visual task scene in the candidate visual task scene set to the patient in turn for acceptance interaction judgment, and receive the patient's judgment results of whether each visual task scene is acceptable, unacceptable, or barely acceptable, to obtain the scene acceptance result set. S3. The subjective expectation expression data and the scene acceptance result set are fused and analyzed to identify the visual needs of patients in different visual task scenarios, and generate a set of visual function needs parameters including near resolution needs, low light recognition needs, glare tolerance needs and visual switching recovery needs. S4. Based on the visual function requirement parameter set, the patient's tolerance for failure consequences in each visual task scenario is quantitatively characterized, and a set of visual task tolerance thresholds corresponding to each visual task scenario is formed. S5. Based on the visual task tolerance threshold set, the patient's preoperative visual function expectation is structured and encoded to obtain the expectation feature vector representing the patient's true preoperative visual function expectation. S6. Generate the expected visual function assessment results of cataract patients before surgery based on the expected feature vector, and output the assessment results to the preoperative lens matching, preoperative communication prompts or postoperative satisfaction risk prediction links.

[0007] As a further embodiment of the present invention, a preset visual task scene library is screened based on preoperative basic examination data to obtain a candidate visual task scene set corresponding to the patient's current visual function status, including the following steps: S11. Input the preoperative basic examination data into the visual function element extraction module to extract the basic visual function parameter set that represents the patient's current visual state. S12. Match and calculate the basic visual function parameter set with the scene requirement parameters corresponding to each visual task scene in the preset visual task scene library to obtain the initial adaptation results corresponding to each visual task scene. S13. Based on the initial fit results, identify visual task scenarios that are significantly mismatched with the patient's current visual state, and remove the identified mismatched visual task scenarios from the preset visual task scenario library to obtain the initial set of visual task scenarios. S14. Perform visual load stratification calculation on each visual task scene in the initial screening visual task scene set to obtain the corresponding scene load level results. S15. Based on the basic visual function parameter set, perform identifiable boundary correction on the scene load level results to obtain the boundary correction results corresponding to each visual task scene. S16. Based on the boundary correction results, retain visual task scenes that are within the range that the patient's current visual ability can perceive and that are discriminative, and obtain a candidate visual task scene set. S17. Output the candidate visual task scene set to the subsequent patient acceptance interaction discrimination stage to perform preoperative visual function expectation acquisition.

[0008] In sections S11-S17, it is necessary to explain how to progressively select a set of candidate visual task scenarios from a pre-set visual task scenario library based on the preoperative baseline examination data of cataract patients. These scenarios should both match the patient's current visual ability level and effectively distinguish the boundaries of their expected true visual function. The "preoperative baseline examination data" preferably includes uncorrected visual acuity, corrected visual acuity, refractive status, corneal astigmatism parameters, contrast sensitivity, glare-related examination results, pupillary parameters, axial length parameters, and examination results related to low-light visual performance. The "basic visual function parameter set" refers to the set of parameters selected from the aforementioned preoperative baseline examination data. The data is a set of parameters extracted and standardized to characterize the patient's current visual ability. The "preset visual task scenario library" refers to a set of scenarios that are pre-established and cover typical life tasks such as reading small print, recognizing road signs at night, identifying object boundaries in low light, observing targets in highly reflective environments, and switching between indoor and outdoor environments. Each visual task scenario is pre-configured with corresponding "scenario requirement parameters". The "scenario requirement parameters" are used to characterize the degree of requirements of the visual task scenario on visual distance resolution, environmental light adaptation, glare tolerance, contrast recognition, and visual switching recovery.

[0009] Specifically, in S11, the preoperative basic examination data is first input into the visual function element extraction module. The original examination results from different sources and with different dimensions are cleaned, normalized, and feature-organized. Then, key parameters that can reflect the patient's current visual state are extracted according to dimensions such as far-field resolution, near-field resolution, low-light identification, glare tolerance, and visual switching recovery. This results in a basic visual function parameter set, so that subsequent matching calculations do not directly rely on messy original data, but rely on a unified parameter expression that is comparable and computable. In S12, the basic visual function parameter set is matched and calculated one by one with the scene requirement parameters corresponding to each visual task scene in the preset visual task scene library. The so-called "matching calculation" refers to comparing the degree of closeness, excess or deficiency between the patient's current visual ability and the scene visual requirements in order to obtain the initial fit result of each visual task scene relative to the patient. The initial fit result can be understood as the basic fit level of the patient when encountering the scene in the current visual state, which is used to reflect whether the scene is obviously too difficult, obviously too easy or in a reasonable range that can be further evaluated for the patient. In S13, visual task scenarios that are significantly mismatched with the patient's current visual state are identified based on the initial fit results. "Significant mismatch" preferably includes two types of situations: one is an overly difficult scenario where the scenario requirements are far higher than the patient's current visual ability and the patient is almost certain to be unable to complete it; the other is an overly easy scenario where the scenario requirements are significantly lower than the patient's current visual ability and cannot reflect the difference in the patient's expectations. Then, these two types of visual task scenarios are removed from the preset visual task scenario library to obtain the initial set of visual task scenarios. The purpose is to remove invalid scenarios first to avoid interference from extreme scenarios in subsequent interaction judgments. In S14, visual load stratification calculation is further performed on each visual task scene in the initial screening visual task scene set. "Visual load" refers to the intensity of the comprehensive visual ability that the patient needs to call when completing the corresponding visual task. It can be comprehensively evaluated from factors such as target size, target distance, illumination conditions, background contrast, glare intensity and visual switching frequency. Through this stratification calculation, the scene load level results corresponding to each visual task scene can be obtained, which can clarify the position of each initial screening scene in the overall difficulty system, so that subsequent screening is not just roughly retained, but further positioned in the difficulty sequence. In S15, the scene load level results are corrected for discernible boundaries based on the basic visual function parameter set. "Discernible boundary" refers to the boundary range of the patient's ability to make a difference judgment of "acceptable", "barely acceptable" or "unacceptable" under the current visual state. In specific execution, it is combined with the patient's current visual ability parameters to determine whether a scene is medium or high in terms of scene load level, but whether it has exceeded the range of perceptible difference for the patient, or whether it is low in load but may still be near the patient's sensitive boundary. Then, the original scene load level results are corrected to fit the individual state of the patient to obtain the boundary correction results corresponding to each visual task scene. In S16, based on the boundary correction results, visual task scenes that are within the perceptible range of the patient's current visual ability and have discriminative characteristics are retained. "Within the perceptible range" means that the scene will not cause the patient to make a uniform negative judgment because it is too difficult, nor will it cause the patient to make a uniform positive judgment because it is too easy. "Having discriminative characteristics" means that the scene can effectively induce the patient to produce meaningful differential feedback between acceptable, unacceptable, and barely acceptable, thereby finally obtaining a set of candidate visual task scenes. The essence of this step is to select the most informative scene near the boundary of the patient's true visual ability. In S17, the candidate visual task scene set is output to the subsequent patient acceptance interaction judgment stage to perform preoperative visual function expectation acquisition. That is, the scenes that have been individually screened and boundary corrected are used as formal interactive input content and presented to the patient on the terminal device for acceptance judgment. This ensures that the subsequent acquisition is not random feedback on irrelevant scenes, but effective feedback on the critical zone of the patient's real visual ability and the difference point of real expectation.

[0010] Through the step-by-step execution of S11 to S17, a continuous computational process is actually completed, starting from the original examination data, through capability extraction, scene matching, extreme elimination, load stratification, boundary correction and candidate retention. This transforms the originally generalized life scene screening into an individualized candidate visual task scene generation process that can directly support the subsequent preoperative visual function expectation acquisition.

[0011] As a further embodiment of the present invention, each visual task scene in the candidate visual task scene set is output to the patient for acceptability judgment, and the patient's acceptance results for each visual task scene are collected to obtain a scene judgment result set corresponding to each visual task scene, including the following steps: S21. Input each visual task scene in the candidate visual task scene set into the scene presentation module, and generate interactive scene display content according to the visual requirement parameters corresponding to each visual task scene to obtain the scene sequence to be judged. S22. Output the sequence of scenes to be judged to the patient in sequence according to the preset interference isolation rules to perform the acceptability judgment operation, and restrict the interference of non-target visual cues during the output of each visual task scene to obtain the patient's initial judgment input for each visual task scene; S23. Collect response delay and monitor discrimination switching of the initial discrimination input to obtain the interaction behavior results corresponding to each visual task scenario; S24. Perform consistency verification between the initial judgment input and the interaction behavior results, identify visual task scenarios with hesitation, repeated switching or abnormal delay, and obtain a set of scenarios to be reviewed. S25. Perform differential enhancement and re-presentation operation on each visual task scene in the scene set to be reviewed, and collect the patient's supplementary discrimination input for the scene set to be reviewed again to obtain the review discrimination result; S26. The initial discrimination input and the verification discrimination result are fused and calculated to generate the final acceptance result that corresponds to each visual task scene. S27. Based on the final acceptance results, establish a scene discrimination result set corresponding to the candidate visual task scene set.

[0012] In sections S21-S27, it is necessary to explain how to transform the candidate visual task scene set obtained in the previous stage into an interactive discrimination process that the patient can actually participate in. Furthermore, based on the patient's subjective choices and combined with their interactive behavior characteristics, the acceptance results of each visual task scene are verified, reviewed, and fused to form a scene discrimination result set that can truly reflect the patient's preoperative visual function expectation boundaries. Here, the "scene presentation module" refers to the functional module used to transform the candidate visual task scenes into an interactive interface that the patient can directly view, compare, and operate. "Visual requirement parameters" refer to parameters that correspond one-to-one with each visual task scene and are used to describe... This scenario is a set of parameters that define the degree of requirements for visual distance, illumination adaptation, glare tolerance, contrast recognition, and visual switching. "Interactive scene display content" refers to the display content generated based on the above visual requirement parameters, which allows patients to make a direct judgment on the terminal that it is acceptable, unacceptable, or barely acceptable. "Initial discrimination input" refers to the direct acceptance level selection made by the patient when viewing the corresponding visual task scene for the first time. "Interactive behavior results" refers to the behavioral information collected synchronously around this acceptance level selection, including at least interactive features that reflect the stability of the patient's judgment, such as response latency, number of discrimination switching, dwell time, and modification actions.

[0013] Specifically, in S21, each visual task scene in the candidate visual task scene set is first input into the scene presentation module. Then, the scene presentation module configures the target object, target size, display distance, background brightness, local contrast, glare simulation degree and screen switching method in the scene according to the visual requirement parameters corresponding to each visual task scene, thereby generating interactive scene display content that matches the visual task scene. The scene to be judged is formed according to the preset presentation order, so that the patient is not facing an abstract task name, but a specific interactive scene that can be directly observed and judged. In S22, the sequence of scenes to be judged is output to the patient sequentially according to preset interference isolation rules to perform an acceptability judgment operation. The "preset interference isolation rules" refer to the output rules used to control the judgment process from being interfered with by irrelevant cues. Preferred rules include presenting each scene individually, restricting direct comparison prompts between adjacent scenes, hiding text descriptions unrelated to the current task, controlling display duration and page layout consistency, etc., to avoid the patient being affected by non-target visual cues, textual inducements, or interface differences. During the output of each visual task scene, the system receives the patient's judgment operation for the scene, indicating whether it is acceptable, unacceptable, or barely acceptable, and records the first judgment result as the patient's initial judgment input for the corresponding visual task scene, thereby completing the first subjective acceptance collection for each scene. In S23, the initial discrimination input is further subjected to response delay acquisition and discrimination switching monitoring. "Response delay acquisition" refers to recording the time interval between the completion of the visual task scene presentation and the patient's acceptance selection. "Discrimination switching monitoring" refers to recording whether the patient changes the option, hesitates back and forth, clicks repeatedly, or cancels and reselects during a discrimination process. These behavioral information are associated with the corresponding initial discrimination input to obtain the interactive behavior results corresponding to each visual task scene. This allows the system to not only see what the patient has selected, but also to determine whether the selection is stable and whether there is hesitation or uncertainty. In S24, the initial discrimination input and the interactive behavior result are checked for consistency. "Consistency check" refers to comparing the patient's subjective choice result with the stability of the behavior when making that result to determine whether the acceptance choice has sufficient credibility. For example, if the patient selects "acceptable" but the response time is significantly longer, the option is switched multiple times during the discrimination process, or the patient shows significant inconsistency in similar scenarios, then the initial discrimination result corresponding to that scenario can be determined to have unstable characteristics. Based on this, visual task scenarios with hesitation, repeated switching, or abnormal delays are identified and classified into the scenario set to be reviewed, thereby distinguishing potentially distorted initial judgments from relatively stable initial judgments. In S25, a difference enhancement and re-presentation operation is performed on each visual task scene in the set of scenes to be reviewed. "Difference enhancement and re-presentation" refers to expressing the key visual requirements of the scene more clearly and with more defined boundaries without changing the essence of the original visual task. For example, appropriately enhancing the difference between the target and the background, highlighting the difference in the strength of glare interference, widening the recognition difficulty gradient between close-range details and low-light conditions, or reconstructing adjacent difficulty versions in sequence to help patients more clearly distinguish their true acceptable boundaries. After the difference enhancement and re-presentation is completed, the patient's supplementary discrimination input for the set of scenes to be reviewed is collected again, and the result of this second input is recorded as the review discrimination result, thereby obtaining the patient's secondary acceptance feedback under the review conditions. In S26, the initial discrimination input and the verification discrimination result are fused and calculated. "Fusion calculation" refers to combining the first discrimination result, the second verification result, and the aforementioned interaction behavior results to comprehensively determine the final acceptance state of each visual task scene. When the initial discrimination input and the verification discrimination result are consistent, the consistent result can be directly used as the final acceptance result of the scene. When the two are inconsistent, the result that better reflects the patient's true judgment boundary is given a higher credibility weight by combining the corresponding response delay, the number of switching, and the stability after verification, thereby generating a final acceptance result that corresponds one-to-one with each visual task scene. In S27, a scene discrimination result set corresponding to the candidate visual task scene set is established based on the final acceptance result. That is, each visual task scene is associated with its final determined acceptance state to form a standardized result set required for subsequent fusion and analysis of subjective demand expression and objective scene feedback.

[0014] Through the step-by-step execution of S21 to S27, a continuous processing process is actually completed, from scene content generation, controlled output, initial discrimination acquisition, behavior stability monitoring, consistency verification, review of difficult scenes, to final result fusion and database construction. This ensures that the obtained scene discrimination result set is not a rough result formed by a single random click, but an effective discrimination result that can more realistically reflect the patient's preoperative visual function expectation boundary after interactive behavior constraints and review and correction.

[0015] As a further embodiment of the present invention, the subjective expectation expression data and the scene acceptance result set are fused and analyzed to identify the visual need types corresponding to patients in different visual task scenarios, and a set of visual function need parameters including near resolution need, low light recognition need, glare tolerance need, and visual switching recovery need is generated, including the following steps: S31. Input the subjective expectation expression data into the semantic parsing module to perform demand semantic decomposition, extract the subjective demand semantic units related to visual distance, ambient illuminance, light interference and scene switching, and obtain the subjective demand semantic set. S32. Input the scene acceptance result set into the scene back-inference module, and perform back-mapping calculation on the patient acceptance result according to the task attribute parameters corresponding to each visual task scene to obtain the objective demand indication set corresponding to each visual task scene. S33. Input the set of subjective needs semantics and the set of objective needs indications into the needs alignment module to perform semantic consistency matching and conflict item identification, and obtain the set of candidate visual needs types corresponding to patients in different visual task scenarios. S34. Perform demand intensity decomposition calculation on the candidate visual demand type set to obtain the sub-demand intensity values ​​corresponding to near-range resolution demand, low-light recognition demand, glare tolerance demand and visual switching recovery demand respectively. S35. Based on the intensity values ​​of sub-items, perform aggregate analysis on the common and differential demand parts among different visual task scenarios to obtain the visual function demand correlation structure corresponding to the patient. S36. Based on the visual functional requirement association structure, parameterize the intensity values ​​of the sub-requirements to generate a visual functional requirement parameter set. S37. Output the set of visual function requirement parameters to the subsequent visual task tolerance threshold quantification calculation stage to perform preoperative visual function expectation assessment.

[0016] In sections S31-S37, it is necessary to explain how to homogenize, make comparable, and structurally integrate the patient's subjective expectation expression data with the aforementioned scenario acceptance result set. This allows for the gradual restoration of implicit needs, originally scattered throughout natural language descriptions and scenario interaction feedback, into a set of visual function requirement parameters that directly represent the patient's true preoperative visual function expectations. Here, "subjective expectation expression data" refers to the original expression content related to preoperative visual goals formed by the patient during consultations, questionnaires, verbal descriptions, or interactive inputs. "Subjective requirement semantic unit" refers to the smallest semantic fragment extracted from this original expression content that can independently represent the meaning of a certain type of visual need. This definition is used because the patient's original expression is often mixed with life scenarios, vague evaluations, and personal habits, and cannot be easily understood. To directly participate in subsequent calculations, it must first be decomposed into identifiable and categorizable semantic components; the "scene acceptance result set" refers to the set of final acceptance states of patients for multiple candidate visual task scenarios; the "objective demand indication set" refers to the demand orientation results derived from the task attribute parameters of each visual task scenario and the patient's acceptance state in the corresponding scenario, used to characterize the strength of the patient's true preference for different visual ability dimensions from a behavioral perspective; the "candidate visual demand type set" refers to the set of demand types that can be used for further intensity decomposition and encoding after mutual verification between subjective semantics and objective indications; the "visual function demand parameter set" is the result set after parameterizing the patient's near resolution demand, low-light recognition demand, glare tolerance demand, and visual switching recovery demand.

[0017] Specifically, in S31, the subjective expectation expression data is first input into the semantic parsing module. The descriptive words, scene words, and evaluative words in the patient's original expression are segmented into sentences, words, and semantically classified. Subjective demand semantic units related to visual distance, ambient illuminance, light interference, and scene switching are extracted from them. For example, expressions such as "want to see the small print on the phone clearly", "don't have blurry vision when looking at road signs at night", and "hope to see clearly as soon as possible after entering the room from the outside" are decomposed into semantic needs for near-distance resolution, low-light recognition, and visual switching recovery direction. These decomposition results are then organized into a set of subjective demand semantics, so that the patient's verbal expression is first transformed into an analyzable demand semantic basis. In S32, the scene acceptance result set is input into the scene back-inference module. Combined with the pre-configured task attribute parameters for each visual task scene, the system performs a back-mapping calculation on the patient's acceptable, unacceptable, or barely acceptable results in each scene. The so-called "back-mapping calculation" refers to inferring the patient's true needs and tendencies in the corresponding visual ability dimension of the scene from the patient's judgment results of the specific scene. For example, if the patient consistently rejects low-contrast scenes at night, it can be inferred that he or she has high requirements for low-light recognition and contrast sensitivity-related abilities; if he or she shows obvious unacceptability to strong reflective scenes, it can be inferred that he or she has high requirements for glare tolerance. In this way, an objective demand indication set corresponding to each visual task scene is obtained, so that the system not only knows "what the patient said", but also knows "what the patient actually prefers in the scene". In S33, the subjective demand semantic set and the objective demand indication set are jointly input into the demand alignment module to perform semantic consistency matching and conflict item identification. "Semantic consistency matching" refers to judging whether the direction of demand presented by the patient in verbal expression is consistent with the direction of demand deduced from the scene behavior. "Conflict item identification" refers to identifying demand items where there is a significant deviation between the patient's verbal expression and scene response. For example, if the patient verbally states that "the requirements for seeing things at night are not high", but continues to make unacceptable judgments in low-light scenes, then this item is marked as a conflict demand. After completing consistency matching and conflict item identification, a set of candidate visual demand types corresponding to the patient in different visual task scenarios is obtained, that is, a set of demand type candidate results that have been verified by both subjective and objective factors. In S34, the demand intensity decomposition calculation is performed on the candidate visual demand type set. The so-called "demand intensity decomposition calculation" means that it no longer stops at the coarse-grained judgment of "whether the patient has this type of demand", but further decomposes the degree of requirement of each candidate visual demand type in different dimensions, and forms sub-demand intensity values ​​corresponding to near resolution demand, low light recognition demand, glare tolerance demand, and visual switching recovery demand. Each sub-demand intensity value can be understood as a quantitative expression of the patient's emphasis, sensitivity, or tolerance level requirements for the corresponding visual ability dimension. In S35, the common and differential demand components among different visual task scenarios are aggregated and analyzed based on the intensity values ​​of the sub-items of demand. The so-called "common demand component" refers to the same type of visual function demand that multiple visual task scenarios point to together. For example, multiple near reading and mobile phone viewing scenarios all point to a high near resolution demand. The so-called "differential demand component" refers to special demands that only appear significantly in specific scenarios. For example, glare tolerance demand that is prominent only in night driving-related scenarios. By aggregating and analyzing these common and differential components, the visual function demand correlation structure corresponding to the patient can be obtained, which is used to reflect the degree of correlation, primary and secondary relationship and scenario dependence between different demand dimensions. In S36, the intensity values ​​of sub-items of visual function requirements are parameterized and encoded according to the visual function requirement association structure. "Parameterized encoding" means that the aforementioned requirement intensity and its association are transformed into a parameter expression in a unified format according to the preset requirement dimensions and encoding rules. This makes near resolution requirement, low light recognition requirement, glare tolerance requirement, and visual switching recovery requirement form corresponding storable, callable, and comparable parameter items, thereby generating a visual function requirement parameter set. This allows subsequent steps to directly call these parameters to quantify the patient's tolerance for the consequences of visual task failure. In S37, the set of visual function requirement parameters is output to the subsequent visual task tolerance threshold quantification calculation stage to perform preoperative visual function expectation assessment. That is, the requirement parameters that have been completed by subjective and objective fusion, intensity decomposition and structured coding are used as the direct input of downstream calculation, so that the subsequent calculation no longer depends on the patient's original verbal expression or scattered scene discrimination results, but on a set of standardized parameters that can truly reflect the characteristics of the patient's preoperative visual function requirement.

[0018] Through the step-by-step execution of S31 to S37, a continuous fusion and analysis process is actually completed, from natural language expectation decomposition, scenario result inference, alignment of subjective and objective needs, conflict identification, demand intensity decomposition, cross-scenario association aggregation to parameterized output. This transforms the patient's originally vague, scattered and easily distorted subjective expectations into a set of visual function demand parameters that can directly support the subsequent preoperative visual function expectation assessment.

[0019] As a further embodiment of the present invention, the patient's tolerance for failure consequences in various visual task scenarios is quantitatively characterized based on a set of visual function requirement parameters, and a set of visual task tolerance thresholds corresponding one-to-one with each visual task scenario is formed, including the following steps: S41. Perform association and matching calculations between the visual function requirement parameter set and the scene task attribute parameters corresponding to each visual task scenario to obtain the requirement function parameter set corresponding to each visual task scenario. S42. Based on the set of demand parameters, the degree of functional impairment, the degree of impact on life, and the degree of subjective rejection of patients when they fail visual tasks in various visual task scenarios are quantitatively calculated to obtain the failure consequence characterization value corresponding to each visual task scenario. S43. Perform scene sensitivity correction calculation on the failure consequence representation value to eliminate the consequence representation shift caused by the difference in scene expression between different visual task scenarios, and obtain the corrected consequence representation value corresponding to each visual task scenario. S44. Based on the correction consequence characterization value, calculate the tolerance boundary for each visual task scenario to obtain the initial tolerance threshold for the patient for each visual task scenario. S45. Perform cross-scenario consistency constraint processing on the initial tolerance threshold to identify and correct abnormal thresholds that significantly deviate from the overall visual function requirements of the patient, and obtain the corrected tolerance threshold. S46. Summarize and encode the modified tolerance thresholds corresponding to each visual task scenario to form a visual task tolerance threshold set that corresponds one-to-one with each visual task scenario. S47. Output the visual task tolerance threshold set to the subsequent preoperative visual function expectation structured assessment stage to perform the representation of the patient's true expectation boundary.

[0020] In sections S41-S47, it is necessary to explain how, based on the aforementioned visual function requirement parameter set, the patient's tolerance for "task failure" in different visual task scenarios is further calculated, and this tolerance is progressively quantified into a set of visual task tolerance thresholds that can be directly used for subsequent structured assessments. Here, the "visual function requirement parameter set" refers to the set of parameters formed after fusing and analyzing subjective expectation expression data with scenario acceptance result sets, used to characterize the intensity and correlation of the patient's actual needs in dimensions such as near-field resolution, low-light identification, glare tolerance, and visual switching recovery. The "scenario task attribute parameters" refer to the pre-configured task characteristic parameters for each visual task scenario, preferably including target observation distance, target... The attributes of size, ambient illumination level, background contrast, glare intensity, observation duration, task switching frequency, and the impact of task failure on specific daily life behaviors are used to describe the degree of dependence of each visual task scenario on different visual function dimensions and the cost of failure. The "tolerance for failure consequences" is defined in this way because what patients are really concerned about is not a certain visual ability indicator itself, but whether they can accept the failure result and its impact on their lives if they cannot see clearly, cannot see steadily, or recover too slowly in a specific daily life task. The "visual task tolerance threshold" is the result of parameterizing this acceptable boundary, used to indicate the degree to which patients can tolerate the consequences of failure in each visual task scenario.

[0021] Specifically, in S41, the visual function requirement parameter set is first correlated and matched with the scene task attribute parameters corresponding to each visual task scenario. The so-called "correlation and matching calculation" refers to the item-by-item correspondence and coupling analysis of the intensity of the patient's needs in different visual function dimensions and the degree of dependence of each visual task scenario on the corresponding ability. For example, the patient's high near resolution needs are combined with the near distance, small target, and long duration attributes in the "reading small text" scenario, and the patient's high glare tolerance needs are combined with the high glare and high contrast change attributes in the "nighttime car headlight interference" scenario. In this way, the requirement action parameter set corresponding to each visual task scenario is obtained, that is, it is clear which requirement dimensions play a dominant role and how strong their role is in each scenario. In S42, based on the set of demand-related parameters, the degree of functional impairment, the degree of impact on daily life, and the degree of subjective rejection are quantitatively calculated for patients experiencing visual task failures in various visual task scenarios. "Degree of functional impairment" refers to the extent to which the patient's objective visual function cannot meet the task requirements in that scenario when the task fails, such as whether the target is completely unrecognizable, only partially recognizable, or the recognition speed is significantly reduced. "Degree of impact on daily life" refers to the degree of inconvenience, risk, or efficiency loss caused by the task failure to the patient's specific daily life behaviors; for example, reading failure only results in a decrease in reading efficiency, while failure to recognize road signs at night may directly affect travel safety. "Degree of subjective rejection" refers to the patient's psychological intolerance of this type of failure result based on their own demand intensity. These three sub-items are defined separately because failure in the same scenario involves objective ability gaps, differences in life consequences, and is also influenced by the patient's personal demand sensitivity. Only by calculating these three separately can we more realistically depict the patient's tolerance for failure results. In practice, the quantitative results of the above three sub-items can be obtained for each visual task scenario, and a corresponding failure consequence representation value can be further formed to uniformly represent the overall unacceptability of task failure in that scenario for the patient. In S43, the failure consequence representation value is calculated for scene sensitivity correction. "Scene sensitivity correction" refers to correcting the consequence representation deviation caused by differences in expression form, presentation method or task appearance between different visual task scenarios. For example, although some scenarios have similar objective needs, they are more likely to be given higher weight by patients because the scene picture is more intuitive and closer to the patient's life experience, while other scenarios may be underestimated because the expression is more abstract. Therefore, it is necessary to combine the scene task attribute parameters, the aforementioned patient acceptance behavior characteristics, and the comparison relationship between similar scenarios to normalize and correct the deviation of the failure consequence representation value, so as to obtain the corrected consequence representation value corresponding to each visual task scenario, so that the failure consequence representation between different scenarios is comparable. In S44, the tolerance boundary is calculated for each visual task scenario based on the correction consequence characterization value. The so-called "tolerance boundary calculation" refers to further determining the critical position from "acceptable failure" to "unacceptable failure" based on the patient's unacceptability of the failure consequence after correction. In other words, it is to find the upper or lower limit of the patient's tolerance for failure in the corresponding visual task scenario. In implementation, the initial tolerance threshold corresponding to each visual task scenario can be determined by combining the aforementioned demand action parameter set and correction consequence characterization value. This initial tolerance threshold can be understood as the patient's initial failure tolerance boundary in that scenario. In S45, the initial tolerance threshold is subjected to cross-scene consistency constraint processing. "Cross-scene consistency constraint processing" refers to comparing the initial tolerance thresholds obtained from multiple visual task scenarios within the same patient's overall visual function needs framework to identify abnormal thresholds that significantly deviate from the patient's overall needs. For example, if a patient exhibits low tolerance in multiple low-light-related scenarios but only shows an abnormally high tolerance threshold in a similar scenario, it can be determined that the initial tolerance threshold for that scenario may be affected by accidental feedback, expression bias, or scenario-specific noise, and needs to be corrected. During execution, the initial tolerance threshold can be identified, smoothed, and corrected based on the patient's overall visual function needs parameter set, visual function needs correlation structure, and threshold distribution of similar scenarios, thereby obtaining a corrected tolerance threshold. This makes the final threshold result more in line with the patient's overall true needs rather than being affected by fluctuations in a single scenario. In S46, the modified tolerance thresholds corresponding to each visual task scenario are summarized and encoded to form a set of visual task tolerance thresholds that correspond one-to-one with each visual task scenario. Here, "summarization encoding" means that the modified thresholds of each scenario are organized into a standardized result set according to unified scenario identification, requirement dimension identification and threshold representation rules, so that they can be directly called, compared and structured for subsequent processing, rather than being retained as scattered single-scenario calculation results. Through this step, the visual task tolerance threshold set is finally obtained, which is a set of parameters that can characterize the patient's tolerance boundary for failure consequences on a scenario-by-scenario basis. In S47, the visual task tolerance threshold set is output to the subsequent preoperative visual function expectation structured assessment stage to perform the representation of the patient's true expectation boundary. That is, the standardized threshold results that have been completed in the previous function matching, failure consequence quantification, sensitivity correction, boundary solution and consistency correction are used as the direct input for subsequent structured coding and expectation feature vector generation. This makes the system represent not the strength of abstract needs, but the true tolerance boundary of the patient determined by calculation at the specific visual task level.

[0022] Through the step-by-step execution of S41 to S47, a continuous calculation process is actually completed, from coupling demand parameters and scene attributes, to the quantification of failure consequences, to scene offset correction, tolerance boundary solution, cross-scene anomaly correction, and finally the output of the threshold summary. This transforms the patient's subjective acceptable boundary for failure results of different visual tasks into a set of visual task tolerance thresholds that can be directly used for the subsequent preoperative visual function expectation structured assessment.

[0023] As a further embodiment of the present invention, the patient's preoperative visual function expectation is structured and encoded according to a visual task tolerance threshold set to obtain an expected feature vector representing the patient's true preoperative visual function expectation, including the following steps: S51. Input the visual task tolerance threshold set into the threshold parsing module, and group and map it according to the visual distance attribute, ambient illumination attribute, light interference attribute and scene switching attribute corresponding to the visual task scene to obtain a multidimensional threshold distribution set. S52. Perform weight normalization calculation on the thresholds of each dimension in the multidimensional threshold distribution set to obtain a standardized threshold parameter set that represents the importance of different visual needs. S53. Based on the standardized threshold parameter set, the threshold items that repeatedly represent the same visual function requirement across visual task scenarios are aggregated and compressed to obtain a redundant requirement parameter set. S54. Perform conflict identification and priority decision calculation on the contradictory threshold items in the redundancy removal requirement parameter set to obtain the consistency requirement parameter set. S55. Based on the consistency requirement parameter set, establish the coding correspondence between visual requirement dimensions and threshold intensity to obtain the structured coding result of the patient's preoperative visual function expectation. S56. Perform vectorization and dimensional ordering on the structured coding results to generate an expected feature vector that represents the patient's true preoperative visual function expectations. S57. Output the expected feature vector to the preoperative plan matching, risk warning or satisfaction prediction stage to perform subsequent evaluation processing.

[0024] In sections S51-S57, it is necessary to explain how to further transform the aforementioned visual task tolerance threshold set from "threshold results scattered according to scenarios" into "structured encoded results uniformly expressed according to the dimension of needs," and finally generate expected feature vectors that can be directly used for preoperative plan matching, risk warning, or satisfaction prediction. Here, the "visual task tolerance threshold set" refers to the set of boundary parameters for the patient's tolerance of failure consequences for each visual task scenario. Essentially, it reflects the patient's acceptable upper limit for visual task failure in different scenarios. However, this result is still stored in a scattered manner with scenarios as the index and cannot be directly used for cross-scenario comparison and downstream model calls. Therefore, it is necessary to reorganize these scattered thresholds into a unified parameter expression centered around the dimension of visual needs through structured encoding.

[0025] Specifically, in S51, the visual task tolerance threshold set is input into the threshold parsing module, and grouped and mapped according to the visual distance attribute, ambient illuminance attribute, light interference attribute and scene switching attribute corresponding to each visual task scene. In other words, the threshold results originally recorded according to specific scenes such as "reading small print", "reading road signs at night", "observing targets with strong reflections", and "switching between indoor and outdoor", are re-merged into the requirement dimensions such as "near distance resolution", "low illuminance identification", "glare tolerance" and "visual switching recovery", thereby obtaining a multi-dimensional threshold distribution set, so that multiple scene thresholds on the same requirement dimension can be processed in the same analysis framework. In S52, the thresholds of each dimension in the multidimensional threshold distribution set are weighted and normalized. The so-called "weight normalization calculation" means that by combining the representativeness of each visual task scene in the patient's life, the importance of the corresponding demand dimension of the scene, and the sensitivity of the threshold itself to the patient's true expectation boundary, a uniform and comparable weight is assigned to the threshold results from different sources. Then, the thresholds of each dimension are scaled to eliminate the differences between the number of different scenes, different threshold dimensions, and different scene influences. Finally, a standardized threshold parameter set representing the importance of different visual needs is obtained, so that subsequent encoding is based on comparable standardized results rather than the original discrete thresholds. In S53, threshold items that repeatedly represent the same visual function requirement across visual task scenarios are aggregated and compressed based on a standardized threshold parameter set. This step is necessary because multiple different scenarios may point to the same visual function requirement. For example, "looking at small text on a mobile phone," "reading instructions on a medicine box," and "reading subtitles" may all reflect near-field resolution requirements. If aggregation is not performed, the subsequent encoding results will repeatedly amplify the influence of the same requirement dimension. Therefore, in this step, threshold items with repeated content, consistent orientation, and similar threshold change trends are merged and compressed to obtain a redundant requirement parameter set, so that each type of visual function requirement is represented by a more refined and representative parameter set. In S54, conflict identification and priority determination calculation are performed on the contradictory threshold items in the redundancy requirement parameter set. The so-called "contradictory threshold items" refer to the thresholds mapped from different scenarios under the same visual function requirement dimension, which show obvious inconsistencies or even opposite boundary characteristics. For example, patients show significant conflicts of high tolerance and low tolerance in two similar low-light scenarios. At this time, it is necessary to combine the aforementioned scenario representativeness, patient interaction stability, scenario sensitivity correction results, and the task weight of the scenario in real life to compare the credibility of the conflict items and determine their priority. The threshold expression that is more representative of the patient's true expectation boundary is retained, thereby obtaining a consistent requirement parameter set, so that subsequent encoding is based on the requirement parameters with consistent internal logic. In S55, a coding correspondence between visual requirement dimensions and threshold intensity is established based on the consistency requirement parameter set. That is, each requirement dimension is configured with a corresponding parameter position, parameter meaning, and threshold intensity expression rule, so that near resolution requirement, low light recognition requirement, glare tolerance requirement, and visual switching recovery requirement have clear coding entry and intensity representation method, thereby obtaining the structured coding result of the patient's preoperative visual function expectation. The role of this structured coding result is to transform "in which aspects the patient has higher requirements and in which aspects the patient has lower tolerance" into a unified, clear, and computable coding form. In S56, the structured coding results are vectorized and ordered by dimensions. "Vectorized arrangement" means that each coding result is arranged into a parameter sequence of fixed length and fixed order according to the pre-set requirement dimension order and parameter position rules. "Dimension ordering" means that the same requirement dimension always occupies a consistent position and consistent expression format among different patients, so that subsequent models or rule engines can directly call and compare it, thereby generating the expected feature vector that represents the patient's true preoperative visual function expectation. In S57, the expected feature vector is output to the preoperative plan matching, risk warning or satisfaction prediction stage to perform subsequent evaluation processing. That is, the results of the aforementioned grouping mapping, normalization, redundancy removal, conflict resolution, structured coding and vectorization processing are used as the standard input of the downstream decision module, so that the subsequent stages can directly perform lens matching constraints, communication focus identification or postoperative satisfaction risk analysis based on the expected feature vector.

[0026] Through the step-by-step execution of S51 to S57, a continuous processing process is actually completed, transforming the scenario-based threshold results into a demand-dimensional, standardized, deredundant, consistent, and vectorized expression. This transforms the patient's real but scattered preoperative visual function expectation boundary into an expectation feature vector that can directly support subsequent clinical auxiliary decision-making.

[0027] As a further embodiment of the present invention, the preoperative visual function expectation assessment result of cataract patients is generated based on the expected feature vector, and the assessment result is output to the preoperative lens matching, preoperative communication prompts, or postoperative satisfaction risk prediction stages, including the following steps: S61. Input the expected feature vector into the assessment calculation module, and perform fractal analysis calculation on the expected feature vector in combination with the preset visual function expected assessment rules to obtain the patient's corresponding sub-item expected assessment value. S62. Based on the sub-item expectation assessment values, the patient’s expected intensity and tolerance boundary in the dimensions of near resolution, low light recognition, glare tolerance and visual switching recovery are comprehensively calculated to obtain a comprehensive expectation assessment parameter set. S63. Based on the comprehensive expectation assessment parameter set, the patient's preoperative visual function expectation is stratified and determined to obtain expectation type results including at least high expectation sensitivity type, scene preference concentration type, strict tolerance threshold type or relatively balanced type; S64. Based on the expected type results and the comprehensive expected assessment parameter set, perform application adaptation calculations to obtain triage assessment results corresponding to the preoperative lens matching stage, the preoperative communication prompting stage, and the postoperative satisfaction risk prediction stage, respectively. S65. Mark and identify high-conflict, high-risk, and high-sensitivity items in the diversion assessment results to obtain assessment prompts. S66. Associate and encapsulate the assessment prompts with the triage assessment results to generate the expected preoperative visual function assessment results for cataract patients. S67. Direct the preoperative visual function expectation assessment results of cataract patients to the preoperative lens matching, preoperative communication prompts, or postoperative satisfaction risk prediction stages to implement corresponding clinical auxiliary decision processing.

[0028] Sections S61-S67 explain how, based on the aforementioned expected feature vector, patients' preoperative visual function expectations are ultimately assessed, categorized, and assigned for specific uses. The assessment results are then transformed into clinical auxiliary decision-making outputs that directly support preoperative lens matching, preoperative communication, and postoperative satisfaction risk prediction. Here, the "expected feature vector" refers to a standardized parameter sequence obtained by structured encoding and vectorization of a visual task tolerance threshold set. Its function is to represent, in a unified format, the patient's true preoperative visual function expectations in dimensions such as near resolution, low-light identification, glare tolerance, and visual switching recovery. The "sub-item expected assessment value" is also relevant. "Expectation Assessment Parameter Set" refers to the single-dimensional assessment result obtained after further analyzing the expected feature vector according to the demand dimension, which is used to reflect the patient's expected level and tolerance boundary characteristics in each visual function dimension; "Comprehensive Expectation Assessment Parameter Set" refers to the summary parameter set formed on the basis of each sub-item expected assessment value for overall decision-making; "Expectation Type Result" is the result after classifying the patient's preoperative visual function expected characteristics, used to identify which expectation mode the patient belongs to; "Triage Assessment Result" refers to the dedicated assessment output generated for different clinical application stages; "Assessment Prompt Result" refers to the prompt information after marking risk items or conflict items that need to be focused on.

[0029] Specifically, in S61, the expected feature vector is input into the evaluation calculation module, and the expected feature vector is analyzed and calculated in a dimensionality-based manner in combination with the preset visual function expectation evaluation rules. The so-called "preset visual function expectation evaluation rules" refers to a set of rules that are pre-set to interpret the meaning of the needs represented by each vector dimension and its strength boundary. Preferably, these rules include threshold interval division rules for different needs dimensions, rules for interpreting combinations between dimensions, and rules for judging needs sensitivity. Through this step, the original expected feature vector is decomposed and calculated according to dimensions such as near resolution, low light recognition, glare tolerance, and visual switching recovery to obtain the patient's corresponding sub-item expectation evaluation values. This allows the system to first clarify the patient's needs status in each specific visual function dimension, rather than directly giving a general overall evaluation. In S62, the patient's expected intensity and tolerance boundary in the dimensions of near resolution, low light recognition, glare tolerance, and visual switching recovery are comprehensively calculated based on the sub-item expected assessment values. "Expected intensity" is used to characterize the patient's attention to and requirements for a certain visual function dimension, and "tolerance boundary" is used to characterize the patient's acceptable range of failure consequences related to that dimension. By jointly analyzing the sub-item results of each dimension, a comprehensive expected assessment parameter set can be obtained. This parameter set is no longer just a simple stacking of multiple single-point results, but can reflect the comprehensive expression of the patient's overall expected structure, the primary and secondary relationships of needs, and key low-tolerance areas. In S63, patients' preoperative visual function expectations are stratified based on a comprehensive expectation assessment parameter set. "Stratification" refers to classifying patients into clinically interpretable expectation categories according to parameter performance. For example, when the expectation intensity of one or more dimensions is significantly high and the tolerance boundary is significantly narrow, the patient can be identified as a high expectation-sensitive type; when the demand is highly concentrated in a few specific life scenarios, the patient can be identified as a scenario-preference-concentrated type; when the failure tolerance thresholds corresponding to multiple scenarios are generally low, the patient can be identified as a strict tolerance threshold type; when the demand distribution of each dimension is relatively balanced and there are few extreme tendencies, the patient can be identified as a relatively balanced type. This yields the expectation type results, allowing for targeted treatment in subsequent clinical stages based on patient type. In S64, the purpose-fit calculation is performed based on the expected type results and the comprehensive expected assessment parameter set. The so-called "purpose-fit calculation" refers to mapping the assessment results of the same patient to three application scenarios: preoperative lens matching, preoperative communication prompts, and postoperative satisfaction risk prediction. The parameters that need the most attention in each application scenario are extracted. For example, in the preoperative lens matching stage, the focus is on the intensity of demand and low tolerance boundary related to different visual tasks such as near, intermediate, and far vision. In the preoperative communication prompt stage, the focus is on the explanation of patients with high sensitivity, high conflict, and high expectations. In the postoperative satisfaction risk prediction stage, the focus is on risk dimensions that are prone to subjective dissatisfaction, such as low light identification, glare tolerance, and visual switching recovery. Thus, the triage assessment results corresponding to the above three stages are obtained respectively. In S65, high-conflict, high-risk, and high-sensitivity items in the triage assessment results are marked and identified. "High-conflict items" refer to items where there is significant tension between different dimensions of needs or between needs and application goals. "High-risk items" refer to items that are more likely to cause dissatisfaction or mismatch after surgery. "High-sensitivity items" refer to items where patients have high subjective attention, low tolerance boundaries, and may have adverse experiences even with slight deviations. By automatically identifying and marking these items, assessment prompts are obtained, so that the final output not only includes general results but also highlights the key points that need to be prioritized in clinical practice. In S66, the assessment prompts and triage assessment results are associated and encapsulated to generate the expected assessment results of visual function before cataract surgery. The so-called "associated encapsulation" means that triage results for different purposes are integrated with corresponding key prompts according to a unified data structure, so that each output result contains the corresponding explanation of needs, risk warnings and application directions, thereby forming a complete assessment result that can be directly displayed or called on the clinical terminal. In S67, the preoperative visual function expectation assessment results of cataract patients are directed to the preoperative lens matching, preoperative communication prompts, or postoperative satisfaction risk prediction stages to execute the corresponding clinical auxiliary decision processing. That is, the assessment results that have been completed by dimensional analysis, comprehensive evaluation, type determination, purpose triage, key marking, and result encapsulation are sent to the corresponding application modules. The lens matching module adjusts the matching constraints and recommendation priorities accordingly, the preoperative communication module generates key information to be communicated, and the satisfaction risk prediction module outputs potential dissatisfaction risk prompts accordingly.

[0030] Through the step-by-step execution of S61 to S67, a continuous processing procedure is actually completed, from expected feature vector parsing to comprehensive evaluation, type classification, usage mapping, key prompt generation, and directional output. This transforms the patient's true preoperative visual function expectation into a preoperative visual function expectation assessment result for cataract surgery that can directly support clinical auxiliary decision-making.

[0031] As a further embodiment of the present invention, S161, each visual task scene in the candidate visual task scene set is a perturbation scene dynamically constructed based on the patient's individual visual ability boundary. The generation of the perturbation scene includes the following steps: S1611. Input the preoperative basic examination data and candidate visual task scene set into the scene parameter generation module, and extract the corresponding target distance parameters, target size parameters, ambient illumination parameters, background contrast parameters, glare interference parameters and viewing angle switching parameters for each candidate visual task scene to obtain the scene basic parameter set. S1612. Based on the basic parameter set of the scene and the patient's current visual ability boundary, perform single-parameter progressive perturbation calculation on each candidate visual task scene to obtain multiple scene perturbation version sets corresponding to different visual load levels. S1613. Perform multi-parameter coupling constraint screening on each scene disturbance version in the scene disturbance version set, remove abnormal combination parameters that exceed the scope of real life scenes, and obtain a set of life-oriented disturbance scenes. S1614. Based on the set of everyday perturbation scenarios, the difficulty change gradient of each visual task scenario is hierarchically sorted to obtain the boundary perturbation scenario sequence corresponding to the critical interval of the patient's visual ability. S1615. Output the boundary perturbation scene sequence to the patient acceptance interaction discrimination stage to perform dynamic scene testing against the patient's true visual function tolerance boundary.

[0032] In S1611-S1615, it is necessary to explain how to generate parameterized perturbations, screen for life-like constraints, and construct boundary sequences for the selected candidate visual task scenarios around the patient's current visual ability boundary, so as to form a dynamic perturbation scenario that can more accurately trigger the patient's true tolerance boundary response. Here, "patient's individual visual ability boundary" refers to the critical ability range determined based on preoperative basic examination data, in which the patient transitions from "can complete" to "difficult to complete" in terms of target recognition, low-light recognition, glare interference tolerance, and perspective switching adaptation. The reason for using this boundary as the basis for perturbation generation is that only when the difficulty of the scenario falls near the patient's critical ability can the subsequent interaction judgment results most accurately reflect the patient's acceptable boundary rather than extremely positive or extremely negative feedback.

[0033] Specifically, in S1611, preoperative basic examination data and candidate visual task scene set are input into the scene parameter generation module. For each candidate visual task scene, the core parameters used to characterize visual load are extracted. The target distance parameter is used to characterize the observation distance between the patient and the target object. The target size parameter is used to characterize the detail scale of the object being identified. The ambient illumination parameter is used to characterize the overall brightness and darkness conditions of the scene. The background contrast parameter is used to characterize the degree of separability between the target and the background. The glare interference parameter is used to characterize the level of interference of strong light or reflected light on target recognition. The viewing angle switching parameter is used to characterize the requirements for visual recovery ability when switching observation in different viewing directions or different spatial areas. By extracting and uniformly representing the above parameters, the scene basic parameter set is obtained, so that each candidate visual task scene is transformed into a computable and adjustable parameterized scene model. In S1612, based on the set of basic scene parameters and the patient's current visual ability boundary, single-parameter progressive perturbation calculation is performed on each candidate visual task scene. The so-called "single-parameter progressive perturbation calculation" means that, while keeping other parameters unchanged, only one basic scene parameter is adjusted step by step according to a preset step size. For example, the target size is gradually reduced, the ambient illuminance is gradually reduced, the glare interference is gradually increased, or the angle switching amplitude is gradually increased, so as to observe the impact of the change of this single variable on the scene difficulty. Multiple scene versions with progressively varying difficulty are generated for each candidate visual task scene, thus obtaining a set of scene perturbation versions corresponding to different visual load levels. The reason for adopting the single-parameter progressive method is to ensure that the source of difficulty change between different perturbation versions is clear, which makes it easier to determine which type of visual load change the patient is most sensitive to. In S1613, multi-parameter coupling constraint screening is performed on each scene perturbation version in the scene perturbation version set. "Multi-parameter coupling constraint screening" means that, based on the aforementioned single-parameter progressive perturbation, further checks whether the combination of multiple parameters in each perturbation version conforms to the objective logic of real-life scenarios. For example, although the combination of extremely low illumination, high glare, and ultra-small target size can be constructed in theory, if it significantly exceeds the range of scenarios that can occur in daily life, it will cause the test results to be distorted. Therefore, it is necessary to eliminate abnormal combination parameters that exceed the range of real-life scenarios based on the prior rules of life scenarios, clinical use conditions, or preset legal range of scenarios, so as to obtain a set of life-oriented perturbation scenarios, so that the retained perturbation scenarios have both boundary testing capabilities and do not deviate from the actual life experience of patients. In S1614, the difficulty gradient of each visual task scene is hierarchically sorted according to the set of everyday perturbation scenes. The so-called "difficulty gradient" refers to the continuous change relationship of the same visual task scene from easy to difficult or from low load to high load under different perturbation versions. By comparing the degree of closeness of each version to the patient's visual ability boundary, the perturbation versions that are closer to the transition interval between "just acceptable" and "just unacceptable" are prioritized, thereby obtaining the boundary perturbation scene sequence corresponding to the critical interval of the patient's visual ability. The purpose of this step is not to simply retain all everyday perturbation scenes, but to select the key versions that are most likely to trigger the patient's true boundary judgment, so that the subsequent interactive tests are more focused on the most informative difficulty interval. In S1615, the boundary perturbation scene sequence is output to the patient acceptance interaction discrimination stage to perform dynamic scene testing on the patient's true visual function tolerance boundary. That is, according to the aforementioned sorting results, the perturbation scenes located near the patient's ability critical zone are presented to the patient in sequence for acceptance judgment, so that the patient can make boundary feedback between different but adjacent difficulty versions, thereby improving the resolution of the subsequent discrimination results on the true tolerance threshold.

[0034] Through the step-by-step execution of S1611 to S1615, a continuous generation process is actually completed, from the extraction of basic scene parameters, single-parameter progressive perturbation, multi-parameter life-oriented constraints, boundary gradient sorting to dynamic test output. This further refines the original candidate visual task scene into a set of perturbation scene sequences dynamically constructed around the critical range of the patient's visual ability, so as to more accurately identify the patient's true preoperative visual function tolerance boundary.

[0035] As a further embodiment of the present invention, S38, when fusing and analyzing the subjective expectation expression data and the scene acceptance result set, also includes rebuttal correction of subjective expression bias, which includes the following steps: S381. Input the set of subjective demand semantics and the set of objective demand indications into the deviation recognition module, and perform a difference comparison calculation on the semantic expression results and scene discrimination results under the same visual demand dimension to obtain the demand deviation itemset. S382. Identify the patient’s underestimation, overestimation and unstable expression in different visual demand dimensions based on the demand deviation itemset to obtain the subjective expression deviation type results. S383. Perform targeted review and retrieval operations on the visual task scenarios corresponding to the subjective expression deviation type results, and generate a set of review visual task scenarios corresponding to the deviation dimension. S384. Output the visual task scene set for review to the patient to perform supplementary acceptance discrimination, and collect the supplementary discrimination results to obtain the deviation review result set; S385. Based on the deviation verification result set, perform confidence correction calculation on the deviation semantic units in the subjective demand semantic set, and perform weight enhancement calculation on the corresponding demand indicator items in the objective demand indicator set to obtain the corrected demand semantic set and the enhanced demand indicator set. S386. Re-execute the fusion parsing of the corrected requirement semantic set and the enhanced requirement indication set to obtain the corrected visual function requirement parameter set.

[0036] In S381-S386, it is necessary to explain how, based on the subjective need semantic extraction, objective need indication back-calculation, and subjective-objective need alignment already completed in S31-S37, further discern the deviation between the patient's verbal expression and scene discrimination, and obtain a corrected post-visual functional need parameter set that more closely approximates the true need boundary through a review and correction mechanism. Here, the "subjective need semantic set" corresponds to the need semantic results obtained from the decomposition of subjective expectation expression data in S31, and the "objective need indication set" corresponds to the need orientation results obtained by back-mapping based on the scene acceptability result set in S32. Therefore, this step is not... Instead of regenerating a separate set of data, the fusion analysis chain in S31-S37 is reinforced by counter-evidence. "Subjective expression bias" is defined as a situation where patients' verbal demands or intensity are inconsistent with their actual judgment in specific visual task scenarios due to habitual generalizations, insufficient awareness of their own visual limitations, unstable expectations of postoperative results, or distorted judgment of the importance of the scene. "Counter-evidence correction" uses objective scene feedback to verify and correct subjective expressions, so that the subsequently generated set of visual function demand parameters no longer relies excessively on the patient's single verbal statement.

[0037] Specifically, in S381, the set of subjective demand semantics and the set of objective demand indications are input into the deviation recognition module. The semantic expression results and scene discrimination results under the same visual demand dimension are compared and calculated. The so-called "same visual demand dimension" means that the two types of data are uniformly mapped to the same demand framework such as near resolution, low light recognition, glare tolerance, and visual switching recovery. Then, the consistency between the patient's "verbal expression of demand direction and intensity" and "actual acceptance boundary and demand tendency in the scene" is compared. For example, if the patient verbally states that the night vision requirement is average, but continuously gives unacceptable judgments in low light and glare scenes, then a difference item is formed in the corresponding dimension. Through this dimension-by-dimensional comparison, the demand deviation item set is obtained, that is, it is clear which demand dimensions have subjective and objective inconsistencies. In S382, based on the demand deviation itemset, patients are identified as exhibiting underestimation, overestimation, and unstable expression across different visual demand dimensions. "Underestimation" refers to a patient's verbal expression of demand intensity being lower than the actual demand intensity reflected in the scene feedback; "overestimation" refers to a patient's verbal expression of demand intensity being higher than the actual demand intensity reflected in the scene feedback; and "unstable expression" refers to inconsistencies in the patient's expression within the same demand dimension, or a lack of stable correspondence between verbal expression and multi-scene discrimination. By further classifying the deviation items, subjective expression deviation types can be obtained, thereby clarifying the nature of each type of deviation, rather than simply knowing that a deviation exists. In S383, a targeted review retrieval operation is performed on the visual task scene corresponding to the subjective expression deviation type result. The so-called "targeted review retrieval" means that instead of repeating the test on all scenes, the review scene that can best distinguish the true boundary of the dimension is retrieved only from the preset visual task scene library or the aforementioned candidate visual task scene set, focusing on the identified deviation dimension. For example, when the low illumination recognition dimension shows an underestimation of expression, relevant scenes under different illumination gradients and different background contrast conditions are retrieved first to generate a set of review visual task scenes corresponding to the deviation dimension, thereby ensuring that the review process focuses on the discovered conflict points. In S384, the visual task scene set for review is output to the patient to perform supplementary acceptance discrimination and the supplementary discrimination results are collected to obtain the deviation review result set. That is, the patient's acceptable, barely acceptable or unacceptable feedback is collected again in the scene directly related to the original deviation dimension, so that the system obtains a set of supplementary evidence specifically used to verify whether the original deviation really exists and whether the deviation direction is stable, thereby avoiding directly correcting the requirement parameters based on only one difference comparison. In S385, confidence correction calculations are performed on the biased semantic units in the subjective demand semantic set based on the bias review result set, and weight enhancement calculations are performed on the corresponding demand indicators in the objective demand indicator set. "Confidence correction calculation" refers to re-evaluating the credibility of the original subjective semantic units based on the review results. If the review results consistently indicate that the patient's actual response in a certain dimension is stronger than their verbal expression, the confidence of that biased semantic unit is reduced; if the review results support the original expression, the credibility of that semantic unit is retained or increased. "Weight enhancement calculation" refers to increasing the influence weight of the corresponding objective demand indicator in subsequent fusion analysis when the review results further confirm the reliability of the scenario feedback, so that the verified objective feedback occupies a more reasonable position in the subjective-objective fusion, thereby obtaining the corrected demand semantic set and the enhanced demand indicator set. In S386, the correction requirement semantic set and the enhancement requirement indication set are re-fused and parsed to obtain the corrected visual function requirement parameter set. That is, the requirement alignment, intensity decomposition, correlation aggregation and parameterized encoding logic in S33-S36 are used. However, the input is no longer the uncorrected subjective requirement semantic set and objective requirement indication set, but the correction result after deviation identification, targeted review and weight adjustment. This makes the final output visual function requirement parameter set more reflective of the patient's true preoperative visual function needs, rather than the initial result interfered with by subjective expression deviation.

[0038] Through the step-by-step execution of S381 to S386, a supplementary correction process is actually completed, which follows S31-S37: "deviation identification - deviation classification - targeted review - result verification - confidence correction - re-fusion". This improves the aforementioned fusion analysis result from "one-time subjective and objective alignment" to "corrected demand parameter output with a counter-evidence mechanism".

[0039] As a further embodiment of the present invention, when the preoperative visual function expectation assessment results of cataract patients are output to the preoperative lens matching, preoperative communication prompts, or postoperative satisfaction risk prediction stages, S68 also includes linkage decision processing based on different application stages. The linkage decision processing includes the following steps: S681. Input the preoperative visual function expectation assessment results of cataract patients into the usage allocation module. Based on the expectation type results, the comprehensive expectation assessment parameter set and the assessment prompt results, generate the matching constraint parameter set corresponding to the preoperative lens matching stage, the notification key parameter set corresponding to the preoperative communication prompt stage, and the risk warning parameter set corresponding to the postoperative satisfaction risk prediction stage. S682. Input the matching constraint parameter set into the lens scheme filtering module, perform constraint matching calculations on each candidate lens scheme in the preset lens scheme library, and obtain the lens adaptation ranking results. S683. Input the set of key parameters into the communication prompt generation module, extract the explanatory content and prioritize the prompts for high-conflict items, high-sensitivity items and low-tolerance threshold items, and obtain the preoperative communication prompt results. S684. Input the risk warning parameter set into the satisfaction risk prediction module, perform sub-item prediction calculations on the patient's postoperative dissatisfaction risk in the dimensions of near resolution, low light recognition, glare tolerance and visual switching recovery, and obtain the postoperative satisfaction risk result. S685. Link and integrate the lens fitting ranking results, preoperative communication prompts, and postoperative satisfaction risk results to generate linked auxiliary decision-making results for the same patient. S686. Output the results of the linkage-assisted decision-making to the clinical terminal to implement preoperative lens matching suggestions, key points of preoperative communication, and postoperative satisfaction risk warning.

[0040] In S681-S686, it is necessary to explain how, based on the expected feature vector analysis, comprehensive evaluation, expected type determination, application triage, and evaluation result encapsulation already completed in S61-S67, the preoperative visual function expected evaluation results of cataract patients can be further transformed into linked decision-making outputs for different clinical application stages. This allows the same evaluation result to simultaneously support preoperative lens matching, preoperative communication prompts, and postoperative satisfaction risk prediction. Among them, the "preoperative visual function expected evaluation results of cataract patients" corresponds to the result generated in S66 and output in S67. It already contains the expected type result, the comprehensive expected evaluation parameter set, and the evaluation prompt result. However, this result is still a standardized output at the overall evaluation level and needs to be further transformed into decision parameters that can be directly called upon by each specific application stage. The reason why "linked decision-making processing" is defined in this way is that it does not process lens matching, communication prompts, and risk warnings separately and in isolation. Instead, it is based on the unified expected evaluation results of the same patient, and simultaneously constrains, prompts, and warns the three stages to avoid inconsistencies in decision-making.

[0041] Specifically, in S681, the preoperative visual function expectation assessment results of cataract patients are input into the usage allocation module. Based on the expectation type results, the comprehensive expectation assessment parameter set, and the assessment prompt results, core constraint information applicable to different application stages is extracted. Among them, the "matching constraint parameter set" refers to the set of parameters used to limit or screen lens options, preferably including the patient's key needs and low tolerance boundaries in near resolution, low light recognition, glare tolerance, and visual switching recovery dimensions; the "information key parameter set" refers to the set of parameters used to generate preoperative communication content, preferably including high conflict items, high sensitivity items, and expectation deviation items that require special explanation; the "risk warning parameter set" refers to the set of parameters used to predict the risk of postoperative subjective dissatisfaction, preferably including low tolerance dimensions and high-risk scenarios that are more likely to cause dissatisfaction after surgery. This completes the first allocation from unified assessment results to dedicated usage parameters. In S682, the matching constraint parameter set is input into the lens scheme screening module, and constraint matching calculation is performed on each candidate lens scheme in the preset lens scheme library. The so-called "constraint matching calculation" refers to comparing the visual function features that each candidate lens scheme can provide with the needs and tolerance boundaries reflected in the patient's assessment results item by item. For example, it is determined whether a certain scheme meets the patient's needs in terms of near vision compensation, low light condition adaptation, glare-related performance and visual switching expectation, and identifies whether the scheme touches the risk points corresponding to the patient's low tolerance boundary. Then, the suitability ranking of each candidate lens scheme is formed, thereby obtaining the lens suitability ranking result, so that the matching suggestion is based on the patient's true expected structure, rather than just based on conventional refractive parameters or experience judgment. In S683, the key parameter set is input into the communication prompt generation module. The module extracts and prioritizes the explanatory content for high-conflict items, high-sensitivity items, and low-tolerance threshold items. "High-conflict items" refer to items where there is significant tension between different patient needs, requiring focused explanation of trade-offs before surgery. "High-sensitivity items" refer to items where patients have a high level of subjective attention, and even slight deviations can cause dissatisfaction. "Low-tolerance threshold items" refer to items where patients have a significantly narrower acceptable range for the consequences of failure in a specific visual task. By extracting the explanatory content corresponding to these items and prioritizing them based on their potential impact on postoperative satisfaction, the preoperative communication prompts are obtained. This allows doctors to focus on explaining the key points that truly affect patient satisfaction during communication, rather than simply providing a generalized overview of all content. In S684, the risk warning parameter set is input into the satisfaction risk prediction module to perform itemized prediction calculations on the patient's postoperative dissatisfaction risk in the dimensions of near resolution, low light recognition, glare tolerance, and visual switching recovery. The so-called "itemized prediction calculation" means that the patient's high demand intensity, low tolerance boundary, and the aforementioned assessment prompts in each visual function dimension are used as inputs to analyze the possibility and sensitivity of the patient to subjective dissatisfaction in the corresponding dimension, thereby obtaining the postoperative satisfaction risk result. This result is not a generalized single risk level, but clearly indicates which dimensions, which life scenarios, or which expected differences are more likely to cause dissatisfaction for the patient. In S685, the lens fitting ranking results, preoperative communication prompts, and postoperative satisfaction risk results are linked and integrated to generate a linked auxiliary decision-making result for the same patient. The so-called "linked integration" means that the three types of outputs are organized in a unified manner according to the needs logic of the same patient, so that each recommended plan not only corresponds to the fitting ranking, but also simultaneously displays the communication focus and potential satisfaction risks related to the plan. For example, although a certain lens plan has a high matching degree, there are still risks that need to be highlighted in terms of low light recognition or glare tolerance. This information will be presented in the integrated result, thereby ensuring that the matching suggestions, communication content, and risk judgments correspond to each other and are not separated. In S686, the results of the linkage-assisted decision-making are output to the clinical terminal to implement preoperative lens matching suggestions, key points of preoperative communication, and postoperative satisfaction risk warnings. In other words, the results that have been allocated, screened, generated, predicted, and integrated are presented on the clinical terminal used by doctors in a callable, viewable, and interactive form, so that doctors can simultaneously complete the recommendation of the plan, the notification of key points, and the risk warning based on the same linkage-assisted decision-making results.

[0042] Through the step-by-step execution of S681 to S686, a continuous processing process is actually completed, which takes the preoperative visual function expectation assessment results generated by S61-S67 as the upstream input, and extends to three downstream application links: lens matching, communication prompts, and satisfaction risk prediction. This process further transforms the aforementioned assessment results into linked clinical auxiliary decision-making results that can be directly implemented.

[0043] The technical effects and advantages of this invention are as follows: 1. The core technical effect of this solution lies in the fusion and analysis of the patient's natural language expectations, scene interaction feedback and preoperative examination data, and further quantification into visual task tolerance thresholds, correcting the "coarse-grained demand input" problem from the source, and substantially improving the authenticity and structure of preoperative expectation assessment.

[0044] 2. By first screening out scenes that are too difficult, too easy, or lack discrimination, and then retaining candidate scenes around the patient's visual ability boundary, this approach focuses subsequent data acquisition on the task range that best reveals the differences in actual expectations, thereby improving the information density and individual fit of scene sampling.

[0045] 3. By performing consistency verification and fusion calculation on the patient's initial discrimination input superimposed with response delay, discrimination switching, and review and re-presentation results, this solution reduces the random deviation caused by a single subjective click, making the scene acceptance result closer to the patient's stable and reproducible true judgment boundary.

[0046] 4. By aligning subjective demand semantics with objective demand indications, identifying conflicts, decomposing intensity, and parameterizing encoding, this scheme can restore scattered and ambiguous everyday demands into calculable demand parameters such as near-field resolution, low-light identification, glare tolerance, and visual switching recovery, thereby enhancing the medical interpretability of the assessment results.

[0047] 5. By quantifying the consequences of failure in different visual tasks, correcting for scene sensitivity, and constraining cross-scene consistency, the tolerance threshold obtained by this scheme is not an accidental result in an isolated scene, but a stable boundary that conforms to the overall needs structure of the patient. Therefore, it is more suitable as a technical basis for subsequent evaluation and matching.

[0048] 6. By further stratifying and distributing the expected feature vector to lens matching, preoperative communication prompts, and postoperative satisfaction risk prediction, this approach achieves the linkage support of the same assessment result for multiple clinical stages, which helps to reduce the deviation in matching the approach and the gap between postoperative expectations and reality. Attached Figure Description

[0049] Figure 1 This is a flowchart outlining the method steps of the present invention. Detailed Implementation

[0050] The following detailed description, in conjunction with embodiments of the present invention, provides a method for assessing preoperative visual function expectations in cataract surgery. It should be understood that the following specific embodiments are for illustrative purposes only and are not intended to limit the scope of protection of the present invention. Without departing from the concept of the present invention, those skilled in the art can make adjustments or substitutions to the specific implementations, all of which should fall within the scope of protection of the present invention.

[0051] In this embodiment, refer to Figure 1 A method for assessing visual function expectations before cataract surgery, comprising the following steps: S1. Collect preoperative basic examination data, subjective expectation expression data and preset visual task scene library of cataract patients, and select a set of candidate visual task scenes that match the patient's current visual ability from the preset visual task scene library based on the preoperative basic examination data. S2. Output each visual task scene in the candidate visual task scene set to the patient in turn for acceptance interaction judgment, and receive the patient's judgment results of whether each visual task scene is acceptable, unacceptable, or barely acceptable, to obtain the scene acceptance result set. S3. The subjective expectation expression data and the scene acceptance result set are fused and analyzed to identify the visual needs of patients in different visual task scenarios, and generate a set of visual function needs parameters including near resolution needs, low light recognition needs, glare tolerance needs and visual switching recovery needs. S4. Based on the visual function requirement parameter set, the patient's tolerance for failure consequences in each visual task scenario is quantitatively characterized, and a set of visual task tolerance thresholds corresponding to each visual task scenario is formed. S5. Based on the visual task tolerance threshold set, the patient's preoperative visual function expectation is structured and encoded to obtain the expectation feature vector representing the patient's true preoperative visual function expectation. S6. Generate the expected visual function assessment results of cataract patients before surgery based on the expected feature vector, and output the assessment results to the preoperative lens matching, preoperative communication prompts or postoperative satisfaction risk prediction links.

[0052] In this embodiment, the preoperative basic examination data preferably includes uncorrected visual acuity, corrected visual acuity, refractive status, corneal astigmatism parameters, contrast sensitivity, glare-related examination results, pupillary parameters, axial length parameters, and low-light visual performance-related examination results. Subjective expectation expression data preferably comes from outpatient consultation records, subjective questionnaires filled out by patients, verbal descriptions by patients, or terminal interactive input content; the preset visual task scenario library preferably includes visual task scenarios closely related to the daily visual experience of cataract patients after surgery, such as reading small print, checking mobile phone information, recognizing road signs at night, identifying object boundaries in low-light environments, observing targets in highly reflective environments, and switching between indoor and outdoor observation. Each visual task scenario is pre-configured with corresponding scenario requirement parameters, which are used to characterize the degree of requirement of the visual task scenario for visual distance resolution ability, environmental illumination adaptation ability, glare tolerance ability, contrast recognition ability, and visual switching recovery ability.

[0053] In this embodiment, the first step is to perform a screening process for the candidate visual task scene set. Specifically, preoperative basic examination data is input into the visual function element extraction module. The original examination results from different sources and with different dimensions are cleaned, normalized, and feature-organized. Key parameters that can reflect the patient's current visual state are extracted according to dimensions such as far-field resolution, near-field resolution, low-light identification, glare tolerance, and visual switching recovery to obtain a basic visual function parameter set that characterizes the patient's current visual state. Then, the basic visual function parameter set is matched and calculated one by one with the scene requirement parameters corresponding to each visual task scene in the preset visual task scene library to compare the degree of closeness, excess, or deficiency between the patient's current visual ability and the visual requirements of the corresponding scene, thereby obtaining the initial fit result corresponding to each visual task scene. Based on the initial fit results, visual task scenarios that are significantly mismatched with the patient's current visual state are identified, and the identified mismatched visual task scenarios are removed from the preset visual task scenario library to obtain a preliminary set of visual task scenarios. Among them, the mismatched visual task scenarios preferably include overly difficult scenarios where the scenario requirements are much higher than the patient's current visual ability and overly easy scenarios where the scenario requirements are significantly lower than the patient's current visual ability. Subsequently, visual load stratification calculation is performed on each visual task scenario in the preliminary set of visual task scenarios to obtain the corresponding scenario load level results. Among them, the visual load is preferably determined by the target size, target distance, illumination conditions, background contrast, glare intensity, and visual switching frequency.

[0054] Furthermore, based on the basic visual function parameter set, the scene load level results are subjected to identifiable boundary correction to obtain the boundary correction results corresponding to each visual task scene. The identifiable boundary refers to the range of the patient's ability to distinguish between "acceptable," "barely acceptable," and "unacceptable" in their current visual state. Finally, based on the boundary correction results, visual task scenes that are within the patient's current perceptible range and possess discriminative qualities are retained to obtain a candidate visual task scene set. This candidate visual task scene set is then output to the subsequent patient acceptance interaction judgment stage. This completes the individualized screening process from raw examination data to a candidate visual task scene set.

[0055] In this embodiment, the candidate visual task scene set is not a static, fixed set of scenes, but rather a dynamically generated perturbation scene around the patient's individual visual ability boundary. Specifically, preoperative basic examination data and the candidate visual task scene set are input into the scene parameter generation module. Target distance parameters, target size parameters, ambient illumination parameters, background contrast parameters, glare interference parameters, and viewing angle switching parameters are extracted for each candidate visual task scene to obtain a basic scene parameter set. Then, based on the basic scene parameter set and combined with the patient's current visual ability boundary, a single-parameter progressive perturbation calculation is performed on each candidate visual task scene. That is, while keeping other parameters unchanged, only one basic scene parameter is adjusted step-by-step according to a preset step size to obtain multiple perturbation version sets of scenes corresponding to different visual load levels.

[0056] Next, multi-parameter coupling constraint screening is performed on each scene perturbation version in the scene perturbation version set to remove abnormal parameter combinations that exceed the range of real-life scenarios, thus obtaining a set of life-oriented perturbation scenes. Then, based on this set, the difficulty change gradient of each visual task scene is hierarchically sorted to obtain a boundary perturbation scene sequence corresponding to the critical interval of the patient's visual ability. This boundary perturbation scene sequence is then output to the patient acceptance interaction discrimination stage to perform dynamic scene testing targeting the patient's true visual function tolerance boundary. Through this process, subsequent interaction discrimination can be more focused on the difficulty interval that best reveals the patient's true expectation boundary.

[0057] In this embodiment, after obtaining the candidate visual task scene set, a patient acceptance interaction discrimination process is performed. Specifically, each visual task scene in the candidate visual task scene set is input into the scene presentation module, and interactive scene display content is generated according to the visual requirement parameters corresponding to each visual task scene to obtain a sequence of scenes to be discriminated. The scene presentation module is preferably deployed in an outpatient terminal, tablet terminal, or other visual interactive terminal to transform abstract scenes into specific visual task scenes that patients can directly view and operate. Then, the sequence of scenes to be discriminated is output to the patient in sequence according to a preset interference isolation rule to perform an acceptance discrimination operation. During the output of each visual task scene, interference from non-target visual cues is restricted to obtain the patient's initial discrimination input for each visual task scene. The preset interference isolation rule preferably includes measures such as presenting scenes one by one, restricting direct comparison prompts between adjacent scenes, hiding explanatory content that is not related to the current task, and controlling the consistency of the display layout.

[0058] Next, response latency and discrimination switching are collected from the initial discrimination input to obtain the interaction behavior results corresponding to each visual task scenario. The interaction behavior results preferably include interaction features reflecting the stability of the patient's judgment, such as response latency, number of discrimination switching, dwell time, and modification actions. Further, the consistency between the initial discrimination input and the interaction behavior results is verified to identify visual task scenarios exhibiting hesitation, repeated switching, or abnormal delays, resulting in a set of scenarios to be reviewed. Subsequently, a differential enhancement re-presentation operation is performed on each visual task scenario in the set of scenarios to be reviewed, and supplementary discrimination input from the patient for the set of scenarios to be reviewed is collected again to obtain the reviewed discrimination results.

[0059] The difference enhancement and re-presentation operation optimization aims to provide clearer boundary expressions for key visual requirement dimensions without altering the essence of the original visual task. Finally, the initial discrimination input and the verification discrimination result are fused and calculated to generate a final acceptance result corresponding to each visual task scene. Based on the final acceptance result, a scene discrimination result set corresponding to the candidate visual task scene set is established. Thus, a complete interactive discrimination process is completed from scene presentation to result verification and then to the establishment of the scene discrimination result set.

[0060] In this embodiment, after obtaining the scene discrimination result set, the subjective expectation expression data and the scene acceptance result set are fused and analyzed. Specifically, the subjective expectation expression data is input into the semantic analysis module for demand semantic decomposition, extracting subjective demand semantic units related to visual distance, ambient illumination, light interference, and scene switching to obtain a subjective demand semantic set. The semantic analysis module is preferably used to segment, word, and semantically classify descriptive words, scene words, and evaluative words in the patient's original expression. Then, the scene acceptance result set is input into the scene back-inference module, which performs reverse mapping calculation on the patient's acceptance results according to the task attribute parameters corresponding to each visual task scene to obtain an objective demand indication set corresponding to each visual task scene. Subsequently, the subjective demand semantic set and the objective demand indication set are input into the demand alignment module to perform semantic consistency matching and conflict item identification to obtain a candidate visual demand type set corresponding to the patient in different visual task scenes.

[0061] Furthermore, the candidate visual demand type set is decomposed and calculated to obtain sub-demand intensity values ​​corresponding to near resolution demand, low-light recognition demand, glare tolerance demand, and visual switching recovery demand, respectively. Then, based on the sub-demand intensity values, the common and differential demand components between different visual task scenarios are aggregated and analyzed to obtain the visual function demand correlation structure corresponding to the patient. Finally, the sub-demand intensity values ​​are parameterized and encoded according to the visual function demand correlation structure to generate a visual function demand parameter set, which is then output to the subsequent visual task tolerance threshold quantification calculation stage. Through the above steps, the patient's originally vague and scattered subjective expectation expression can be transformed into a calculable, callable, and comparable standardized visual function demand parameter set.

[0062] In this embodiment, to further reduce the impact of patients' subjective expression bias on the fusion analysis results, the process of fusing and analyzing subjective expectation expression data and scene acceptance result set may also include a process of rebuttal correction of subjective expression bias. Specifically, the subjective demand semantic set and the objective demand indicator set are input into the bias identification module, and the semantic expression results and scene discrimination results under the same visual demand dimension are compared and calculated to obtain the demand bias itemset. Then, based on the demand deviation itemset, the system identifies the patient's underestimation, overestimation, and unstable expression across different visual demand dimensions to obtain subjective expression deviation type results. Next, a targeted review retrieval operation is performed on the visual task scenarios corresponding to the subjective expression deviation type results to generate a review visual task scenario set corresponding to the deviation dimension. This review visual task scenario set is then output to the patient for supplementary acceptability discrimination to collect the supplementary discrimination results, thus obtaining a deviation review result set. Based on the deviation review result set, confidence correction calculations are performed on the deviation semantic units in the subjective demand semantic set, and weight enhancement calculations are performed on the corresponding demand indicator items in the objective demand indicator set to obtain a corrected demand semantic set and an enhanced demand indicator set. Finally, the corrected demand semantic set and the enhanced demand indicator set are re-fused and parsed to obtain a corrected visual function demand parameter set. Thus, a counter-evidence reinforcement mechanism is introduced on the basis of the original alignment of subjective and objective demands, making the output results closer to the patient's true preoperative visual function needs.

[0063] In this embodiment, after obtaining the visual function requirement parameter set, the patient's tolerance for failure consequences in each visual task scenario is quantitatively characterized based on the visual function requirement parameter set. Specifically, the visual function requirement parameter set is correlated and matched with the scenario task attribute parameters corresponding to each visual task scenario to obtain the requirement effect parameter set corresponding to each visual task scenario. Subsequently, based on the requirement effect parameter set, the degree of functional impairment, the degree of impact on life, and the degree of subjective rejection corresponding to the patient's visual task failure in each visual task scenario are quantitatively calculated separately to obtain the failure consequence characterization value corresponding to each visual task scenario. Among them, the degree of functional impairment is used to characterize the degree to which visual function cannot meet the task requirements, the degree of impact on life is used to characterize the degree of inconvenience, risk, or efficiency loss caused by task failure to specific life behaviors, and the degree of subjective rejection is used to characterize the degree of non-acceptance of this type of failure result by the patient based on the intensity of their own needs. Then, scenario sensitivity correction calculation is performed on the failure consequence characterization value to eliminate the consequence characterization bias caused by the difference in scenario expression forms between different visual task scenarios, and the corrected consequence characterization value corresponding to each visual task scenario is obtained.

[0064] Furthermore, the tolerance boundaries for each visual task scenario are calculated based on the corrected consequence representation values ​​to obtain the initial tolerance thresholds for each visual task scenario. Then, cross-scenario consistency constraints are applied to the initial tolerance thresholds to identify and correct abnormal thresholds that significantly deviate from the overall visual function requirements of the patient, thereby obtaining the corrected tolerance thresholds. Finally, the corrected tolerance thresholds corresponding to each visual task scenario are summarized and encoded to form a set of visual task tolerance thresholds that correspond one-to-one with each visual task scenario. The set of visual task tolerance thresholds is then output to the subsequent preoperative structured assessment of visual function expectations. Through the above processing, the acceptable boundaries of the patient's failure results for different visual tasks can be expressed in the form of standardized thresholds.

[0065] In this embodiment, after obtaining the visual task tolerance threshold set, the patient's preoperative visual function expectations are structured and encoded according to the visual task tolerance threshold set. Specifically, the visual task tolerance threshold set is input into the threshold parsing module, and grouped and mapped according to the visual distance attribute, ambient illumination attribute, light interference attribute, and scene switching attribute corresponding to the visual task scene to obtain a multidimensional threshold distribution set. Then, the thresholds of each dimension in the multidimensional threshold distribution set are weighted and normalized to obtain a standardized threshold parameter set representing the importance of different visual needs. Subsequently, the threshold items that repeatedly represent the same visual function needs across visual task scenes are aggregated and compressed according to the standardized threshold parameter set to obtain a redundant requirement parameter set. Finally, conflict identification and priority adjudication calculations are performed on the contradictory threshold items in the redundant requirement parameter set to obtain a consistent requirement parameter set.

[0066] Furthermore, based on the consistent requirement parameter set, a coding correspondence between visual requirement dimensions and threshold intensity is established to obtain the structured coding result of the patient's preoperative visual function expectation. Finally, the structured coding result is vectorized and ordered by dimensions to generate the expected feature vector representing the patient's true preoperative visual function expectation. The expected feature vector is then output to the preoperative plan matching, risk warning, or satisfaction prediction stage. Through the above steps, the threshold results that were originally stored separately according to the scenario can be converted into a standardized parameter sequence that is uniformly expressed according to the requirement dimension.

[0067] In this embodiment, after obtaining the expected feature vector, a preoperative visual function expectation assessment result for cataract patients is generated based on the expected feature vector. Specifically, the expected feature vector is input into the assessment calculation module, and the expected feature vector is analyzed by fractal analysis in combination with preset visual function expectation assessment rules to obtain the corresponding sub-item expectation assessment value for the patient. Then, based on the sub-item expectation assessment value, the patient's expectation intensity and tolerance boundary in the dimensions of near resolution, low light recognition, glare tolerance, and visual switching recovery are comprehensively calculated to obtain a comprehensive expectation assessment parameter set. Subsequently, the patient's preoperative visual function expectation is stratified and determined according to the comprehensive expectation assessment parameter set to obtain expectation type results that include at least high expectation sensitivity type, scene preference concentration type, strict tolerance threshold type, or relatively balanced type.

[0068] Furthermore, based on the expected type results and the comprehensive expected assessment parameter set, a usage adaptation calculation is performed to obtain triage assessment results corresponding to the preoperative lens matching stage, the preoperative communication prompting stage, and the postoperative satisfaction risk prediction stage, respectively. Then, high-conflict, high-risk, and high-sensitivity items in the triage assessment results are labeled and identified to obtain assessment prompt results. Finally, the assessment prompt results and triage assessment results are associated and encapsulated to generate the preoperative visual function expectation assessment results for cataract patients. These results are then directed to the preoperative lens matching, preoperative communication prompting, or postoperative satisfaction risk prediction stages to execute corresponding clinical auxiliary decision processing. Thus, the patient's true preoperative visual function expectations can be transformed into assessment outputs that can directly serve clinical auxiliary decision-making.

[0069] In this embodiment, when the preoperative visual function expectation assessment results of cataract patients are output to the preoperative lens matching, preoperative communication prompts, or postoperative satisfaction risk prediction stages, further linkage decision processing based on different application stages can be performed. Specifically, the preoperative visual function expectation assessment results of cataract patients are input into the usage allocation module. Based on the expectation type results, the comprehensive expectation assessment parameter set, and the assessment prompt results, a matching constraint parameter set corresponding to the preoperative lens matching stage, a notification key parameter set corresponding to the preoperative communication prompt stage, and a risk warning parameter set corresponding to the postoperative satisfaction risk prediction stage are generated. Then, the matching constraint parameter set is input into the lens scheme screening module to perform constraint matching calculations on each candidate lens scheme in the preset lens scheme library to obtain the lens adaptation ranking results. At the same time, the notification key parameter set is input into the communication prompt generation module to classify high-conflict items, high-sensitivity items, and low-risk items. The tolerance threshold is used to extract and prioritize information to obtain preoperative communication prompts. The risk warning parameter set is then input into the satisfaction risk prediction module to calculate postoperative dissatisfaction risks in near vision, low-light perception, glare tolerance, and visual switching recovery, yielding postoperative satisfaction risk results. Subsequently, the lens fitting ranking results, preoperative communication prompts, and postoperative satisfaction risk results are integrated to generate a linked auxiliary decision-making result for the same patient. Finally, the linked auxiliary decision-making result is output to the clinical terminal to implement preoperative lens matching recommendations, key preoperative communication prompts, and postoperative satisfaction risk warnings. Through this linked decision-making process, lens matching recommendations, preoperative communication content, and postoperative risk warnings are built upon the same preoperative visual function expectation assessment result, thereby improving consistency and synergy between different clinical application stages.

[0070] In this embodiment, the method can be executed by a processor calling program instructions from memory. The processor is preferably located in a data processing system formed by a medical terminal, a dedicated assessment terminal, a cloud server, or a combination thereof. Each module, including a visual function element extraction module, a scene presentation module, a semantic parsing module, a scene reverse inference module, a demand alignment module, a deviation identification module, a threshold analysis module, an assessment calculation module, a usage allocation module, a lens treatment plan screening module, a communication prompt generation module, and a satisfaction risk prediction module, can be implemented by independent software modules or by several integrated functional modules working together. The data call relationships between modules are established sequentially based on the aforementioned step chain, thereby ensuring consistency in terminology and processing logic from the input of preoperative basic examination data to the output of linked auxiliary decision-making results.

[0071] In summary, this implementation method achieves a structured, computable, and applicable assessment of the preoperative visual function expectations of cataract patients through a continuous processing process involving candidate visual task scene set screening, patient acceptance interaction judgment, subjective and objective demand fusion analysis, visual task tolerance threshold calculation, expected feature vector generation, and multi-stage linkage decision output. This provides a unified data foundation and decision basis for preoperative lens matching, preoperative communication prompts, and postoperative satisfaction risk prediction.

[0072] The above are merely preferred embodiments of the present invention and are 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 method for assessing visual function expectations before cataract surgery, characterized in that, Includes the following steps: S1. Collect preoperative basic examination data, subjective expectation expression data and preset visual task scene library of cataract patients, and select a set of candidate visual task scenes that match the patient's current visual ability from the preset visual task scene library based on the preoperative basic examination data. S2. Output each visual task scene in the candidate visual task scene set to the patient in turn for acceptance interaction judgment, and receive the patient's judgment results of whether each visual task scene is acceptable, unacceptable, or barely acceptable, to obtain the scene acceptance result set. S3. The subjective expectation expression data and the scene acceptance result set are fused and analyzed to identify the visual needs of patients in different visual task scenarios, and generate a set of visual function needs parameters including near resolution needs, low light recognition needs, glare tolerance needs and visual switching recovery needs. S4. Based on the visual function requirement parameter set, the patient's tolerance for failure consequences in each visual task scenario is quantitatively characterized, and a set of visual task tolerance thresholds corresponding to each visual task scenario is formed. S5. Based on the visual task tolerance threshold set, the patient's preoperative visual function expectation is structured and encoded to obtain the expectation feature vector representing the patient's true preoperative visual function expectation. S6. Generate the expected visual function assessment results of cataract patients before surgery based on the expected feature vector, and output the assessment results to the preoperative lens matching, preoperative communication prompts or postoperative satisfaction risk prediction links.

2. A method for assessing preoperative visual function expectations in cataract surgery according to claim 1, characterized in that: Based on preoperative baseline examination data, a pre-set visual task scenario library is filtered to obtain a set of candidate visual task scenarios corresponding to the patient's current visual function status, including the following steps: S11. Input the preoperative basic examination data into the visual function element extraction module to extract the basic visual function parameter set that represents the patient's current visual state. S12. Match and calculate the basic visual function parameter set with the scene requirement parameters corresponding to each visual task scene in the preset visual task scene library to obtain the initial adaptation results corresponding to each visual task scene. S13. Based on the initial fit results, identify visual task scenarios that are significantly mismatched with the patient's current visual state, and remove the identified mismatched visual task scenarios from the preset visual task scenario library to obtain the initial set of visual task scenarios. S14. Perform visual load stratification calculation on each visual task scene in the initial screening visual task scene set to obtain the corresponding scene load level results. S15. Based on the basic visual function parameter set, perform identifiable boundary correction on the scene load level results to obtain the boundary correction results corresponding to each visual task scene. S16. Based on the boundary correction results, retain visual task scenes that are within the range that the patient's current visual ability can perceive and that are discriminative, and obtain a candidate visual task scene set. S17. Output the candidate visual task scene set to the subsequent patient acceptance interaction discrimination stage to perform preoperative visual function expectation acquisition.

3. A method for assessing preoperative visual function expectations in cataract surgery according to claim 2, characterized in that: The visual task scenes in the candidate visual task scene set are output to the patient for acceptability assessment, and the patient's acceptance results for each visual task scene are collected to obtain the scene discrimination result set corresponding to each visual task scene. The process includes the following steps: S21. Input each visual task scene in the candidate visual task scene set into the scene presentation module, and generate interactive scene display content according to the visual requirement parameters corresponding to each visual task scene to obtain the scene sequence to be judged. S22. Output the sequence of scenes to be judged to the patient in sequence according to the preset interference isolation rules to perform the acceptability judgment operation, and restrict the interference of non-target visual cues during the output of each visual task scene to obtain the patient's initial judgment input for each visual task scene; S23. Collect response delay and monitor discrimination switching of the initial discrimination input to obtain the interaction behavior results corresponding to each visual task scenario; S24. Perform consistency verification between the initial judgment input and the interaction behavior results, identify visual task scenarios with hesitation, repeated switching or abnormal delay, and obtain a set of scenarios to be reviewed. S25. Perform differential enhancement and re-presentation operation on each visual task scene in the scene set to be reviewed, and collect the patient's supplementary discrimination input for the scene set to be reviewed again to obtain the review discrimination result; S26. The initial discrimination input and the verification discrimination result are fused and calculated to generate the final acceptance result that corresponds to each visual task scene. S27. Based on the final acceptance results, establish a scene discrimination result set corresponding to the candidate visual task scene set.

4. A method for assessing preoperative visual function expectations in cataract surgery according to claim 3, characterized in that: The subjective expectation expression data and the scenario acceptance result set are fused and analyzed to identify the visual need types of patients in different visual task scenarios, and to generate a set of visual function need parameters including near resolution need, low light recognition need, glare tolerance need, and visual switching recovery need, including the following steps: S31. Input the subjective expectation expression data into the semantic parsing module to perform demand semantic decomposition, extract the subjective demand semantic units related to visual distance, ambient illuminance, light interference and scene switching, and obtain the subjective demand semantic set. S32. Input the scene acceptance result set into the scene back-inference module, and perform back-mapping calculation on the patient acceptance result according to the task attribute parameters corresponding to each visual task scene to obtain the objective demand indication set corresponding to each visual task scene. S33. Input the set of subjective needs semantics and the set of objective needs indications into the needs alignment module to perform semantic consistency matching and conflict item identification, and obtain the set of candidate visual needs types corresponding to patients in different visual task scenarios. S34. Perform demand intensity decomposition calculation on the candidate visual demand type set to obtain the sub-demand intensity values ​​corresponding to near-range resolution demand, low-light recognition demand, glare tolerance demand and visual switching recovery demand respectively. S35. Based on the intensity values ​​of sub-items, perform aggregate analysis on the common and differential demand parts among different visual task scenarios to obtain the visual function demand correlation structure corresponding to the patient. S36. Based on the visual functional requirement association structure, parameterize the intensity values ​​of the sub-requirements to generate a visual functional requirement parameter set. S37. Output the set of visual function requirement parameters to the subsequent visual task tolerance threshold quantification calculation stage to perform preoperative visual function expectation assessment.

5. A method for assessing preoperative visual function expectations in cataract surgery according to claim 4, characterized in that: The tolerance of patients for failure in various visual task scenarios is quantitatively characterized based on a set of visual function requirement parameters, and a set of visual task tolerance thresholds corresponding to each visual task scenario is formed. This includes the following steps: S41. Perform association and matching calculations between the visual function requirement parameter set and the scene task attribute parameters corresponding to each visual task scenario to obtain the requirement function parameter set corresponding to each visual task scenario. S42. Based on the set of demand parameters, the degree of functional impairment, the degree of impact on life, and the degree of subjective rejection of patients when they fail visual tasks in various visual task scenarios are quantitatively calculated to obtain the failure consequence characterization value corresponding to each visual task scenario. S43. Perform scene sensitivity correction calculation on the failure consequence representation value to eliminate the consequence representation shift caused by the difference in scene expression between different visual task scenarios, and obtain the corrected consequence representation value corresponding to each visual task scenario. S44. Based on the correction consequence characterization value, calculate the tolerance boundary for each visual task scenario to obtain the initial tolerance threshold for the patient for each visual task scenario. S45. Perform cross-scenario consistency constraint processing on the initial tolerance threshold to identify and correct abnormal thresholds that significantly deviate from the overall visual function requirements of the patient, and obtain the corrected tolerance threshold. S46. Summarize and encode the modified tolerance thresholds corresponding to each visual task scenario to form a visual task tolerance threshold set that corresponds one-to-one with each visual task scenario. S47. Output the visual task tolerance threshold set to the subsequent preoperative visual function expectation structured assessment stage to perform the representation of the patient's true expectation boundary.

6. A method for assessing preoperative visual function expectations in cataract surgery according to claim 5, characterized in that: The patient's preoperative visual function expectation is structured and encoded based on the visual task tolerance threshold set to obtain the expected feature vector representing the patient's true preoperative visual function expectation. This includes the following steps: S51. Input the visual task tolerance threshold set into the threshold parsing module, and group and map it according to the visual distance attribute, ambient illumination attribute, light interference attribute and scene switching attribute corresponding to the visual task scene to obtain a multidimensional threshold distribution set. S52. Perform weight normalization calculation on the thresholds of each dimension in the multidimensional threshold distribution set to obtain a standardized threshold parameter set that represents the importance of different visual needs. S53. Based on the standardized threshold parameter set, the threshold items that repeatedly represent the same visual function requirement across visual task scenarios are aggregated and compressed to obtain a redundant requirement parameter set. S54. Perform conflict identification and priority decision calculation on the contradictory threshold items in the redundancy removal requirement parameter set to obtain the consistency requirement parameter set. S55. Based on the consistency requirement parameter set, establish the coding correspondence between visual requirement dimensions and threshold intensity to obtain the structured coding result of the patient's preoperative visual function expectation. S56. Perform vectorization and dimensional ordering on the structured coding results to generate an expected feature vector that represents the patient's true preoperative visual function expectations. S57. Output the expected feature vector to the preoperative plan matching, risk warning or satisfaction prediction stage to perform subsequent evaluation processing.

7. A method for assessing preoperative visual function expectations in cataract surgery according to claim 6, characterized in that: The system generates a preoperative visual function expectation assessment result for cataract patients based on the expected feature vector, and outputs the assessment result to the preoperative lens matching, preoperative communication prompts, or postoperative satisfaction risk prediction stages, including the following steps: S61. Input the expected feature vector into the assessment calculation module, and perform fractal analysis calculation on the expected feature vector in combination with the preset visual function expected assessment rules to obtain the patient's corresponding sub-item expected assessment value. S62. Based on the sub-item expectation assessment values, the patient’s expected intensity and tolerance boundary in the dimensions of near resolution, low light recognition, glare tolerance and visual switching recovery are comprehensively calculated to obtain a comprehensive expectation assessment parameter set. S63. Based on the comprehensive expectation assessment parameter set, the patient's preoperative visual function expectation is stratified and determined to obtain expectation type results including at least high expectation sensitivity type, scene preference concentration type, strict tolerance threshold type or relatively balanced type; S64. Based on the expected type results and the comprehensive expected assessment parameter set, perform application adaptation calculations to obtain triage assessment results corresponding to the preoperative lens matching stage, the preoperative communication prompting stage, and the postoperative satisfaction risk prediction stage, respectively. S65. Mark and identify high-conflict, high-risk, and high-sensitivity items in the diversion assessment results to obtain assessment prompts. S66. Associate and encapsulate the assessment prompts with the triage assessment results to generate the expected preoperative visual function assessment results for cataract patients. S67. Direct the preoperative visual function expectation assessment results of cataract patients to the preoperative lens matching, preoperative communication prompts, or postoperative satisfaction risk prediction stages to implement corresponding clinical auxiliary decision processing.

8. A method for assessing preoperative visual function expectations in cataract surgery according to claim 2, characterized in that: S161. Each visual task scene in the candidate visual task scene set is a perturbation scene dynamically constructed based on the individual patient's visual ability boundary. The generation of the perturbation scene includes the following steps: S1611. Input the preoperative basic examination data and candidate visual task scene set into the scene parameter generation module, and extract the corresponding target distance parameters, target size parameters, ambient illumination parameters, background contrast parameters, glare interference parameters and viewing angle switching parameters for each candidate visual task scene to obtain the scene basic parameter set. S1612. Based on the basic parameter set of the scene and the patient's current visual ability boundary, perform single-parameter progressive perturbation calculation on each candidate visual task scene to obtain multiple scene perturbation version sets corresponding to different visual load levels. S1613. Perform multi-parameter coupling constraint screening on each scene disturbance version in the scene disturbance version set, remove abnormal combination parameters that exceed the scope of real life scenes, and obtain a set of life-oriented disturbance scenes. S1614. Based on the set of everyday perturbation scenarios, the difficulty change gradient of each visual task scenario is hierarchically sorted to obtain the boundary perturbation scenario sequence corresponding to the critical interval of the patient's visual ability. S1615. Output the boundary perturbation scene sequence to the patient acceptance interaction discrimination stage to perform dynamic scene testing against the patient's true visual function tolerance boundary.

9. A method for assessing preoperative visual function expectations in cataract surgery according to claim 4, characterized in that: S38. When fusing and analyzing the subjective expectation expression data and the scene acceptance result set, the process also includes rebuttal correction of subjective expression bias. Rebuttal correction includes the following steps: S381. Input the set of subjective demand semantics and the set of objective demand indications into the deviation recognition module, and perform a difference comparison calculation on the semantic expression results and scene discrimination results under the same visual demand dimension to obtain the demand deviation itemset. S382. Identify the patient’s underestimation, overestimation and unstable expression in different visual demand dimensions based on the demand deviation itemset to obtain the subjective expression deviation type results. S383. Perform targeted review and retrieval operations on the visual task scenarios corresponding to the subjective expression deviation type results, and generate a set of review visual task scenarios corresponding to the deviation dimension. S384. Output the visual task scene set for review to the patient to perform supplementary acceptance discrimination, and collect the supplementary discrimination results to obtain the deviation review result set; S385. Based on the deviation verification result set, perform confidence correction calculation on the deviation semantic units in the subjective demand semantic set, and perform weight enhancement calculation on the corresponding demand indicator items in the objective demand indicator set to obtain the corrected demand semantic set and the enhanced demand indicator set. S386. Re-execute the fusion parsing of the corrected requirement semantic set and the enhanced requirement indication set to obtain the corrected visual function requirement parameter set.

10. A method for assessing preoperative visual function expectations in cataract surgery according to claim 7, characterized in that: S68. When the preoperative visual function expectation assessment results of cataract patients are output to the preoperative lens matching, preoperative communication prompts, or postoperative satisfaction risk prediction stages, it also includes linkage decision processing based on different application stages. The linkage decision processing includes the following steps: S681. Input the preoperative visual function expectation assessment results of cataract patients into the usage allocation module. Based on the expectation type results, the comprehensive expectation assessment parameter set and the assessment prompt results, generate the matching constraint parameter set corresponding to the preoperative lens matching stage, the notification key parameter set corresponding to the preoperative communication prompt stage, and the risk warning parameter set corresponding to the postoperative satisfaction risk prediction stage. S682. Input the matching constraint parameter set into the lens scheme filtering module, perform constraint matching calculations on each candidate lens scheme in the preset lens scheme library, and obtain the lens adaptation ranking results. S683. Input the set of key parameters into the communication prompt generation module, extract the explanatory content and prioritize the prompts for high-conflict items, high-sensitivity items and low-tolerance threshold items, and obtain the preoperative communication prompt results. S684. Input the risk warning parameter set into the satisfaction risk prediction module, perform sub-item prediction calculations on the patient's postoperative dissatisfaction risk in the dimensions of near resolution, low light recognition, glare tolerance and visual switching recovery, and obtain the postoperative satisfaction risk result. S685. Link and integrate the lens fitting ranking results, preoperative communication prompts, and postoperative satisfaction risk results to generate linked auxiliary decision-making results for the same patient. S686. Output the results of the linkage-assisted decision-making to the clinical terminal to implement preoperative lens matching suggestions, key points of preoperative communication, and postoperative satisfaction risk warning.

Images