Parkinson's disease gait impairment assessment model and method and storage medium

Abstract

The invention relates to the technical field of medical image processing, in particular to a Parkinson's disease gait damage assessment model and method and a storage medium. The method comprises the following steps: continuously monitoring gait images of a Parkinson's disease patient; when it is detected that gait continuity of the gait image is interrupted, it is judged that the gait image is a gait freezing event, and freezing event characteristics, freezing occurrence time periods and freezing duration of the gait freezing event are recorded; determining gait feature parameters based on the gait image; extracting gait spatial-temporal characteristic parameters of the gait characteristic parameters, and constructing a gait characteristic curve according to the gait spatial-temporal characteristic parameters; according to the gait characteristic curve, recognizing the predecessor attribute of the freezing event characteristic, and obtaining the predecessor attribute of the freezing event. Through the data processing technology and the image processing technology, gait monitoring of the Parkinson's disease patient is achieved, the gait freezing features are recognized, the accuracy of Parkinson's disease gait damage assessment is improved, and the time used in the damage assessment process is shortened.

Description

Technical Field

[0001] The present invention relates to the technical field of medical image processing, and in particular to a model, method and storage medium for establishing a Parkinson's disease gait impairment assessment model. Background Art

[0002] In the early stage, the gait assessment of Parkinson's disease mainly relied on the experience of clinicians and the subjective description of patients. Doctors judged the degree of gait impairment by observing the patient's walking posture and gait characteristics. This method is highly subjective, lacks a unified quantitative standard, and is difficult to accurately assess changes in the disease. With the gradual introduction of gait analysis technology into the assessment of Parkinson's disease, gait analysis can quantify gait abnormalities by collecting parameters in the walking cycle. The technology at this stage is mostly based on laboratory environments, using pressure sensors, optical motion capture systems and other equipment. However, when detecting gait freezing events in Parkinson's patients, existing technologies cannot accurately identify the specific characteristics and duration of gait interruptions. In addition, existing gait analysis methods only focus on a single gait parameter, and lack a comprehensive analysis of gait spatiotemporal characteristic parameters, making it difficult to monitor gait characteristics, resulting in low accuracy in the assessment of gait impairment in Parkinson's disease. Summary of the invention

[0003] Based on this, it is necessary to provide a model, method and storage medium for establishing Parkinson's disease gait impairment assessment to solve at least one of the above technical problems.

[0004] To achieve the above object, a method for assessing gait impairment in Parkinson's disease is established, the method comprising the following steps: Step S1: continuously monitoring the gait image of the Parkinson's disease patient; when the gait continuity of the gait image is detected to be interrupted, it is determined as a gait freezing event, and the freezing event characteristics, freezing occurrence time period and freezing duration of the gait freezing event are recorded; Step S2: determining gait characteristic parameters based on the gait image; extracting gait spatiotemporal characteristic parameters of the gait characteristic parameters, and constructing a gait characteristic curve according to the gait spatiotemporal characteristic parameters; identifying the precursor attribute of the freezing event feature according to the gait characteristic curve, and obtaining the precursor attribute of the freezing event; Step S3: classify the precursor attributes of the freezing event into gait gradual change types; perform gait freezing cycle detection on the gait gradual change types according to the freezing occurrence time period to generate a gait freezing cycle; quantify the freezing cumulative frequency of the gait freezing cycle based on the freezing duration to obtain a gait freezing index; Step S4: determining the gait change trend based on the freezing cycle data and the gait freezing index, and evaluating the degree of gait impairment of Parkinson's disease on the gait change trend, and generating a gait impairment evaluation report.

[0005] The present invention uses a high-resolution image acquisition device to monitor the gait of Parkinson's patients in real time, ensuring that a complete gait image sequence can be captured at any time point, and can continuously record the patient's walking process, providing a comprehensive data basis for subsequent analysis; the gait image is analyzed based on the image processing algorithm, and when a gait continuity interruption is detected, it is automatically determined as a gait freezing event. At the same time, the characteristics of the freezing event are recorded, including the time period when the freezing occurs, the duration of the freezing, and the morphological characteristics of the freezing event, ensuring the accuracy and integrity of the data and avoiding errors caused by manual intervention. Gait spatiotemporal characteristic parameters are extracted from the gait image, including key parameters such as step length, step width, step speed, and gait cycle. These parameters are calculated through image analysis algorithms, ensuring the objectivity and repeatability of the data. A gait characteristic curve is constructed based on the extracted spatiotemporal characteristic parameters of gait. The curve takes time as the horizontal axis and gait characteristic parameters as the vertical axis, which can intuitively reflect the changing trend of gait parameters over time. Through this curve processing, the system can capture the dynamic process of gait changes. By analyzing the gait characteristic curve, the characteristic changes before the freezing event, that is, the precursor attributes of the freezing event, are identified. This identification process is based on the morphological analysis of the curve and the parameter change trend, and can accurately locate the precursor characteristics of the freezing event. The precursor attributes of freezing events are divided into different gait gradient types. This division is based on the morphological characteristics and parameter change rules of the gait characteristic curve, and can classify and process complex gait changes. The gait gradient types are periodically detected according to the time period when the freezing occurs. By analyzing the periodic changes of the gait characteristic curve, the periodic pattern of gait freezing can be accurately identified, thereby revealing the regularity of gait freezing events. The gait freezing cycle is quantitatively analyzed based on the duration of freezing, and the cumulative frequency and duration of freezing events are calculated. The weighted sum is obtained to comprehensively reflect the severity of gait freezing, providing a unified quantitative standard for the assessment of the degree of gait impairment. Based on the freezing cycle data and the gait freezing index, the trend of gait changes is analyzed by a mathematical model, and the regularity of the freezing cycle and the changing trend of the freezing index can be comprehensively considered, so as to accurately judge the overall situation of gait impairment; according to the gait change trend, the degree of gait impairment is quantitatively evaluated, and the evaluation process is based on a preset quantitative standard, and by calculating the difference between the gait freezing index and the normal gait parameters, an objective and quantitative evaluation result can be provided; according to the evaluation result, a detailed gait impairment evaluation report is generated, and the report contains a detailed record of the freezing event, a gait characteristic curve, a freezing cycle analysis result, and a quantitative evaluation of the degree of gait impairment. The report is presented in the form of charts and data, which can intuitively show the details of gait changes and provide comprehensive data support for related research. Therefore, the present invention realizes gait monitoring of Parkinson's patients through data processing technology and image processing technology, and identifies gait freezing characteristics, so as to improve the accuracy of Parkinson's gait impairment evaluation and shorten the time used for the impairment evaluation process.

[0006] Preferably, step S1 comprises the following steps: Step S11: setting the frame rate of the binocular camera to 40-60 frames / second, the resolution to 1920×1080 pixels, and setting the angle between the binocular camera and the patient's walking path to 45°; Step S12: collecting the complete walking cycle of the patient each time, including the initial gait, the middle gait and the final gait, and each collection lasts for 3 minutes; Step S13: activating the automatic tracking function of the binocular camera to continuously monitor the walking path, number of steps, walking speed and walking movements of the Parkinson's patient to obtain a gait image; Step S14: recording the number of swings of the patient's gait image, marking the patient's gait swing number as zero for two consecutive cycles; determining the gait interruption point for the two consecutive cycles of zero times, and obtaining gait continuity interruption data; Step S15: performing gait freezing judgment on the gait image according to the gait continuity interruption data, obtaining a gait freezing event, and recording the freezing event characteristics, freezing occurrence time period and freezing duration of the gait freezing event.

[0007] The present invention sets the frame rate of the binocular camera to 40-60 frames per second, the resolution to 1920×1080 pixels, and sets the angle between the binocular camera and the patient's walking path to 45°; setting the frame rate of the binocular camera to 40-60 frames per second, and the resolution to 1920×1080 pixels can ensure that clear and continuous image data is captured during gait monitoring, avoiding image blur and information loss caused by insufficient frame rate or too low resolution. Setting the angle between the binocular camera and the patient's walking path to 45° can effectively capture the side profile and dynamic changes of the patient's gait, while reducing occlusion and deformation caused by viewing angle problems, ensuring accurate extraction of gait features; collecting the patient's complete walking cycle each time, including the starting gait, intermediate gait and ending gait, can comprehensively reflect the patient's gait characteristics at different stages and avoid analysis bias caused by fragmented collection; each collection lasts for 3 minutes to ensure that a sufficient amount of gait data is collected, while avoiding insufficient or redundant data due to too short or too long collection time; starting the automatic tracking function of the binocular camera can monitor the patient's walking path, number of steps, speed and movements in real time, ensuring that clear gait images are always captured during the patient's walking process, avoiding errors caused by manual adjustments. Simultaneously monitoring walking path, number of steps, speed and movements can provide a rich data basis for subsequent gait analysis and support accurate identification of gait freezing events; recording the number of patient step swings in gait images can quantify changes in the patient's gait rhythm and provide key data for identifying gait freezing events; marking two consecutive cycles of zero times as gait interruption points can accurately locate the location of gait freezing events and ensure the accuracy of subsequent analysis; gait freezing judgment based on gait continuity interruption data can accurately identify gait freezing events and avoid errors caused by subjective judgment; recording the freezing event characteristics, freezing time period and freezing duration of gait freezing events can provide comprehensive and accurate data support for subsequent gait impairment assessments.

[0008] Determining gait characteristic parameters based on the gait image includes: Divide the gait image into a plurality of continuous frame sequences, wherein each frame sequence corresponds to a gait cycle; In each frame sequence, the foot joint, knee joint and ankle joint are extracted to obtain the human joint point information; the displacement time change of the human joint point information in adjacent frames is calculated to generate the gait motion cycle; Calculate the distance between the foot joint and the contact point in two consecutive gait motion cycles to obtain the step length data; measure the lateral distance between the left and right foot joints and the contact point to obtain the step width data; Calculate gait speed using step length data and gait motion cycle; determine gait symmetry for gait motion cycle based on gait speed and step width data; Calculate the flexion and extension phase angles of the knee joint and the ankle joint in two consecutive gait motion cycles, and map the flexion and extension phase angles to flexion and extension discreteness to obtain joint flexion and extension discreteness; The gait motion cycle, step length data, step width data, gait speed, gait symmetry and joint flexion and extension discreteness are combined into gait characteristic parameters.

[0009] The present invention divides the gait image into multiple continuous frame sequences, and makes each frame sequence correspond to a gait cycle, so as to ensure that the gait analysis is based on a complete gait action unit, and provide a clear time division for subsequent feature extraction; By extracting the information of the foot joint, knee joint and ankle joint in each frame sequence, the movement trajectory of the key parts of the human body can be accurately located, providing basic data for the quantitative analysis of gait characteristics; by calculating the displacement time changes of the human body joint points in adjacent frames, the dynamic process of joint movement can be accurately reflected, providing a time reference for the subsequent quantitative analysis of gait characteristics; by calculating the distance between the foot joint and the contact point in two consecutive gait movement cycles, the step length data can be directly obtained, providing a key parameter for the quantitative analysis of gait characteristics; by measuring the lateral distance between the left and right foot joints and the contact point, the step width data can be accurately obtained, providing an important lateral parameter for the quantitative analysis of gait characteristics; combining the step length data with the gait movement cycle to calculate the gait speed , which can directly reflect the patient's walking speed and provide speed parameters for the quantitative analysis of gait characteristics; by analyzing the gait movement cycle based on gait speed and step width data, the gait symmetry can be determined, providing a quantitative basis for evaluating the coordination and balance of gait; by calculating the flexion and extension phase angles of the knee and ankle joints in two consecutive gait movement cycles and mapping them into flexion and extension discreteness, the coordination and consistency of joint movement can be quantified, providing important parameters for the comprehensive analysis of gait characteristics; the gait movement cycle, step length data, step width data, gait speed, gait symmetry and joint flexion and extension discreteness are combined into gait characteristic parameters, which can comprehensively reflect the dynamic characteristics of gait and provide comprehensive quantitative data for subsequent gait impairment assessment.

[0010] The step of extracting the gait spatiotemporal characteristic parameters of the gait characteristic parameters and constructing a gait characteristic curve according to the gait spatiotemporal characteristic parameters comprises: Extracting the periodic time information of the gait characteristic parameters, and dividing the periodic time information into a time period sequence; performing sequence parameterization on the time period sequence to obtain the time period sequence parameters; Based on the time period sequence parameters, gait spatial feature recognition is performed on gait feature parameters, and feature parameters of gait spatial features are quantified to obtain time period sequence parameters; Arrange the time period series parameters in chronological order to form chronological order data; The time-arranged sequence data is used as the horizontal axis and the time period sequence parameters are used as the vertical axis to construct the initial gait characteristic curve; The initial gait characteristic curve is processed for curve smoothness to generate a gait characteristic curve.

[0011] The present invention can subdivide the time dimension of gait data by extracting the periodic time information of gait characteristic parameters and dividing it into time period sequences, so as to provide a clear time frame for subsequent parameterization processing; parameterize the time period sequence, so as to convert the time information into a quantifiable data form, and provide basic data support for further analyzing gait characteristics; perform spatial feature recognition on gait characteristic parameters based on time period sequence parameters, and perform quantitative processing, so as to convert the spatial characteristics of gait into specific numerical parameters; arrange the time period sequence parameters in time order to form time arrangement sequence data, so as to ensure the time coherence of data, and provide an orderly data basis for constructing gait characteristic curves; take the time arrangement sequence data as the horizontal axis and the time period sequence parameters as the vertical axis to construct the initial gait characteristic curve; take the time arrangement sequence data as the horizontal axis and the time period sequence parameters as the vertical axis to construct the initial gait characteristic curve, so as to intuitively show the changing trend of gait characteristics over time, and provide a visualization basis for subsequent curve processing. Perform smoothness processing on the initial gait characteristic curve to generate a smooth gait characteristic curve, so as to remove noise and abnormal fluctuations in the curve.

[0012] The preceding attributes of identifying the freezing event characteristics according to the gait characteristic curve include: On the gait characteristic curve, mark the time point when the gait speed drops below 0.5 m / s as the starting point of the speed freezing event; calculate the rate of change of gait speed within 3 seconds before and after the marked point based on the starting point of the speed freezing event. If the rate of change is less than -0.2 m / s², it is identified as a speed precursor feature; On the gait characteristic curve, mark the time point when the gait symmetry index drops below 0.8 as the starting point of the symmetry freezing event; compare the change in gait symmetry within 5 seconds before and after the marked point based on the starting point of the symmetry freezing event. If the symmetry change exceeds 0.2, record the time interval of the gait symmetry change. If the interval is less than 1 second, it is identified as a symmetry precursor feature. On the gait characteristic curve, mark the time point when the stride length is shortened to less than 68 cm as the starting point of the stride length freezing event; compare the change in stride length within 3 seconds before and after the marked point based on the starting point of the stride length freezing event. If the stride length is shortened by more than 5 cm, it is identified as a stride length precursor feature; On the gait characteristic curve, mark the time point when the step width changes by more than 3.5 cm as the starting point of the step width freezing event; compare the change of step width within 3 seconds before and after the marked point based on the starting point of the step width freezing event. If the step width changes by more than 2 cm, it is identified as a step width precursor feature; On the gait characteristic curve, mark the time point when the joint flexion and extension discreteness changes by more than 10% as the starting point of the flexion and extension freezing event; compare the change of joint flexion and extension discreteness within 3 seconds before and after the marked point based on the starting point of the flexion and extension freezing event. If the change exceeds 5%, it is identified as the precursor feature of joint flexion and extension; The speed precursor features, symmetry precursor features, step width precursor features and joint flexion and extension precursor features are combined to obtain the freezing event precursor attributes.

[0013] The present invention can accurately locate the starting moment of the speed freezing event by marking the time point when the gait speed drops below 0.5 m / s on the gait characteristic curve, calculate the change rate of the gait speed within 3 seconds before and after the marking point, and identify it as a speed precursor feature when the change rate is less than -0.2 m / s², which can quantify the severity of the gait speed change; by marking the time point when the gait symmetry index drops below 0.8 on the gait characteristic curve, it can accurately locate the starting moment of the symmetry freezing event, providing a clear time reference for subsequent analysis; by comparing the change of gait symmetry within 5 seconds before and after the marking point, when the symmetry change exceeds 0.2 and the time interval is less than 1 second, it is identified as a symmetry precursor feature, which can quantify the severity and time characteristics of the gait symmetry change, and provide an objective basis for the early identification of freezing events; by marking the time point when the stride length is shortened to less than 68 cm on the gait characteristic curve, it can accurately locate the starting moment of the stride length freezing event; by comparing the change of stride length within 3 seconds before and after the marking point, when When the step length is shortened by more than 5 cm, it is identified as a stride length precursor feature, which can quantify the severity of the step length change and provide an objective basis for the early identification of freezing events; by marking the time point when the step width changes by more than 3.5 cm on the gait characteristic curve, the starting time of the step width freezing event can be accurately located; by comparing the change in step width within 3 seconds before and after the marking point, when the step width changes by more than 2 cm, it is identified as a stride width precursor feature, which can quantify the severity of the step width change; by marking the time point when the joint flexion and extension discreteness changes by more than 10% on the gait characteristic curve, the starting time of the flexion and extension freezing event can be accurately located; by comparing the change in joint flexion and extension discreteness within 3 seconds before and after the marking point, when the change exceeds 5%, it is identified as a joint flexion and extension precursor feature, which can quantify the severity of the joint flexion and extension change; the speed precursor feature, symmetry precursor feature, step length precursor feature, step width precursor feature and joint flexion and extension precursor feature are combined to obtain the precursor attribute of the freezing event, which can comprehensively reflect the multi-dimensional feature changes before the gait freezing event occurs.

[0014] Preferably, step S3 comprises the following steps: Step S31: extracting gait interruption precursor features from the freezing event precursor attributes to obtain gait interruption precursor features; determining the interruption type from the gait interruption precursor features to generate a gait interruption type; Step S32: performing inter-type similarity comparison on the gait interruption types to obtain inter-type similarity data; and classifying the precursor attributes of the freezing event into gait gradual change types according to the inter-type similarity data; Step S33: segmenting the freezing time period according to the freezing event characteristics, taking the freezing event characteristic occurrence point as the center, extending forward and backward by 5 seconds each, and dividing it into a pre-freezing period and a post-freezing period; Step S34: marking the gait type normal interval according to the pre-freezing interval for the gait gradual change type; marking the gait type frozen interval according to the post-freezing interval; Step S35: gait freezing period is calculated based on the normal gait type interval and the gait type freezing interval to generate a gait freezing period; Step S36: quantifying the cumulative freezing frequency of the gait freezing period based on the freezing duration to obtain a gait freezing index.

[0015] The present invention extracts features from the precursor attributes of freezing events, and can accurately identify early signals of gait interruption; classifies the precursor features of gait interruption, generates gait interruption types, and provides a clear classification basis for subsequent analysis; by comparing the similarity of gait interruption types, the correlation of different interruption types can be quantified. Based on similar data between types, the precursor attributes of freezing events are classified into gait gradual change types, providing a classification basis for subsequent periodic measurement; with the occurrence point of the freezing event feature as the center, extending forward and backward for 5 seconds each, it is divided into a pre-freezing interval and a post-freezing interval, and the time range of the freezing event can be accurately located; based on the pre-freezing interval, the gait gradual change type is marked, and the characteristic interval of the normal gait can be clarified; based on the post-freezing interval, the gait gradual change type is marked, and the characteristic interval of the freezing gait can be clarified; by comparing the normal interval of the gait type with the freezing interval, the gait freezing period is measured, and the periodic characteristics of the freezing event can be quantified; based on the duration of freezing, the gait freezing period is quantified, and the severity of the freezing event can be fully reflected.

[0016] Preferably, step S36 includes the following steps: Step S361: performing periodic segmentation processing on the freezing duration to obtain the freezing type segmentation duration; Step S362: performing duration normalization processing on the frozen type segmentation duration to obtain duration normalized data; Step S363: performing freezing frequency statistics on the gait freezing period according to the normalized duration data, and determining the occurrence frequency of each freezing period to obtain freezing frequency distribution data; Step S364: performing cumulative calculation on the freezing frequency distribution data, and accumulating the frequencies of each freezing period in chronological order to form a freezing cumulative frequency curve; Step S365: performing frozen frequency parameter mapping on the frozen cumulative frequency curve, and dividing the frozen frequency parameter into a plurality of quantization intervals, each interval corresponding to a quantization value, to generate a frozen frequency quantization parameter; Step S366: performing weighted calculation on the freezing frequency quantization parameters to obtain a final gait freezing index.

[0017] The present invention performs periodic segmentation on the duration of freezing, can subdivide freezing events according to periodic characteristics, and provide basic data for subsequent duration normalization processing; normalization processing is performed on the duration of freezing type segmentation, which can eliminate the influence of different cycle lengths on the analysis and make the data comparable; based on the duration normalization data, freezing frequency statistics are performed on the gait freezing cycle to determine the occurrence frequency of each freezing cycle, which can quantify the periodic occurrence law of freezing events; cumulative calculation is performed on the freezing frequency distribution data, and the frequency of each freezing cycle is accumulated in chronological order to form a freezing cumulative frequency curve, which can intuitively display the cumulative change trend of freezing events over time; parameter mapping is performed on the freezing cumulative frequency curve, and the freezing frequency parameters are divided into multiple quantization intervals, each interval corresponds to a quantization value, which can convert continuous frequency data into quantifiable discrete parameters; weighted calculation is performed on the freezing frequency quantization parameters, which can comprehensively reflect the severity of gait freezing events and provide a quantitative basis for gait impairment assessment.

[0018] Preferably, step S4 comprises the following steps: Step S41: Arrange the freezing cycle data in chronological order, and draw a freezing cycle curve with the start time of the freezing cycle as the horizontal axis and the freezing duration as the vertical axis; Step S42: Calculate the slope of the freezing cycle curve according to the gait freezing index. When the slope is greater than 0, it is determined that the gait freezing is increasing; when the slope is less than 0, it is determined that the gait freezing is decreasing. Step S43: merging the gait freezing enhancement trend and the gait freezing reduction trend into gait trend features to generate gait change trend data; Step S44: dividing the gait change trend into three intervals of mild, moderate and severe gait impairment, and marking them with different colors to obtain gait change impairment colors; Step S45: Mapping the color of the gait change impairment to the degree of gait impairment of Parkinson's disease, and visualizing the degree of gait impairment to generate a gait impairment assessment report.

[0019] The present invention arranges the freezing cycle data in chronological order and draws a curve, which can intuitively display the changing trend of the starting time and duration of the freezing cycle, providing a visual basis for subsequent trend analysis; the slope of the freezing cycle curve is calculated based on the gait freezing index, and the changing trend of gait freezing can be quantified. When the slope is greater than 0, it is determined as a gait freezing strengthening trend, and when the slope is less than 0, it is determined as a gait freezing weakening trend, providing a clear basis for subsequent trend analysis; the gait freezing strengthening trend and the weakening trend are merged to generate gait change trend data, which can comprehensively reflect the overall change of gait freezing; the gait change trend is divided into three intervals of mild, moderate and severe impairment, and marked with different colors respectively, which can intuitively distinguish gait changes of different impairment degrees, and provide clear visual identification for visual analysis. By mapping the gait change impairment color with the gait impairment degree of Parkinson's disease and performing visualization processing, an intuitive gait impairment assessment report can be generated, providing a clear and accurate visualization basis for related analysis.

[0020] The present invention also provides a model for establishing a Parkinson's disease gait impairment assessment model, which is used for the above-mentioned method for establishing a Parkinson's disease gait impairment assessment model. The model for establishing a Parkinson's disease gait impairment assessment model comprises: An image monitoring module is used to continuously monitor the gait images of Parkinson's patients; when a gait continuity interruption in the gait image is detected, it is determined as a gait freezing event, and the freezing event characteristics, freezing occurrence time period and freezing duration of the gait freezing event are recorded; A gait characteristic curve construction module is used to determine gait characteristic parameters based on gait images; extract gait spatiotemporal characteristic parameters of the gait characteristic parameters, and construct a gait characteristic curve according to the gait spatiotemporal characteristic parameters; identify the precursor attributes of the freezing event characteristics according to the gait characteristic curve, and obtain the precursor attributes of the freezing event; The gait freezing quantification module is used to classify the precursor attributes of the freezing event into gait gradual change types; perform gait freezing cycle detection on the gait gradual change types according to the freezing time period to generate the gait freezing cycle; perform freezing cumulative frequency quantification on the gait freezing cycle based on the freezing duration to obtain the gait freezing index; The impairment degree assessment module is used to determine the gait change trend based on the freezing cycle data and the gait freezing index, and to assess the gait impairment degree of Parkinson's disease on the gait change trend, and to generate a gait impairment assessment report.

[0021] The image monitoring module of the present invention monitors the gait of Parkinson's patients in real time through a high-resolution image acquisition device, ensuring that a complete gait image sequence can be captured at any time point, and can continuously record the patient's walking process, providing a comprehensive data basis for subsequent analysis; the gait image is analyzed based on the image processing algorithm, and when a gait continuity interruption is detected, it is automatically determined as a gait freezing event. At the same time, the characteristics of the freezing event are recorded, including the time period when the freezing occurs, the duration of the freezing, and the morphological characteristics of the freezing event, ensuring the accuracy and integrity of the data and avoiding errors caused by manual intervention. Through the gait characteristic curve construction module, the gait spatiotemporal characteristic parameters of the gait are extracted from the gait image, including key parameters such as step length, step width, step speed, and gait cycle. These parameters are calculated through image analysis algorithms, ensuring the objectivity and repeatability of the data. A gait characteristic curve is constructed based on the extracted spatiotemporal characteristic parameters of gait. The curve takes time as the horizontal axis and gait characteristic parameters as the vertical axis, which can intuitively reflect the changing trend of gait parameters over time. Through this curve processing, the system can capture the dynamic process of gait changes. By analyzing the gait characteristic curve, the characteristic changes before the freezing event, that is, the precursor attributes of the freezing event, are identified. This identification process is based on the morphological analysis of the curve and the parameter change trend, and can accurately locate the precursor characteristics of the freezing event. Through the gait freezing quantification module, the precursor attributes of the freezing event are divided into different gait gradient types. This division is based on the morphological characteristics and parameter change rules of the gait characteristic curve, and can classify and process complex gait changes; and according to the time period when the freezing occurs, the gait gradient type is periodically detected. By analyzing the periodic changes of the gait characteristic curve, the periodic pattern of gait freezing can be accurately identified, thereby revealing the regularity of gait freezing events; the gait freezing cycle is quantitatively analyzed based on the duration of freezing, and the cumulative frequency of the freezing event and the weighted sum of the duration are calculated to comprehensively reflect the severity of gait freezing, providing a unified quantitative standard for the assessment of the degree of gait impairment. Through the impairment assessment module, based on the freezing cycle data and gait freezing index, the trend of gait changes is analyzed through a mathematical model, which can comprehensively consider the regularity of the freezing cycle and the changing trend of the freezing index, so as to accurately judge the overall situation of gait impairment; according to the gait change trend, the degree of gait impairment is quantitatively assessed. This assessment process is based on the preset quantitative standards, and by calculating the difference between the gait freezing index and the normal gait parameters, it can provide objective and quantitative assessment results; based on the assessment results, a detailed gait impairment assessment report is generated, which contains detailed records of freezing events, gait characteristic curves, freezing cycle analysis results, and quantitative assessment of gait impairment. The report is presented in the form of charts and data, which can intuitively show the details of gait changes and provide comprehensive data support for related research.Therefore, the present invention realizes gait monitoring of Parkinson's patients and identifies gait freezing characteristics through data processing technology and image processing technology, so as to improve the accuracy of Parkinson's disease gait impairment assessment and shorten the time used in the impairment assessment process.

[0022] A computer-readable storage medium stores a computer program, wherein the computer program is used to execute the method for establishing a Parkinson's disease gait impairment assessment method. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] Figure 1 A schematic diagram of the steps for establishing a method for assessing gait impairment in Parkinson's disease; Figure 2 for Figure 1 Detailed implementation steps of step S1 in FIG. Figure 3 for Figure 1 Detailed implementation steps of step S4 in FIG. The realization of the purpose, functional features and advantages of the present invention will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0024] The technical method of the present invention is described clearly and completely below in conjunction with the accompanying drawings. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by technicians in this field without creative work are within the scope of protection of the present invention.

[0025] In addition, the accompanying drawings are only schematic illustrations of the present invention and are not necessarily drawn to scale. The same reference numerals in the figures represent the same or similar parts, and their repeated description will be omitted. Some of the block diagrams shown in the accompanying drawings are functional entities and do not necessarily correspond to physically or logically independent entities. The functional entities can be implemented in software form, or implemented in one or more hardware modules or integrated circuits, or implemented in different networks and / or processor methods and / or microcontroller methods.

[0026] It should be understood that, although the terms "first", "second", etc. may be used herein to describe various units, these units should not be limited by these terms. These terms are used only to distinguish one unit from another unit. For example, without departing from the scope of the exemplary embodiments, the first unit may be referred to as the second unit, and similarly the second unit may be referred to as the first unit. The term "and / or" used herein includes any and all combinations of one or more of the listed associated items.

[0027] To achieve this, please refer to Figures 1 to 3 , a method for establishing a Parkinson's disease gait impairment assessment method, the method comprising the following steps: Step S1: continuously monitoring the gait image of the Parkinson's disease patient; when the gait continuity of the gait image is detected to be interrupted, it is determined as a gait freezing event, and the freezing event characteristics, freezing occurrence time period and freezing duration of the gait freezing event are recorded; Step S2: determining gait characteristic parameters based on the gait image; extracting gait spatiotemporal characteristic parameters of the gait characteristic parameters, and constructing a gait characteristic curve according to the gait spatiotemporal characteristic parameters; identifying the precursor attribute of the freezing event feature according to the gait characteristic curve, and obtaining the precursor attribute of the freezing event; Step S3: classify the precursor attributes of the freezing event into gait gradual change types; perform gait freezing cycle detection on the gait gradual change types according to the freezing occurrence time period to generate a gait freezing cycle; quantify the freezing cumulative frequency of the gait freezing cycle based on the freezing duration to obtain a gait freezing index; Step S4: determining the gait change trend based on the freezing cycle data and the gait freezing index, and evaluating the degree of gait impairment of Parkinson's disease on the gait change trend, and generating a gait impairment evaluation report.

[0028] The present invention uses a high-resolution image acquisition device to monitor the gait of Parkinson's patients in real time, ensuring that a complete gait image sequence can be captured at any time point, and can continuously record the patient's walking process, providing a comprehensive data basis for subsequent analysis; the gait image is analyzed based on the image processing algorithm, and when a gait continuity interruption is detected, it is automatically determined as a gait freezing event. At the same time, the characteristics of the freezing event are recorded, including the time period when the freezing occurs, the duration of the freezing, and the morphological characteristics of the freezing event, ensuring the accuracy and integrity of the data and avoiding errors caused by manual intervention. Gait spatiotemporal characteristic parameters are extracted from the gait image, including key parameters such as step length, step width, step speed, and gait cycle. These parameters are calculated through image analysis algorithms, ensuring the objectivity and repeatability of the data. A gait characteristic curve is constructed based on the extracted spatiotemporal characteristic parameters of gait. The curve takes time as the horizontal axis and gait characteristic parameters as the vertical axis, which can intuitively reflect the changing trend of gait parameters over time. Through this curve processing, the system can capture the dynamic process of gait changes. By analyzing the gait characteristic curve, the characteristic changes before the freezing event, that is, the precursor attributes of the freezing event, are identified. This identification process is based on the morphological analysis of the curve and the parameter change trend, and can accurately locate the precursor characteristics of the freezing event. The precursor attributes of freezing events are divided into different gait gradient types. This division is based on the morphological characteristics and parameter change rules of the gait characteristic curve, and can classify and process complex gait changes. The gait gradient types are periodically detected according to the time period when the freezing occurs. By analyzing the periodic changes of the gait characteristic curve, the periodic pattern of gait freezing can be accurately identified, thereby revealing the regularity of gait freezing events. The gait freezing cycle is quantitatively analyzed based on the duration of freezing, and the cumulative frequency and duration of freezing events are calculated. The weighted sum is obtained to comprehensively reflect the severity of gait freezing, providing a unified quantitative standard for the assessment of the degree of gait impairment. Based on the freezing cycle data and the gait freezing index, the trend of gait changes is analyzed by a mathematical model, and the regularity of the freezing cycle and the changing trend of the freezing index can be comprehensively considered, so as to accurately judge the overall situation of gait impairment; according to the gait change trend, the degree of gait impairment is quantitatively evaluated, and the evaluation process is based on a preset quantitative standard, and by calculating the difference between the gait freezing index and the normal gait parameters, an objective and quantitative evaluation result can be provided; according to the evaluation result, a detailed gait impairment evaluation report is generated, and the report contains a detailed record of the freezing event, a gait characteristic curve, a freezing cycle analysis result, and a quantitative evaluation of the degree of gait impairment. The report is presented in the form of charts and data, which can intuitively show the details of gait changes and provide comprehensive data support for related research. Therefore, the present invention realizes gait monitoring of Parkinson's patients through data processing technology and image processing technology, and identifies gait freezing characteristics, so as to improve the accuracy of Parkinson's gait impairment evaluation and shorten the time used for the impairment evaluation process.

[0029] In the embodiment of the present invention, reference Figure 1 FIG. 1 is a flow chart of steps for establishing a method for assessing gait impairment in Parkinson's disease according to the present invention. In this example, the method for assessing gait impairment in Parkinson's disease includes the following steps: Step S1: continuously monitoring the gait image of the Parkinson's disease patient; when the gait continuity of the gait image is detected to be interrupted, it is determined as a gait freezing event, and the freezing event characteristics, freezing occurrence time period and freezing duration of the gait freezing event are recorded; In an embodiment of the present invention, micro inertial measurement unit (IMU) sensors are installed on the feet, knees and waist of a Parkinson's patient respectively to collect acceleration, angular velocity and posture information in real time. At the same time, multiple high-resolution cameras are installed above the patient's walking area to form a stereoscopic visual monitoring system for capturing the patient's gait image. The frame rate of the camera is 60fps and the resolution is 1920×1080 pixels to ensure that every detail of the gait can be clearly recorded; the collected gait image is preprocessed, including grayscale, denoising and edge detection. First, the color image is converted into a grayscale image to reduce the amount of data and improve processing efficiency. Then, the grayscale image is denoised using Gaussian filtering to remove random noise in the image. Finally, the Canny edge detection algorithm is used to extract edge information in the image for subsequent gait feature extraction; gait features are extracted based on the preprocessed image data. First, the patient's body contour and leg motion trajectory are identified by background subtraction. Then, the key parameters of the gait are calculated, including step length, step width, step speed, gait symmetry, swing time and support time. At the same time, combined with IMU sensor data, the patient's acceleration, angular velocity and posture angle changes are calculated to fully reflect the dynamic characteristics of gait; the gait characteristics are analyzed using deep learning algorithms to detect gait continuity interruptions. The specific method is to build a gait freezing detection model based on a convolutional neural network (CNN). The model inputs gait images and sensor data, and outputs the determination results of gait freezing events. The model learns the characteristic differences between normal gait and freezing state through the training data set. When the gait characteristics are detected to be similar to the freezing state, it is determined as a gait freezing event; when a gait freezing event is detected, the system automatically records the relevant features of the freezing event. The freezing event features include changes in gait parameters (such as shortened step length, reduced step speed, extended double support time, etc.), the timestamp of freezing (accurate to milliseconds), and the duration of freezing (obtained by calculating the difference between the start and end timestamps of freezing). These feature data are stored in the local database for subsequent analysis and research.

[0030] Step S2: determining gait characteristic parameters based on the gait image; extracting gait spatiotemporal characteristic parameters of the gait characteristic parameters, and constructing a gait characteristic curve according to the gait spatiotemporal characteristic parameters; identifying the precursor attribute of the freezing event feature according to the gait characteristic curve, and obtaining the precursor attribute of the freezing event; In an embodiment of the present invention, first, the gait image is analyzed using the OpenPose human posture recognition technology to extract the two-dimensional coordinate information of the key points of the human body. OpenPose detects the key points of the human body in the image, including ankles, knees, hips, etc., through a deep learning model, and outputs the precise positions of these key points. These key point data will serve as the basis for the subsequent extraction of gait feature parameters. Next, based on the extracted key point coordinates, the gait feature parameters are calculated. Specifically, the step length is obtained by calculating the distance between the center points of the two ankles; the step width is calculated by measuring the lateral distance between the center points of the two ankles; the walking speed is determined by the product of the step length and the step frequency; the single support time refers to the time when one foot touches the ground, which is calculated by analyzing the motion trajectory of the ankle key points; the double support time refers to the time when both feet touch the ground at the same time, which is also obtained by analyzing the motion trajectory of the ankle key points. These parameters can fully reflect the gait characteristics of the patient. Subsequently, the gait spatiotemporal feature parameters are extracted, and a gait feature curve is constructed. Taking the parameters such as step length, step frequency, and single support time as examples, the changes of these parameters over time are plotted as curves. In order to smooth the curve and remove noise, the Savitzky-Golay filter was used for processing. The filter smoothes the data by polynomial fitting while retaining the basic shape of the signal. In practical applications, the window size of the filter is set to 21 and the polynomial order is 2. This parameter configuration can better retain the detailed characteristics of the signal while removing noise. Finally, the change trend of the gait characteristic curve before the freezing event occurs is analyzed to identify the precursor attributes of the freezing event. The specific parameters analyzed include the rate of change of step length (step length shortened by more than 10%), the rate of change of step frequency (step frequency decreased by more than 20%), and the rate of change of double support time (double support time increased by more than 30%).

[0031] Step S3: classify the precursor attributes of the freezing event into gait gradual change types; perform gait freezing cycle detection on the gait gradual change types according to the freezing occurrence time period to generate a gait freezing cycle; quantify the freezing cumulative frequency of the gait freezing cycle based on the freezing duration to obtain a gait freezing index; In the embodiment of the present invention, first, the precursor attributes of the freezing event are classified to divide the gait gradual change type. By analyzing the changes in the step length and the step frequency in the gait characteristic curve, the gait gradual change type is divided into "slow type" and "fast type". The specific operation is: calculate the step length shortening rate and the step frequency reduction rate, the step length shortening rate is obtained by comparing the step length changes before and after freezing, and the step frequency reduction rate is obtained by comparing the step frequency changes before and after freezing. When the step length shortening rate is less than 10% and the step frequency reduction rate is less than 20%, it is determined to be a "slow type" gait gradual change; when the step length shortening rate is greater than or equal to 10% and the step frequency reduction rate is greater than or equal to 20%, it is determined to be a "fast type" gait gradual change. Subsequently, the gait gradual change type is subjected to gait freezing cycle detection according to the freezing time period. The starting point and the end point of the gait freezing cycle are determined by using the time series analysis technology and combining the periodic changes of the gait characteristic curve. The specific operation is: in the gait characteristic curve, the starting point and the end point of the freezing cycle are identified by detecting the significant change points of the step length and the step frequency. The starting point is defined as the moment when the stride length shortening rate exceeds 10% and the step frequency decrease rate exceeds 20%; the ending point is defined as the moment when the stride length recovers to 80% of the time before shortening and the step frequency recovers to 80% of the time before decreasing. In this way, the gait freezing cycle is extracted from the gait characteristic curve, and the duration of each freezing cycle is recorded. Finally, the cumulative freezing frequency of the gait freezing cycle is quantified based on the freezing duration to obtain the gait freezing index. The specific operation is as follows: the freezing duration in each gait freezing cycle is counted, and its proportion in the total gait cycle is calculated. The proportion values ​​of all gait freezing cycles are accumulated to obtain the gait freezing index. This index reflects the cumulative frequency and severity of gait freezing in patients within a certain period of time, and can be used to quantitatively evaluate the gait impairment of Parkinson's patients.

[0032] Step S4: determining the gait change trend based on the freezing cycle data and the gait freezing index, and evaluating the degree of gait impairment of Parkinson's disease on the gait change trend, and generating a gait impairment evaluation report.

[0033] In an embodiment of the present invention, the freezing cycle data is analyzed to extract the key features of the gait change trend. The gait freezing index (FI) is used as a quantitative index to determine the gait change trend by comparing the changes in the gait parameters of the patient before and after the freezing event. The specific operation includes: calculating the average value and coefficient of variation of parameters such as step length, step speed, and step frequency during the freezing cycle, and analyzing the change range of these parameters before and after the freezing event. For example, indicators such as the step length shortening rate, the step speed reduction rate, and the increase rate of the double support time can be used to reflect the trend of gait change. Secondly, the degree of gait impairment in Parkinson's disease is evaluated according to the gait change trend. The multi-output random forest algorithm (MRFR) is used to classify and quantify the degree of gait impairment. The algorithm divides patients into three categories of mild, moderate and severe impairment by analyzing the change pattern of gait parameters. The specific operation is: inputting key features such as gait freezing index, step length shortening rate, and step speed reduction rate, and evaluating the degree of gait impairment of patients through the classification model obtained by algorithm training. Finally, a gait impairment assessment report is generated. The assessment report includes the patient's basic information, gait freezing index, gait change trend analysis results and classification results of the degree of gait impairment. The report records in detail the changes in gait parameters, such as stride length shortening rate, gait speed decrease rate, etc.

[0034] As an example of the present invention, refer to Figure 2 As shown, in this example, step S1 includes: Step S11: setting the frame rate of the binocular camera to 40-60 frames / second, the resolution to 1920×1080 pixels, and setting the angle between the binocular camera and the patient's walking path to 45°; Step S12: collecting the complete walking cycle of the patient each time, including the initial gait, the middle gait and the final gait, and each collection lasts for 3 minutes; Step S13: activating the automatic tracking function of the binocular camera to continuously monitor the walking path, number of steps, walking speed and walking movements of the Parkinson's patient to obtain a gait image; Step S14: recording the number of swings of the patient's gait image, marking the patient's gait swing number as zero for two consecutive cycles; determining the gait interruption point for the two consecutive cycles of zero times, and obtaining gait continuity interruption data; Step S15: performing gait freezing judgment on the gait image according to the gait continuity interruption data, obtaining a gait freezing event, and recording the freezing event characteristics, freezing occurrence time period and freezing duration of the gait freezing event.

[0035] In the embodiment of the present invention, first, according to the requirements of step S11, the frame rate of the binocular camera is set to 50 frames per second, the resolution is 1920×1080 pixels, and the angle between the binocular camera and the patient's walking path is set to 45°. This setting can ensure that a clear and comprehensive gait image is obtained during the acquisition process, providing a high-quality data basis for subsequent analysis. Then, in step S12, the complete cycle of each walking of the patient is collected, including the starting gait, the intermediate gait and the ending gait, and each acquisition lasts for 3 minutes. In this way, the gait characteristics of the patient at different walking stages can be fully recorded to ensure the integrity and representativeness of the data. In step S13, the automatic tracking function of the binocular camera is started to continuously monitor the walking path, walking steps, walking speed and walking movements of the Parkinson's patient to obtain a gait image. The automatic tracking function can capture the patient's motion trajectory in real time to ensure that key information is not lost during the acquisition process. In the specific operation, the depth information and visual tracking algorithm of the binocular camera are used to track the patient's limb movements in real time and record their gait characteristics. Subsequently, in step S14, the number of swings of the patient's gait image is recorded, and the number of swings of the patient's gait is marked as zero for two consecutive cycles. By analyzing the limb movement data in the gait image, the situation where the number of swings is zero in two consecutive cycles is identified and marked as a potential gait interruption point. These interruption points are an important basis for judging the interruption of gait continuity. In the specific operation, the gait image is processed by gait analysis software, the number of swings of each step is extracted, and the situation of zero for two consecutive cycles is determined by time series analysis. Finally, in step S15, the gait freezing judgment is performed on the gait image according to the gait continuity interruption data, and the gait freezing event is obtained, and the freezing event characteristics, freezing occurrence time period and freezing duration of the gait freezing event are recorded. By analyzing the time-space gait parameter changes before and after the interruption point, the specific characteristics of the gait freezing event are determined, including the time point when the freezing occurs, the duration of the freezing, and the changes in the gait parameters during the freezing period. These data provide key information for the subsequent gait impairment assessment. In specific operations, the gait analysis system is used to analyze the time series data before and after the interruption point, and the change rate of the gait parameters is calculated to determine the characteristics of the gait freezing event.

[0036] Determining gait characteristic parameters based on the gait image includes: Divide the gait image into a plurality of continuous frame sequences, wherein each frame sequence corresponds to a gait cycle; In each frame sequence, the foot joint, knee joint and ankle joint are extracted to obtain the human joint point information; the displacement time change of the human joint point information in adjacent frames is calculated to generate the gait motion cycle; Calculate the distance between the foot joint and the contact point in two consecutive gait motion cycles to obtain the step length data; measure the lateral distance between the left and right foot joints and the contact point to obtain the step width data; Calculate gait speed using step length data and gait motion cycle; determine gait symmetry for gait motion cycle based on gait speed and step width data; Calculate the flexion and extension phase angles of the knee joint and the ankle joint in two consecutive gait motion cycles, and map the flexion and extension phase angles to flexion and extension discreteness to obtain joint flexion and extension discreteness; The gait motion cycle, step length data, step width data, gait speed, gait symmetry and joint flexion and extension discreteness are combined into gait characteristic parameters.

[0037] In an embodiment of the present invention, the gait image is divided into a plurality of continuous frame sequences, each frame sequence corresponding to a gait cycle. Computer vision technology is used in combination with the OpenPose framework to perform human posture recognition on the gait image, and the position information of the foot joint, knee joint and ankle joint is extracted. By detecting the key points in each frame image, the two-dimensional coordinate data of the human joint points is obtained. Then, the displacement time change of the human joint point information in the adjacent frames is calculated to generate the gait motion cycle. The specific method is to calculate the displacement of the key points in each frame, and convert the displacement into time series data in combination with the frame rate (such as 50 frames / second), so as to generate the gait motion cycle. Subsequently, the distance between the foot joint and the contact point in two consecutive gait motion cycles is calculated to obtain the step length data. At the same time, the lateral distance between the left and right foot joints and the contact point is measured to obtain the step width data. The step length is obtained by calculating the distance between the center points of the two heels, and the step width is calculated by measuring the lateral distance between the center points of the two ankles. The step length data and the gait motion cycle are used to measure the gait speed, and the specific formula is the step length divided by the time of the gait motion cycle. The gait symmetry is determined based on the gait speed and step width data for the gait motion cycle, and the gait symmetry is evaluated by calculating the difference in the length and step width of the left and right steps. Furthermore, the flexion and extension phase angles of the knee and ankle joints in two consecutive gait motion cycles are calculated, and the flexion and extension phase angles are mapped to the flexion and extension discreteness to obtain the joint flexion and extension discreteness. The flexion and extension discreteness of the joints is quantified by analyzing the range of change and degree of fluctuation of the joint angles. Finally, the gait motion cycle, step length data, step width data, gait speed, gait symmetry, and joint flexion and extension discreteness are combined into gait characteristic parameters, which are integrated through the gait analysis system and used for subsequent gait impairment assessment.

[0038] The step of extracting the gait spatiotemporal characteristic parameters of the gait characteristic parameters and constructing a gait characteristic curve according to the gait spatiotemporal characteristic parameters comprises: Extracting the periodic time information of the gait characteristic parameters, and dividing the periodic time information into a time period sequence; performing sequence parameterization on the time period sequence to obtain the time period sequence parameters; Based on the time period sequence parameters, gait spatial feature recognition is performed on gait feature parameters, and feature parameters of gait spatial features are quantified to obtain time period sequence parameters; Arrange the time period series parameters in chronological order to form chronological order data; The time-arranged sequence data is used as the horizontal axis and the time period sequence parameters are used as the vertical axis to construct the initial gait characteristic curve; The initial gait characteristic curve is processed for curve smoothness to generate a gait characteristic curve.

[0039] In an embodiment of the present invention, the gait image is processed by computer vision technology, the cycle time information of the gait characteristic parameters is extracted, and the gait characteristic parameters are divided into a time period sequence; specifically, the gait image is subjected to human posture recognition through the OpenPose framework, and the position information of the foot joint, knee joint and ankle joint is extracted. Based on these key point information, the change of the gait characteristic parameters over time is calculated, and the time series data of the gait motion cycle is generated; then, the time period sequence is parameterized to obtain the time period sequence parameters. The specific method is to calculate the parameters such as the step length, step width, and gait speed in each time period, and use these parameters as the characteristic values ​​of the time period sequence. For example, the step length is obtained by calculating the distance between the center points of the two ankles, and the step width is calculated by measuring the lateral distance between the center points of the two ankles. Subsequently, the gait spatial feature recognition is performed on the gait characteristic parameters based on the time period sequence parameters, and the characteristic parameters of the gait spatial feature are quantified. By analyzing the changes of the gait characteristic parameters in different time periods, indicators such as gait symmetry and joint flexion and extension discreteness are calculated. For example, gait symmetry can be evaluated by calculating the difference in left and right footstep length and width, and joint flexion and extension dispersion can be quantified by analyzing the range and fluctuation of the flexion and extension phase angles of the knee and ankle joints. Afterwards, the time period sequence parameters are arranged in chronological order to form time-ordered data. These data reflect the changing trend of gait characteristic parameters over time, providing a basis for the subsequent construction of gait characteristic curves. Finally, the initial gait characteristic curve is constructed with the time-ordered data as the horizontal axis and the time period sequence parameters as the vertical axis. The initial gait characteristic curve is processed for curve smoothness, and the Savitzky-Golay filter is used to smooth the curve. The filter fits the data through local polynomial regression, which can smooth the noise while retaining the key features of the signal. The specific parameters are set as a window size of 21 and a polynomial order of 2.

[0040] The preceding attributes of identifying the freezing event characteristics according to the gait characteristic curve include: On the gait characteristic curve, mark the time point when the gait speed drops below 0.5 m / s as the starting point of the speed freezing event; calculate the rate of change of gait speed within 3 seconds before and after the marked point based on the starting point of the speed freezing event. If the rate of change is less than -0.2 m / s², it is identified as a speed precursor feature; On the gait characteristic curve, mark the time point when the gait symmetry index drops below 0.8 as the starting point of the symmetry freezing event; compare the change in gait symmetry within 5 seconds before and after the marked point based on the starting point of the symmetry freezing event. If the symmetry change exceeds 0.2, record the time interval of the gait symmetry change. If the interval is less than 1 second, it is identified as a symmetry precursor feature. On the gait characteristic curve, mark the time point when the stride length is shortened to less than 68 cm as the starting point of the stride length freezing event; compare the change in stride length within 3 seconds before and after the marked point based on the starting point of the stride length freezing event. If the stride length is shortened by more than 5 cm, it is identified as a stride length precursor feature; On the gait characteristic curve, mark the time point when the step width changes by more than 3.5 cm as the starting point of the step width freezing event; compare the change of step width within 3 seconds before and after the marked point based on the starting point of the step width freezing event. If the step width changes by more than 2 cm, it is identified as a step width precursor feature; On the gait characteristic curve, mark the time point when the joint flexion and extension discreteness changes by more than 10% as the starting point of the flexion and extension freezing event; compare the change of joint flexion and extension discreteness within 3 seconds before and after the marked point based on the starting point of the flexion and extension freezing event. If the change exceeds 5%, it is identified as the precursor feature of joint flexion and extension; The speed precursor features, symmetry precursor features, step width precursor features and joint flexion and extension precursor features are combined to obtain the freezing event precursor attributes.

[0041] In an embodiment of the present invention, key indicators are extracted from the gait characteristic curve. The OpenPose framework is used to perform human posture recognition on gait video data, and the position information of key points such as ankles, knee joints and ankle joints is extracted. Based on these key point information, gait characteristic parameters are calculated, including gait speed, gait symmetry, step length, step width and joint flexion and extension discreteness. On the gait characteristic curve, when the gait speed drops below 0.5 m / s, the time point is marked as the starting point of the speed freezing event. For the marked point, the rate of change of the gait speed within 3 seconds before and after is calculated. The specific operation is to calculate the change in speed by the difference method divided by the time interval. If the rate of change is less than -0.2 m / s², it is identified as a speed precursor feature. On the gait characteristic curve, when the gait symmetry index drops below 0.8, the time point is marked as the starting point of the symmetry freezing event. Based on the starting point, the change in gait symmetry within 5 seconds before and after the marked point is compared. If the symmetry change exceeds 0.2, the time interval of the gait symmetry change is recorded. If the interval is less than 1 second, it is considered a symmetrical precursor feature. On the gait characteristic curve, when the stride length is shortened to less than 68 cm, mark this time point as the starting point of the stride length freezing event. For this starting point, compare the change in stride length within 3 seconds before and after the marked point. If the stride length is shortened by more than 5 cm, it is considered a stride length precursor feature. On the gait characteristic curve, when the step width changes by more than 3.5 cm, mark this time point as the starting point of the step width freezing event. For this starting point, compare the change in step width within 3 seconds before and after the marked point. If the step width changes by more than 2 cm, it is considered a step width precursor feature. On the gait characteristic curve, when the joint flexion and extension discreteness changes by more than 10%, mark this time point as the starting point of the flexion and extension freezing event. For this starting point, compare the change in joint flexion and extension discreteness within 3 seconds before and after the marked point. If the change exceeds 5%, it is considered a joint flexion and extension precursor feature.

[0042] Preferably, step S3 comprises the following steps: Step S31: extracting gait interruption precursor features from the freezing event precursor attributes to obtain gait interruption precursor features; determining the interruption type from the gait interruption precursor features to generate a gait interruption type; Step S32: performing inter-type similarity comparison on the gait interruption types to obtain inter-type similarity data; and classifying the precursor attributes of the freezing event into gait gradual change types according to the inter-type similarity data; Step S33: segmenting the freezing time period according to the freezing event characteristics, taking the freezing event characteristic occurrence point as the center, extending forward and backward by 5 seconds each, and dividing it into a pre-freezing period and a post-freezing period; Step S34: marking the gait type normal interval according to the pre-freezing interval for the gait gradual change type; marking the gait type frozen interval according to the post-freezing interval; Step S35: gait freezing period is calculated based on the normal gait type interval and the gait type freezing interval to generate a gait freezing period; Step S36: quantifying the cumulative freezing frequency of the gait freezing period based on the freezing duration to obtain a gait freezing index.

[0043] In an embodiment of the present invention, a gait interruption precursor feature is extracted from a gait characteristic curve. Using time series analysis technology, key indicators in the gait characteristic curve (such as gait speed, gait symmetry, step length, step width, and joint flexion and extension discreteness) are monitored. When the gait speed drops below 0.5 m / s, the time point is marked as the starting point of the speed freezing event, and the rate of change of the speed within 3 seconds before and after is calculated. If the rate of change is less than -0.2 m / s², it is identified as a speed precursor feature. At the same time, when the gait symmetry index drops below 0.8, it is marked as the starting point of the symmetry freezing event, and the change in symmetry within 5 seconds before and after the marked point is compared. If the change exceeds 0.2 and the time interval is less than 1 second, it is identified as a symmetry precursor feature. In addition, when the step length is shortened to less than 68 cm, it is marked as the starting point of the step length freezing event; when the step width changes by more than 3.5 cm, it is marked as the starting point of the step width freezing event. Compare the changes within 3 seconds before and after the marked point. If the step length is shortened by more than 5 cm or the step width changes by more than 2 cm, it is identified as the corresponding precursor feature. Finally, the above precursor features are integrated to determine the type of gait interruption. Based on the extracted gait interruption type, the similarity between types is calculated. By calculating the Euclidean distance or cosine similarity between different types of precursor features, similar data between types are generated. Based on the similar data, the precursor attributes of the freezing event are divided into different gait gradient types, such as "slow type" and "fast type". According to the freezing event characteristics, with the occurrence point of the freezing event as the center, it extends forward and backward for 5 seconds each, and is divided into the pre-freezing interval and the post-freezing interval. This operation is achieved through time series analysis to ensure that the data intervals before and after the freezing can completely cover the start and end of the freezing event. In the pre-freezing interval, the gait gradient type is marked to determine the normal interval of the gait type; in the post-freezing interval, the gait type freezing interval is marked. This process is achieved by comparing the changes in gait characteristic parameters before and after freezing, such as gait speed, step length, and gait symmetry; comparing and analyzing the normal interval of gait type with the freezing interval of gait type, and generating a gait freezing cycle by calculating the frequency of change of gait characteristic parameters in the two intervals. The specific operation includes calculating the rate of change of gait speed, step length, and gait symmetry before and after freezing, and determining the freezing cycle based on this. Based on the duration of freezing, the gait freezing cycle is quantitatively analyzed. By counting the duration of each gait freezing cycle and calculating its proportion in the total gait cycle, this index reflects the cumulative frequency and severity of gait freezing events, providing a quantitative basis for the assessment of gait impairment in Parkinson's disease.

[0044] Preferably, step S36 includes the following steps: Step S361: performing periodic segmentation processing on the freezing duration to obtain the freezing type segmentation duration; Step S362: performing duration normalization processing on the frozen type segmentation duration to obtain duration normalized data; Step S363: performing freezing frequency statistics on the gait freezing period according to the normalized duration data, and determining the occurrence frequency of each freezing period to obtain freezing frequency distribution data; Step S364: performing cumulative calculation on the freezing frequency distribution data, and accumulating the frequencies of each freezing period in chronological order to form a freezing cumulative frequency curve; Step S365: performing frozen frequency parameter mapping on the frozen cumulative frequency curve, and dividing the frozen frequency parameter into a plurality of quantization intervals, each interval corresponding to a quantization value, to generate a frozen frequency quantization parameter; Step S366: performing weighted calculation on the freezing frequency quantization parameters to obtain a final gait freezing index.

[0045] In an embodiment of the present invention, the duration of the freezing event is periodically segmented. The duration of the freezing event is divided into a plurality of independent freezing periods by a time series analysis technique. The duration of each freezing period starts from the starting point of the freezing event and ends at the ending point of the freezing event. The purpose of the segmentation process is to decompose the continuous freezing event into a plurality of independent segments for subsequent analysis. The segmented freezing period duration is normalized. The purpose of the normalization process is to unify freezing periods of different lengths into a standardized range for subsequent statistical analysis. The specific operation is to divide the duration of each freezing period by the sum of the durations of all freezing periods to obtain the normalized duration data. Based on the normalized duration data, the frequency of the gait freezing period is counted. By counting the occurrence frequency of each freezing period, the freezing frequency distribution data is generated. The specific operation is to calculate the proportion of each freezing period in the total gait period and record it as the freezing frequency distribution data. The freezing frequency distribution data is cumulatively calculated, and the frequencies of each freezing period are accumulated in chronological order to form a freezing cumulative frequency curve. The cumulative frequency curve reflects the cumulative occurrence probability of freezing events in the gait cycle and can intuitively show the severity of freezing events. The freezing cumulative frequency curve is parameter mapped and the freezing frequency parameters are divided into multiple quantitative intervals. Each interval corresponds to a quantitative value, which is used to quantitatively evaluate the severity of the freezing event. For example, the freezing frequency parameter can be divided into three quantitative intervals: mild (0-20%), moderate (20%-50%), and severe (50%-100%). Finally, the freezing frequency quantitative parameters are weighted and calculated to obtain the final gait freezing index. The purpose of weighted calculation is to comprehensively consider the severity of freezing events in different quantitative intervals and generate a quantitative index that can comprehensively reflect the gait freezing situation. The specific operation is to weight the freezing frequency quantitative parameters according to the weight of each quantitative interval, and finally obtain the gait freezing index.

[0046] As an example of the present invention, refer to Figure 3 As shown, in this example, step S4 includes: Step S41: Arrange the freezing cycle data in chronological order, and draw a freezing cycle curve with the starting time of the freezing cycle as the horizontal axis and the freezing duration as the vertical axis; Step S42: Calculate the slope of the freezing cycle curve according to the gait freezing index. When the slope is greater than 0, it is determined that the gait freezing is increasing; when the slope is less than 0, it is determined that the gait freezing is decreasing. Step S43: merging the gait freezing enhancement trend and the gait freezing reduction trend into gait trend features to generate gait change trend data; Step S44: dividing the gait change trend into three intervals of mild, moderate and severe gait impairment, and marking them with different colors to obtain gait change impairment colors; Step S45: Mapping the color of the gait change impairment to the degree of gait impairment of Parkinson's disease, and visualizing the degree of gait impairment to generate a gait impairment assessment report.

[0047] In the embodiment of the present invention, the freezing cycle data are arranged in chronological order, and the freezing cycle curve is drawn with the starting time of the freezing cycle as the horizontal coordinate and the freezing duration as the vertical coordinate. The starting time and duration of each freezing cycle are extracted by time series analysis technology and visualized on a two-dimensional plane. This process utilizes the gait freezing index and the time series analysis method, which can intuitively display the time distribution of freezing events. The slope of the freezing cycle curve is calculated according to the gait freezing index. The specific operation is to fit the freezing cycle curve by the linear regression method and calculate its slope value. When the slope is greater than 0, it is determined to be a gait freezing strengthening trend; when the slope is less than 0, it is determined to be a gait freezing weakening trend. This determination process is based on the change trend of the freezing cycle curve and can reflect the dynamic change of gait freezing. The gait freezing strengthening trend and the gait freezing weakening trend are merged to generate gait change trend data. By integrating the characteristic parameters of the strengthening and weakening trends, a comprehensive gait change trend data set is formed to provide a basis for subsequent evaluation. The gait change trend is divided into three intervals of mild, moderate and severe gait impairment, and marked with different colors respectively to obtain the gait change impairment color. The specific operation is to divide the gait change trend into three intervals according to the range of the gait freezing index: mild impairment (0-33%), moderate impairment (34%-66%), and severe impairment (67%-100%), and mark them with green, yellow, and red respectively. The impaired color of gait changes is mapped to the degree of gait impairment in Parkinson's disease, and the degree of gait impairment is visualized to generate a gait impairment assessment report. Through visualization technology, the impaired color of gait changes is mapped to the assessment report to intuitively show the patient's gait impairment.

[0048] It is particularly important that step S44 includes the following steps: Step S44: dividing the gait change trend into three intervals of mild, moderate and severe gait impairment, and marking them with different colors to obtain gait change impairment colors; Step S441: classify the gait change trend into gait impairment. If the gait movement cycle change is less than 10%, the step length change is less than 5% and the difference between the left and right step lengths is less than 3%, the step width change is less than 2 cm, the gait speed decrease is less than 10%, the gait symmetry index is greater than 0.9, and the joint flexion and extension dispersion is less than 10°, then it is judged that the gait is mildly impaired; Step S442: the gait change trend is classified into gait impairment. If the gait movement cycle changes between 10% and 20%, the step length changes between 5% and 10% and the difference between the left and right step lengths is between 3% and 5%, the step width changes between 2 cm and 4 cm, the gait speed decreases between 10% and 20%, the gait symmetry index is between 0.7 and 0.9, and the joint flexion and extension dispersion is between 10° and 20°, it is judged as moderate gait impairment. Step S443: The gait change trend is classified into gait impairment. If the gait movement cycle change is greater than 20%, the step length change is greater than 10% and the difference between the left and right step lengths is greater than 5%, the step width change is greater than 4 cm, the gait speed decreases by more than 20%, the gait symmetry index is less than 0.7, and the joint flexion and extension dispersion is greater than 20°, then it is judged as severe gait impairment; Step S444: Mark mild gait impairment as light green, moderate gait impairment as yellow, and severe gait impairment as red, so as to obtain the gait variation impairment color.

[0049] In an embodiment of the present invention, a high-precision gait analysis system, such as the GAITRite electronic carpet system, is used in combination with a wearable inertial sensor. The GAITRite system can record the time-space parameters of gait in real time, including gait motion cycle, step length, step width, gait speed, etc.; the inertial sensor is used to measure the change of joint angle to obtain the discreteness of joint flexion and extension. The subject walks on the GAITRite electronic carpet with a natural gait, and the system automatically records the gait data. At the same time, the inertial sensor worn by the subject synchronously collects the joint angle data. During the collection process, the integrity and accuracy of the data are ensured to avoid data loss due to equipment failure or abnormal walking of the subject. The time-space parameters collected by the GAITRite system are time-aligned with the joint angle data collected by the inertial sensor to ensure the consistency of the two sets of data in time. The joint angle data collected by the inertial sensor is filtered to remove high-frequency noise and retain effective joint angle change information. A low-pass filter is used with a cutoff frequency set to 10 Hz to smooth the data and reduce interference. The timestamps of the start and end points of each step are extracted from the GAITRite system to calculate the duration of the gait cycle. Based on the footprint position recorded by the GAITRite system, measure the straight-line distance between the starting and ending points of each step. Calculate the stride length of the left and right feet separately, and find the ratio of the difference between the two to the average stride length. Step width: measure the horizontal distance between the center points of the two feet. Gait speed: determine the ratio of step length to gait cycle Gait symmetry index: compare the ratio of the difference in gait parameters of the left and right feet to the sum. Joint flexion and extension dispersion: extract the standard deviation of joint angle changes from inertial sensor data. Mild impairment: when the gait movement cycle changes less than 10%, the step length changes less than 5% and the difference between the left and right step lengths is less than 3%, the step width changes less than 2cm, the gait speed decreases less than 10%, the gait symmetry index is greater than 0.9, and the joint flexion and extension dispersion is less than 10°, it is marked as light green. Moderate impairment: When the gait cycle changes between 10% and 20%, the step length changes between 5% and 10% and the difference between the left and right step lengths is between 3% and 5%, the step width changes between 2 cm and 4 cm, the gait speed decreases between 10% and 20%, the gait symmetry index is between 0.7 and 0.9, and the joint flexion and extension discreteness is between 10° and 20°, it is marked as yellow. Severe impairment: When the gait cycle changes by more than 20%, the step length changes by more than 10% and the difference between the left and right step lengths is more than 5%, the step width changes by more than 4 cm, the gait speed decreases by more than 20%, the gait symmetry index is less than 0.7, and the joint flexion and extension discreteness is greater than 20°, it is marked as red. Use Python's Matplotlib library or professional gait analysis software (such as the analysis software provided by GAITRite) for data visualization. Color marking: According to the degree of impairment, the gait data is marked with different colors in the visualization interface. Mildly damaged areas are marked in light green, moderately damaged areas are marked in yellow, and severely damaged areas are marked in red.Result display: Generate time domain curve graphs and heat maps of gait parameters to intuitively display the changing trend of the degree of gait impairment.

[0050] The present invention also provides a model for establishing a Parkinson's disease gait impairment assessment model, which is used for the above-mentioned method for establishing a Parkinson's disease gait impairment assessment model. The model for establishing a Parkinson's disease gait impairment assessment model comprises: An image monitoring module is used to continuously monitor the gait images of Parkinson's patients; when a gait continuity interruption in the gait image is detected, it is determined as a gait freezing event, and the freezing event characteristics, freezing occurrence time period and freezing duration of the gait freezing event are recorded; A gait characteristic curve construction module is used to determine gait characteristic parameters based on gait images; extract gait spatiotemporal characteristic parameters of the gait characteristic parameters, and construct a gait characteristic curve according to the gait spatiotemporal characteristic parameters; identify the precursor attributes of the freezing event characteristics according to the gait characteristic curve, and obtain the precursor attributes of the freezing event; The gait freezing quantification module is used to classify the precursor attributes of the freezing event into gait gradual change types; perform gait freezing cycle detection on the gait gradual change types according to the freezing time period to generate the gait freezing cycle; perform freezing cumulative frequency quantification on the gait freezing cycle based on the freezing duration to obtain the gait freezing index; The impairment degree assessment module is used to determine the gait change trend based on the freezing cycle data and the gait freezing index, and to assess the gait impairment degree of Parkinson's disease on the gait change trend, and to generate a gait impairment assessment report.

[0051] The image monitoring module of the present invention monitors the gait of Parkinson's patients in real time through a high-resolution image acquisition device, ensuring that a complete gait image sequence can be captured at any time point, and can continuously record the patient's walking process, providing a comprehensive data basis for subsequent analysis; the gait image is analyzed based on the image processing algorithm, and when a gait continuity interruption is detected, it is automatically determined as a gait freezing event. At the same time, the characteristics of the freezing event are recorded, including the time period when the freezing occurs, the duration of the freezing, and the morphological characteristics of the freezing event, ensuring the accuracy and integrity of the data and avoiding errors caused by manual intervention. Through the gait characteristic curve construction module, the gait spatiotemporal characteristic parameters of the gait are extracted from the gait image, including key parameters such as step length, step width, step speed, and gait cycle. These parameters are calculated through image analysis algorithms, ensuring the objectivity and repeatability of the data. A gait characteristic curve is constructed based on the extracted spatiotemporal characteristic parameters of gait. The curve takes time as the horizontal axis and gait characteristic parameters as the vertical axis, which can intuitively reflect the changing trend of gait parameters over time. Through this curve processing, the system can capture the dynamic process of gait changes. By analyzing the gait characteristic curve, the characteristic changes before the freezing event, that is, the precursor attributes of the freezing event, are identified. This identification process is based on the morphological analysis of the curve and the parameter change trend, and can accurately locate the precursor characteristics of the freezing event. Through the gait freezing quantification module, the precursor attributes of the freezing event are divided into different gait gradient types. This division is based on the morphological characteristics and parameter change rules of the gait characteristic curve, and can classify and process complex gait changes; and according to the time period when the freezing occurs, the gait gradient type is periodically detected. By analyzing the periodic changes of the gait characteristic curve, the periodic pattern of gait freezing can be accurately identified, thereby revealing the regularity of gait freezing events; the gait freezing cycle is quantitatively analyzed based on the duration of freezing, and the cumulative frequency of the freezing event and the weighted sum of the duration are calculated to comprehensively reflect the severity of gait freezing, providing a unified quantitative standard for the assessment of the degree of gait impairment. Through the impairment assessment module, based on the freezing cycle data and gait freezing index, the trend of gait changes is analyzed through a mathematical model, which can comprehensively consider the regularity of the freezing cycle and the changing trend of the freezing index, so as to accurately judge the overall situation of gait impairment; according to the gait change trend, the degree of gait impairment is quantitatively assessed. This assessment process is based on the preset quantitative standards, and by calculating the difference between the gait freezing index and the normal gait parameters, it can provide objective and quantitative assessment results; based on the assessment results, a detailed gait impairment assessment report is generated, which contains detailed records of freezing events, gait characteristic curves, freezing cycle analysis results, and quantitative assessment of gait impairment. The report is presented in the form of charts and data, which can intuitively show the details of gait changes and provide comprehensive data support for related research.Therefore, the present invention realizes gait monitoring of Parkinson's patients and identifies gait freezing characteristics through data processing technology and image processing technology, so as to improve the accuracy of Parkinson's disease gait impairment assessment and shorten the time used in the impairment assessment process.

[0052] A computer-readable storage medium stores a computer program, wherein the computer program is used to execute the method for establishing a Parkinson's disease gait impairment assessment method.

[0053] Therefore, the embodiments should be regarded as illustrative and non-restrictive from all points, and the scope of the present invention is limited by the appended claims rather than the above description, and it is therefore intended that all changes falling within the meaning and range of equivalent elements of the application documents are included in the present invention.

[0054] The above description is only a specific embodiment of the present invention, so that those skilled in the art can understand or implement the present invention. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but should conform to the widest scope consistent with the principles and novel features invented herein.

Claims

1. A method for establishing a Parkinson's disease gait impairment assessment method, characterized in that: The following steps are involved: Step S1: continuously monitoring the gait image of the Parkinson's disease patient; when the gait continuity of the gait image is detected to be interrupted, it is determined as a gait freezing event, and the freezing event characteristics, freezing occurrence time period and freezing duration of the gait freezing event are recorded; Step S2: determining gait characteristic parameters based on the gait image; extracting gait spatiotemporal characteristic parameters of the gait characteristic parameters, and constructing a gait characteristic curve according to the gait spatiotemporal characteristic parameters; identifying the precursor attribute of the freezing event feature according to the gait characteristic curve, and obtaining the precursor attribute of the freezing event; Step S3: Classify the freezing event precursor attributes into gait gradient types; Performing gait freezing cycle detection on the gait gradual change type according to the freezing occurrence time period, and generating a gait freezing cycle; The cumulative freezing frequency of gait freezing cycles was quantified based on the freezing duration to obtain the gait freezing index. Step S4: determining the gait change trend based on the freezing cycle data and the gait freezing index, and evaluating the degree of gait impairment of Parkinson's disease on the gait change trend, and generating a gait impairment evaluation report.

2. The method for establishing a Parkinson's disease gait impairment assessment method according to claim 1, characterized in that: Step S1 includes the following steps: Step S11: setting the frame rate of the binocular camera to 40-60 frames / second, the resolution to 1920×1080 pixels, and setting the angle between the binocular camera and the patient's walking path to 45°; Step S12: collecting the complete cycle of each walking of the patient, including the initial gait, the middle gait and the final gait, and each collection lasts for 3 minutes; Step S13: activating the automatic tracking function of the binocular camera to continuously monitor the walking path, number of steps, walking speed and walking movements of the Parkinson's patient to obtain a gait image; Step S14: recording the number of swings of the patient's gait image, marking the patient's gait swing number as zero for two consecutive cycles; determining the gait interruption point for the two consecutive cycles of zero times, and obtaining gait continuity interruption data; Step S15: performing gait freezing judgment on the gait image according to the gait continuity interruption data, obtaining a gait freezing event, and recording the freezing event characteristics, freezing occurrence time period and freezing duration of the gait freezing event.

3. The method for establishing a Parkinson's disease gait impairment assessment method according to claim 1, characterized in that: Determining gait characteristic parameters based on the gait image includes: Divide the gait image into a plurality of continuous frame sequences, wherein each frame sequence corresponds to a gait cycle; In each frame sequence, the foot joint, knee joint and ankle joint are extracted to obtain the human joint point information; the displacement time change of the human joint point information in adjacent frames is calculated to generate the gait motion cycle; Calculate the distance between the foot joint and the contact point in two consecutive gait motion cycles to obtain the step length data; measure the lateral distance between the left and right foot joints and the contact point to obtain the step width data; Calculate gait speed using step length data and gait motion cycle; determine gait symmetry for gait motion cycle based on gait speed and step width data; Calculate the flexion and extension phase angles of the knee joint and the ankle joint in two consecutive gait motion cycles, and map the flexion and extension phase angles to flexion and extension discreteness to obtain joint flexion and extension discreteness; The gait motion cycle, step length data, step width data, gait speed, gait symmetry and joint flexion and extension discreteness are combined into gait characteristic parameters.

4. The method for establishing a Parkinson's disease gait impairment assessment method according to claim 3, characterized in that: The step of extracting the gait spatiotemporal characteristic parameters of the gait characteristic parameters and constructing a gait characteristic curve according to the gait spatiotemporal characteristic parameters comprises: Extracting the periodic time information of the gait characteristic parameters, and dividing the periodic time information into a time period sequence; performing sequence parameterization on the time period sequence to obtain the time period sequence parameters; Based on the time period sequence parameters, gait spatial feature recognition is performed on gait feature parameters, and feature parameters of gait spatial features are quantified to obtain time period sequence parameters; Arrange the time period series parameters in chronological order to form chronological order data; The time-arranged sequence data is used as the horizontal axis and the time period sequence parameters are used as the vertical axis to construct the initial gait characteristic curve; The initial gait characteristic curve is processed for curve smoothness to generate a gait characteristic curve.

5. The method for establishing a Parkinson's disease gait impairment assessment method according to claim 1, characterized in that: The preceding attributes of identifying the freezing event characteristics according to the gait characteristic curve include: On the gait characteristic curve, mark the time point when the gait speed drops below 0.5 m / s as the starting point of the speed freezing event; calculate the rate of change of gait speed within 3 seconds before and after the marked point based on the starting point of the speed freezing event. If the rate of change is less than -0.2 m / s², it is identified as a speed precursor feature; On the gait characteristic curve, mark the time point when the gait symmetry index drops below 0.8 as the starting point of the symmetry freezing event; compare the change in gait symmetry within 5 seconds before and after the marked point based on the starting point of the symmetry freezing event. If the symmetry change exceeds 0.2, record the time interval of the gait symmetry change. If the interval is less than 1 second, it is identified as a symmetry precursor feature. On the gait characteristic curve, mark the time point when the stride length is shortened to less than 68 cm as the starting point of the stride length freezing event; compare the change in stride length within 3 seconds before and after the marked point based on the starting point of the stride length freezing event. If the stride length is shortened by more than 5 cm, it is identified as a stride length precursor feature; On the gait characteristic curve, mark the time point when the step width changes by more than 3.5 cm as the starting point of the step width freezing event; compare the change of step width within 3 seconds before and after the marked point based on the starting point of the step width freezing event. If the step width changes by more than 2 cm, it is identified as a step width precursor feature; On the gait characteristic curve, mark the time point when the joint flexion and extension discreteness changes by more than 10% as the starting point of the flexion and extension freezing event; compare the change of joint flexion and extension discreteness within 3 seconds before and after the marked point based on the starting point of the flexion and extension freezing event. If the change exceeds 5%, it is identified as the precursor feature of joint flexion and extension; The speed precursor features, symmetry precursor features, step width precursor features and joint flexion and extension precursor features are combined to obtain the freezing event precursor attributes.

6. The method for establishing a Parkinson's disease gait impairment assessment method according to claim 1, characterized in that: Step S3 includes the following steps: Step S31: extracting gait interruption precursor features from the freezing event precursor attributes to obtain gait interruption precursor features; determining the interruption type from the gait interruption precursor features to generate a gait interruption type; Step S32: performing inter-type similarity comparison on the gait interruption types to obtain inter-type similarity data; and classifying the precursor attributes of the freezing event into gait gradual change types according to the inter-type similarity data; Step S33: segmenting the freezing time period according to the freezing event characteristics, taking the freezing event characteristic occurrence point as the center, extending forward and backward by 5 seconds each, and dividing it into a pre-freezing period and a post-freezing period; Step S34: marking the gait type normal interval according to the pre-freezing interval for the gait gradual change type; marking the gait type frozen interval according to the post-freezing interval; Step S35: gait freezing period is calculated based on the normal gait type interval and the gait type freezing interval to generate a gait freezing period; Step S36: quantifying the cumulative freezing frequency of the gait freezing period based on the freezing duration to obtain a gait freezing index.

7. The method for establishing a Parkinson's disease gait impairment assessment method according to claim 6, characterized in that: Step S36 includes the following steps: Step S361: performing periodic segmentation processing on the freezing duration to obtain the freezing type segmentation duration; Step S362: performing duration normalization processing on the frozen type segmentation duration to obtain duration normalized data; Step S363: performing freezing frequency statistics on the gait freezing period according to the normalized duration data, and determining the occurrence frequency of each freezing period to obtain freezing frequency distribution data; Step S364: performing cumulative calculation on the freezing frequency distribution data, and accumulating the frequencies of each freezing period in chronological order to form a freezing cumulative frequency curve; Step S365: performing frozen frequency parameter mapping on the frozen cumulative frequency curve, and dividing the frozen frequency parameter into a plurality of quantization intervals, each interval corresponding to a quantization value, to generate a frozen frequency quantization parameter; Step S366: performing weighted calculation on the freezing frequency quantization parameters to obtain a final gait freezing index.

8. The method for establishing a Parkinson's disease gait impairment assessment method according to claim 1, characterized in that: Step S4 includes the following steps: Step S41: Arrange the freezing cycle data in chronological order, and draw a freezing cycle curve with the start time of the freezing cycle as the horizontal axis and the freezing duration as the vertical axis; Step S42: Calculate the slope of the freezing cycle curve according to the gait freezing index. When the slope is greater than 0, it is determined that the gait freezing is increasing; when the slope is less than 0, it is determined that the gait freezing is decreasing. Step S43: merging the gait freezing enhancement trend and the gait freezing reduction trend into gait trend features to generate gait change trend data; Step S44: dividing the gait change trend into three intervals of mild, moderate and severe gait impairment, and marking them with different colors to obtain gait change impairment colors; Step S45: Mapping the color of the gait change impairment to the degree of gait impairment of Parkinson's disease, and visualizing the degree of gait impairment to generate a gait impairment assessment report.

9. A model for establishing gait impairment assessment in Parkinson's disease, characterized in that: Used to execute the method for establishing a Parkinson's disease gait impairment assessment model as claimed in claim 1, wherein establishing a Parkinson's disease gait impairment assessment model comprises: An image monitoring module is used to continuously monitor the gait images of Parkinson's patients; when a gait continuity interruption in the gait image is detected, it is determined as a gait freezing event, and the freezing event characteristics, freezing occurrence time period and freezing duration of the gait freezing event are recorded; A gait characteristic curve construction module is used to determine gait characteristic parameters based on gait images; extract gait spatiotemporal characteristic parameters of the gait characteristic parameters, and construct a gait characteristic curve according to the gait spatiotemporal characteristic parameters; identify the precursor attributes of the freezing event characteristics according to the gait characteristic curve, and obtain the precursor attributes of the freezing event; The gait freezing quantification module is used to classify the precursor attributes of the freezing event into gait gradual change types; perform gait freezing cycle detection on the gait gradual change types according to the freezing time period to generate the gait freezing cycle; perform freezing cumulative frequency quantification on the gait freezing cycle based on the freezing duration to obtain the gait freezing index; The impairment degree assessment module is used to determine the gait change trend based on the freezing cycle data and the gait freezing index, and to assess the gait impairment degree of Parkinson's disease on the gait change trend, and to generate a gait impairment assessment report.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed, the method for establishing a Parkinson's disease gait impairment assessment method as claimed in any one of claims 1 to 8 is implemented.

Images