CN-121662400-B - Objective evaluation system for dynamic characteristics of pupil light reflex

Abstract

An objectification evaluation system for pupil light reflex dynamic characteristics comprises an undoubted type area and an interference area of a color pupil image, a pupil diameter-time change curve of a pupil sub-block coverage area in the undoubted type area is built, key judgment parameters of the pupil diameter-time change curve of a patient under different scene information conditions are obtained, key judgment parameter safety threshold intervals in an anesthesia depth detection mode and key judgment parameter individuation derivative parameters in a neurology follow-up mode are additionally generated, a clinical evaluation comparison table is built, dynamic parameter sets are evaluated according to the clinical evaluation comparison table, reflection grades corresponding to the dynamic parameter sets are obtained, standardized diagnosis and treatment advice and clinical early warning signals of the patient are generated according to the reflection grades, a nerve function individuation evaluation conclusion of the patient is generated according to the key judgment parameter individuation derivative parameters, and a fine reference basis is provided for early screening and illness severity judgment of diseases.

Inventors

• LI BIXIA

Assignees

• 福州大学附属省立医院

Dates

Publication Date

20260508

Application Date

20260209

Claims (6)

1. 1. The objective evaluation system for the dynamic characteristics of pupil light reflection is characterized by comprising a cloud, a detection terminal, a segmentation threshold judgment module, a data processing module, a reflection evaluation module and an evaluation early warning module; The cloud end is in communication connection with a detection terminal in a preset range, the detection terminal is used for uploading the acquired color pupil image sequence and the crowd type of the patient to the detection terminal, a segmentation threshold judgment module, a data processing module, a reflection evaluation module, an evaluation early warning module and a database are arranged in the cloud end, a plurality of block chain nodes are arranged in the database, the block chain nodes are mutually linked to form a block chain network, each block chain node is linked with a data link end, and the data link end is used for storing the dynamic parameter set and the reflection grade of the patient generated by the reflection evaluation module to the block chain nodes and marking the uplink time; The segmentation threshold judging module is used for carrying out frame-by-frame graying on the color pupil image sequence, carrying out personalized denoising on each frame of gray image, dividing the denoised gray image into a plurality of rectangular sub-blocks, obtaining segmentation threshold values of the rectangular sub-blocks, carrying out region feature recognition on the rectangular sub-blocks based on the segmentation threshold values, and obtaining an undoubted type region and an interference region, wherein the undoubted type region comprises pupil sub-blocks, iris sub-blocks and sclera sub-blocks; The data processing module is used for carrying out center positioning on the coverage area of the pupil sub-block in the gray image of the current frame and obtaining the center coordinates and the diameter of the coverage area of the pupil sub-block; Extracting the diameter of the pupil sub-block coverage area of each frame gray level image in real time, marking a corresponding timestamp, generating a pupil diameter-time change curve, Acquiring the scene information of a patient, wherein the scene information comprises an emergency rapid screening mode, a conventional detection mode, an anesthesia depth detection mode and a neurology department follow-up mode; If the patient is in the conventional detection mode, extracting the static characteristic of the pupil diameter-time change curve, acquiring a dynamic parameter set of the pupil diameter-time change curve according to the static characteristic and the pupil diameter-time change curve, and taking the static characteristic and the dynamic parameter set as key judgment parameters; If the patient is in an emergency rapid screening mode, shielding static characteristics of a pupil diameter-time change curve, extracting the maximum contraction rate and the complete recovery time in a dynamic parameter set as key judgment parameters, and additionally generating a key judgment parameter safety threshold interval in an anesthesia depth detection mode and key judgment parameter personalized derivative parameters in a neurology department follow-up mode; The reflection evaluation module is used for evaluating key judgment parameters of a patient with the scene information of an emergency rapid screening mode, a conventional detection mode or an anesthesia depth detection mode according to a clinical evaluation comparison table, and obtaining a reflection grade corresponding to the key judgment parameters, and comprises the following steps: setting key judgment parameters of a pupil diameter-time change curve and the crowd type of a patient as evaluation indexes, setting index weights of the evaluation indexes, taking a clinical evaluation comparison table as a fuzzy evaluation rule table, acquiring membership matrixes of the evaluation indexes for different reflection grades through fuzzy comprehensive evaluation, and acquiring the reflection grades corresponding to the evaluation indexes according to the membership matrixes and the index weights; the assessment early warning module is used for generating standardized diagnosis and treatment advice and clinical early warning signals of the patient according to the reflection grade, and generating a nerve function personalized assessment conclusion of the patient according to the key judgment parameter personalized derivative parameters.
2. 2. The objective assessment system for dynamic characteristics of pupillary light reflex according to claim 1, wherein the process of obtaining the undoubted type of region and the interference region comprises: Setting the size of a local area, splitting the gray image of the current frame into a plurality of rectangular sub-blocks according to the size of the local area, and acquiring the primary type of each rectangular sub-block according to the gray average value and the gray extremum difference of each rectangular sub-block; Acquiring three-dimensional core characteristics of each rectangular sub-block, acquiring gray characteristic references of different preliminary types, judging whether the three-dimensional core characteristics of each rectangular sub-block are matched with the gray characteristic references of the preliminary types of each rectangular sub-block, if so, judging the peak types of the rectangular sub-blocks according to the three-dimensional core characteristics of the rectangular sub-blocks, and acquiring segmentation thresholds of the rectangular sub-blocks based on the peak types and the preliminary types of the rectangular sub-blocks; According to the preliminary type and the segmentation threshold value of each rectangular sub-block, obtaining the preliminary type segmentation area of each rectangular sub-block, obtaining edge pixels in the preliminary type segmentation area, obtaining gradient direction variance of the preliminary type segmentation area based on the gray gradient direction of the edge pixels, judging whether the gradient direction variance of each preliminary type segmentation area is matched with the regional gradient feature standard of the same preliminary type, if so, marking the preliminary type segmentation area as an undoubted type area, and if not, marking the preliminary type segmentation area as an interference area.
3. 3. The objective evaluation system for dynamic characteristics of pupillary light reflex according to claim 2, wherein the process of additionally generating the safety threshold interval of the critical decision parameters in the anesthesia depth detection mode comprises: if the patient is in an anesthesia depth detection mode, presetting an anesthesia depth-pupil parameter safety threshold mapping library, wherein the anesthesia depth-pupil parameter safety threshold mapping library comprises key judgment parameter safety thresholds corresponding to different anesthesia depths and crowd types, taking a dynamic parameter set of a pupil diameter and pupil diameter-time change curve as key judgment parameters, inputting the current anesthesia depth of the patient and the crowd type to the anesthesia depth-pupil parameter safety threshold mapping library, outputting a key judgment parameter safety threshold interval corresponding to the crowd type and the current anesthesia depth, comparing the key judgment parameters with the corresponding judgment parameter safety threshold interval, if any key judgment parameter exceeds the corresponding judgment parameter safety threshold interval, immediately triggering voice warning, marking any key judgment parameter as an abnormal parameter, comparing the abnormal parameter with the judgment parameter safety threshold interval, generating an abnormal parameter deviation degree, and feeding the abnormal parameter and the abnormal parameter deviation degree back to a detection terminal.
4. 4. An objective evaluation system for dynamic characteristics of pupil light reflex according to claim 3, wherein the process of additionally generating personalized derivative parameters of key decision parameters in the neurologic follow-up mode comprises: If the patient is in a neurology department follow-up mode, extracting total historical data of the patient from blockchain nodes, wherein the total historical data comprises crowd types of the patient, individual key judgment parameter baseline data, last and last follow-up detection data, individual baseline threshold intervals and nerve function index weights of corresponding disease types, taking static characteristics and dynamic parameter sets of a current pupil diameter-time change curve as key judgment parameters, acquiring absolute differences and relative change rates of the current key judgment parameters and the individual key judgment parameter baseline data, and absolute differences and relative change rates of the current key judgment parameters and the last follow-up detection data, acquiring change trend and change rate of each current key judgment parameter by adopting a linear regression algorithm based on the current key judgment parameters and the last follow-up detection data, simultaneously comparing the static characteristics and the dynamic parameter sets of pupil diameter-time change curves of pupils on the left side and the right side of the patient, acquiring bilateral parameter symmetry, and comparing bilateral parameter symmetry with the last follow-up detection data, and acquiring an asymmetry change coefficient; And comparing the current key judgment parameters with the individual baseline threshold interval, and marking the key judgment parameters exceeding the individual baseline threshold interval as abnormal parameters.
5. 5. The objective assessment system for dynamic characteristics of pupils' light reflex according to claim 4, wherein the process of constructing the clinical assessment look-up table comprises: Clustering static features and dynamic parameter sets of a plurality of different crowd types and different scenery information in a database by taking the different crowd types and the scenery information as a clustering center, acquiring the static features and the dynamic parameter sets of the different crowd types under the different scenery information conditions, carrying out distribution type inspection on each group of dynamic parameters in the static features and the dynamic parameter sets of the different crowd types, acquiring the distribution type of the static features and each group of dynamic parameters of the different crowd types under the different scenery information conditions, acquiring the static features and the common reference range of each group of dynamic parameters of the different crowd types under the different scenery information conditions according to the distribution type, acquiring the clinical classification type of the static features and each group of dynamic parameters, and acquiring the reflection grade critical interval of the static features and each group of dynamic parameters according to the clinical classification type and the common reference range of the static features and each group of dynamic parameters; and constructing a clinical evaluation comparison table according to the static characteristics of different crowd types under different scene information conditions and the reflection grade critical interval of each group of dynamic parameters.
6. 6. The objective evaluation system for dynamic characteristics of pupil light reflex according to claim 5, wherein the process of generating standardized diagnosis and treatment advice and clinical warning signals of the patient according to the reflex scale, and generating the individualized evaluation conclusion of the neural function of the patient according to the individualized derivative parameters of the critical decision parameters comprises: Presetting an inline clinical knowledge base, wherein the inline clinical knowledge base comprises standardized diagnosis and treatment suggestions and clinical early warning signals corresponding to different reflection grades and crowd types, and if a patient is in an emergency rapid screening mode, a conventional detection mode or an anesthesia depth detection mode, inputting the crowd types and the reflection grades of the patient into the inline clinical knowledge base, and outputting the standardized diagnosis and treatment suggestions and the clinical early warning signals of the patient; If the patient is in the neurology department follow-up mode, presetting a nerve function grading quantitative determination rule, and generating a nerve function personalized assessment conclusion of the patient according to the nerve function grading quantitative determination rule, the absolute difference value of key determination parameters, the relative change rate, the change trend, the change rate, the asymmetric change coefficient and whether abnormal parameters exist.

Description

Objective evaluation system for dynamic characteristics of pupil light reflex Technical Field The invention relates to the technical field of pupil light reflection characteristic evaluation, in particular to an objectification evaluation system for dynamic characteristics of pupil light reflection. Background Pupil light reflex is a core physiological reflex under the regulation of central nervous system functions, and dynamic characteristics such as contraction/relaxation amplitude, speed, latency, recovery time and the like are key indexes for judging consciousness states, damage degree of central nervous system, anesthesia depth and disease prognosis of patients. In clinical diagnosis and treatment, objectivity and accuracy of pupil light reflex assessment directly influence formulation and adjustment of diagnosis and treatment scheme. In the existing pupil light reflex assessment technology, a software module is used as a core processing unit, so that a plurality of technical defects are caused, and the key bottleneck for limiting assessment precision and clinical practicability is that a single parameter reference range is adopted in clinical assessment, physiological and pathological feature differences of people with different ages and diseases are not considered, misjudgment and missed judgment are easy to occur due to the fact that single parameter judgment results are mostly relied on, adaptability of assessment results and clinical actual diagnosis and treatment requirements is insufficient, few software with a dynamic analysis function only can calculate basic parameters of contraction amplitude and recovery time of 2-3, quantification of core dynamic features such as latency, speed and stability is lacking, parameter calculation logic is coarse, fine changes of pupil reflex cannot be reflected, and clinical precision assessment requirements are difficult to meet. Disclosure of Invention In order to solve the technical problems, the invention aims to provide an objectification evaluation system for dynamic characteristics of pupil light reflection, which comprises a cloud end, a detection terminal, a segmentation threshold judgment module, a data processing module, a reflection evaluation module and an evaluation early warning module; The segmentation threshold judging module is used for carrying out frame-by-frame graying on the color pupil image sequence, carrying out personalized denoising on each frame of gray image, dividing the denoised gray image into a plurality of rectangular sub-blocks, obtaining the segmentation threshold of each rectangular sub-block, carrying out region feature recognition on each rectangular sub-block based on the segmentation threshold, and obtaining an undoubted type region and an interference region; the data processing module is used for constructing a pupil diameter-time change curve of a pupil sub-block coverage area in the undoubted type area, acquiring key judgment parameters of the pupil diameter-time change curve of a patient under different scene information conditions, and additionally generating a key judgment parameter safety threshold interval under an anesthesia depth detection mode and key judgment parameter personalized derivative parameters under a neurology follow-up mode; The reflection evaluation module is used for constructing a clinical evaluation comparison table, evaluating the dynamic parameter set according to the clinical evaluation comparison table and obtaining the reflection grade corresponding to the dynamic parameter set; the assessment early warning module is used for generating standardized diagnosis and treatment advice and clinical early warning signals of the patient according to the reflection grade, and generating a nerve function personalized assessment conclusion of the patient according to the key judgment parameter personalized derivative parameters. Further, the cloud end is in communication connection with a detection terminal in a preset range, the detection terminal is used for uploading the collected color pupil image sequence and the crowd type of the patient to the detection terminal, a segmentation threshold judgment module, a data processing module, a reflection evaluation module, an evaluation early warning module and a database are arranged in the cloud end, a plurality of block chain nodes are arranged in the database, the block chain nodes are mutually linked to form a block chain network, each block chain node is linked with a data uploading end, and the data uploading end is used for storing the dynamic parameter set and the reflection grade of the patient generated by the reflection evaluation module to the block chain nodes and marking the uploading time. Further, the process of acquiring the undoubted type area and the interference area includes: Setting the size of a local area, splitting the gray image of the current frame into a plurality of rectangular sub-blocks according to the size of the local