Stroke early warning system fusing traditional Chinese medicine characteristics, western medicine indexes, psychological states, video behaviors, voice voiceprints and traditional Chinese and western medicine inquiry information

Abstract

The invention provides a stroke early warning system fusing traditional Chinese medicine characteristics, western medicine indexes, psychological states, video behaviors, voice voiceprints and traditional Chinese and western medicine inquiry information, and the system comprises a traditional Chinese medicine characteristic analysis module which carries out the weighted calculation of a stroke risk score based on the quantitative scores of a face image, eye diagnosis, a tongue image and a pulse image; the western medicine index analysis module is used for calculating a stroke risk score based on the electronic medical record, the family medical history and the past medical history; the psychological state analysis module is used for calculating a stroke risk score through characteristics such as low emotion; the video behavior analysis module is used for calculating a stroke risk score through behavior characteristics such as tumble; the voice voiceprint analysis module is used for calculating a stroke risk score through characteristics such as aphasia; the traditional Chinese and western medicine inquiry analysis module is used for calculating a stroke risk score according to symptoms such as dizziness; and finally, the six types of risk scores are integrated through a comprehensive evaluation module, and accurate layered early warning is realized.

Description

[Technical Field]

[0001] This invention relates to the fields of artificial intelligence and medical information processing. Specifically, it relates to a stroke early warning system that integrates traditional Chinese medicine characteristics, Western medicine indicators, psychological state, video behavior, voiceprint, and traditional Chinese and Western medicine consultation. More specifically, it relates to a stroke risk assessment method implemented through a weighted scoring mechanism. [Background Technology]

[0002] Stroke (cerebrovascular accident) is an acute cerebrovascular disease that seriously endangers human health, characterized by its sudden onset, high disability rate, and high mortality rate. Current clinical diagnosis of stroke largely relies on emergency treatment and imaging examinations after symptom onset, which lacks sufficient early warning capabilities and often leads to delays in intervention. While traditional Western medical methods such as the NIHSS score, ABCD2 score, electrocardiogram, CT scan, and MRI play important roles in diagnosis, they are difficult to implement for home-based, community-based, and continuous risk screening. Traditional Chinese medicine has accumulated rich clinical experience in identifying stroke warning signs, but its diagnostic criteria are highly subjective and lack digital representation, making it difficult to effectively integrate into modern intelligent systems.

[0003] Patent document CN119586975A (application number: 202411704696.6) discloses a stroke early warning system based on voice, image, motion data, and electrocardiogram (ECG) data. This system assesses the clarity and fluency of the user's speech through a voice analysis module, detects the accuracy and coherence of finger-to-nose gestures using a video analysis module to determine ataxia, analyzes the symmetry of the corners of the mouth using image recognition technology to assess the degree of facial asymmetry, monitors limb weakness using motion sensor data, and analyzes ECG signals using a neural network model to identify atrial fibrillation risk. Finally, it generates a stroke risk score based on multi-dimensional data. However, this system still has certain limitations, such as not integrating traditional Chinese medicine diagnostic features (e.g., tongue appearance, eye diagnosis) and psychological state assessment, and lacking a dynamic weighted fusion mechanism based on machine learning. [Summary of the Invention]

[0004] To address the shortcomings of existing technologies, the present invention aims to provide a stroke early warning system that integrates traditional Chinese medicine characteristics, Western medicine indicators, psychological state, video behavior, voiceprint, and information from both traditional Chinese and Western medicine consultations. The system integrates multiple information dimensions such as traditional Chinese medicine, Western medicine, behavior, psychology, voiceprint, and consultations, constructs a feature weighting mechanism, and achieves early identification and warning of stroke risk through intra-module scoring and global weighted synthesis.

[0005] A stroke early warning system according to the present invention, integrating traditional Chinese medicine characteristics, Western medicine indicators, psychological state, video behavior, voiceprint, and information from both traditional Chinese and Western medicine consultations, includes:

[0006] The TCM feature analysis module, based on four types of features—facial features, eye diagnosis, tongue appearance, and pulse appearance—quantifies the stroke risk scores of each part and uses a weighted fusion algorithm to perform comprehensive calculations to generate a stroke risk assessment score based on TCM features.

[0007] The Western medicine indicator analysis module, based on three types of clinical data—electronic medical records, family medical history, and past medical history—quantitatively assesses stroke risk scores across various dimensions and uses a weighted fusion algorithm for comprehensive analysis to generate a stroke risk prediction score based on evidence-based medicine.

[0008] The psychological state analysis module is used to identify users' psychological states such as low mood, anxiety, and mood swings, and to calculate stroke risk scores.

[0009] The video behavior analysis module is used to analyze individual behavioral characteristics, such as falls, convulsions, and gait asymmetry, and to calculate stroke risk scores.

[0010] The voiceprint analysis module is used to analyze information such as unclear pronunciation, aphasia, and abnormal tone in speech, and to calculate stroke risk scores;

[0011] The integrated traditional Chinese and Western medicine consultation and analysis module is used to identify subjective discomfort descriptions from users, such as "dizziness" and calculate stroke risk scores.

[0012] The stroke risk comprehensive assessment module is used to assess stroke risk based on six modules: traditional Chinese medicine characteristics, Western medicine indicators, psychological state, video behavior, voiceprint, and traditional Chinese and Western medicine consultation. It comprehensively obtains the overall stroke risk score and realizes risk stratification and early warning.

[0013] Preferably, the TCM feature analysis module includes:

[0014] Based on facial image data captured by a camera, feature parameters with stroke warning characteristics are selected, such as facial color features and facial symmetry features. When facial flushing or asymmetry features are detected, a feature warning mechanism is triggered, and the overall stroke risk score A1 of the facial image is calculated, as shown in the following expression: Where, n 1 S represents the number of facial warning features. 1 This represents the set of all facial features that can trigger an alert. Indicates the weight of facial warning features. This indicates whether facial features trigger an alert; 1 indicates triggering and 0 indicates not triggering.

[0015] Preferably, the TCM feature analysis module includes:

[0016] Based on eye image data captured by a camera, feature parameters with stroke warning characteristics are selected, such as eye color. When red eye color is detected, a feature warning mechanism is triggered. The overall stroke risk score A2 for visual examination is calculated as follows: Where, n 2 S represents the number of eye warning features. 2 This represents the set of all visual diagnostic features that can trigger an alert. Indicates the weight of eye warning features. This indicates whether the eye feature triggers an alert; 1 indicates triggering and 0 indicates not triggering.

[0017] Preferably, the TCM feature analysis module includes:

[0018] Based on tongue image data captured by a camera, feature parameters with stroke warning characteristics are selected, such as tongue deviation and tongue petechiae. When the tongue petechiae feature is detected as present or the tongue deviation feature is detected as deviation, the feature warning mechanism is triggered, and the overall stroke risk score A3 of the tongue image is calculated, as shown in the following expression: Where, n 3 S represents the number of tongue appearance warning features. 3 This represents the set of all tongue features that can trigger an alert. Indicates the weight of tongue appearance warning features. This indicates whether the tongue appearance features trigger an alert; 1 indicates triggering and 0 indicates not triggering.

[0019] Preferably, the TCM feature analysis module includes:

[0020] Based on pulse data collected by wearable devices, characteristic parameters with stroke warning features are selected. For example, when a wiry pulse or a wiry and slippery pulse is detected, a feature warning mechanism is triggered, and the overall stroke risk score A4 of the pulse is calculated, as shown in the following expression: Where, n 4 S represents the number of pulse warning features. 4 This represents the set of all pulse characteristics that can trigger an early warning. Indicates the weight of pulse warning features. This indicates whether the pulse characteristics trigger an alert; 1 indicates triggering and 0 indicates not triggering.

[0021] Preferably, the TCM feature analysis module includes:

[0022] Based on the stroke risk scores {A1, A2, A3, A4} from four aspects—facial appearance, eye diagnosis, tongue appearance, and pulse diagnosis—a weighted average is used to obtain the stroke risk score P1 based on the overall characteristics of traditional Chinese medicine. The expression is as follows: in, The weights of stroke risk scores are determined by four components: facial appearance, eye diagnosis, tongue appearance, and pulse diagnosis.

[0023] Preferably, the Western medicine indicator analysis module includes:

[0024] Based on the collected electronic medical records, indicators with stroke early warning capabilities are selected, such as coagulation function and arachnoid status indicators. When abnormal coagulation function (e.g., elevated INR) or bloody cerebrospinal fluid is detected in subarachnoid hemorrhage, the indicator early warning mechanism is triggered. The overall stroke risk score B1 contained in the electronic medical records is calculated, as shown in the following expression: Where, m 1 T represents the number of early warning indicators for electronic medical records. 1 This represents the set of all electronic medical record indicators that can trigger alerts. Indicates the weight of the early warning indicators for electronic medical records. This indicates whether an electronic medical record indicator has triggered an alert; 1 indicates a triggered alert, and 0 indicates no triggered alert.

[0025] Preferably, the Western medicine indicator analysis module includes:

[0026] Based on collected family medical history, indicators with stroke early warning capabilities are selected. For example, when there is a family history of stroke or cerebral thrombosis, an indicator early warning mechanism is triggered, and the overall stroke risk score B2 of the family medical history is calculated, as shown in the following expression: Where, m 2 Indicates the number of warning indicators related to family medical history, T 2 This represents the set of all family medical history indicators that can trigger an alert. Indicates the weight of the family medical history warning indicator. This indicates whether the family medical history indicator has triggered an alert; 1 indicates a triggered alert, and 0 indicates no triggered alert.

[0027] Preferably, the Western medicine indicator analysis module includes:

[0028] Based on the collected medical history, indicators with stroke early warning capabilities are selected. For example, when the medical history includes coronary heart disease or hypertension, the indicator early warning mechanism is triggered, and the overall stroke risk score B3 of the medical history is calculated, as shown in the following expression: Where, m3 Indicates the number of warning indicators related to family medical history, T 3 This represents the set of all family medical history indicators that can trigger an alert. Indicates the weight of the family medical history warning indicator. This indicates whether the family medical history indicator has triggered an alert; 1 indicates a triggered alert, and 0 indicates no triggered alert.

[0029] Preferably, the Western medicine feature analysis module includes:

[0030] Based on the stroke risk scores {B1, B2, B3} from electronic medical records, family medical history, and past medical history, a weighted average was used to obtain the overall stroke risk score P2 from Western medicine, as expressed below: in, The weights of stroke risk scores are determined by the three components: electronic medical records, family medical history, and past medical history.

[0031] Preferably, the psychological state analysis module includes:

[0032] Based on the collected psychological state data, state parameters with stroke warning characteristics are selected, such as whether the mood is depressed or whether the mood changes rapidly. When a video shows a depressed mood or a rapid change in mood, the warning mechanism is triggered, and the stroke risk score P3 of the overall psychological state is calculated, as shown in the following expression: Where k represents the number of psychological warning states, β represents the set of all psychological states that can trigger a warning, and W K K represents the weight of the psychological warning state. i This indicates whether a psychological state triggers a warning; 1 indicates a trigger and 0 indicates no trigger.

[0033] Preferably, the video behavior analysis module includes:

[0034] Based on video behavior captured by cameras, behavioral parameters with stroke warning characteristics are filtered out, such as whether the gait is symmetrical and whether a fall occurs. When gait asymmetry or sudden fall is observed in the video, the warning mechanism is triggered. The overall stroke risk score P4 of the behaviors contained in the video is calculated, as shown in the following expression: Where z represents the number of warning behaviors in the video, R represents the set of all video behaviors that can trigger warnings, and W R Z represents the weight of the video alert behavior. i This indicates whether the behavior in the video triggered a warning; 1 indicates a warning and 0 indicates no warning.

[0035] Preferably, the voiceprint analysis module includes:

[0036] Based on the collected voiceprints, voiceprint features with stroke warning characteristics are selected, such as whether there is slurred speech, difficulty in understanding, aphasia, or unclear pronunciation. When slurred speech, difficulty in understanding, aphasia, or unclear pronunciation or substitute sounds are present in the video, the warning mechanism is triggered. The overall voiceprint stroke risk score P5 contained in the speech is calculated, as shown in the following expression: Where q represents the number of warning voiceprints in the speech, δ represents the set of all voiceprints that can trigger warnings, and W δ Indicates the voiceprint weight for voice warnings, Q i This indicates whether the voiceprint triggers an alert; 1 indicates triggering and 0 indicates not triggering.

[0037] Preferably, the integrated traditional Chinese and Western medicine consultation and analysis module includes:

[0038] Based on the collected data from both traditional Chinese medicine and Western medicine consultations, physiological characteristics that could indicate stroke were identified. For example, when dizziness or slurred speech was present in the consultation results, an early warning mechanism was triggered. The stroke risk score P6 of the consultation results was calculated, as shown in the following expression: Where d represents the number of warning physiological features in TCM and Western medicine consultations, γ represents the set of all physiological features that can trigger warnings, and W γ D represents the weight of early warning physiological characteristics in traditional Chinese medicine and Western medicine consultations. i This indicates whether physiological characteristics during consultations between traditional Chinese medicine and Western medicine trigger an alert; 1 indicates triggering and 0 indicates not triggering.

[0039] Preferably, the stroke risk comprehensive assessment module includes:

[0040] Based on the stroke risk scores {P1, P2, P3, P4, P5, P6} from six modules—Traditional Chinese Medicine (TCM) feature analysis, Western Medicine (WC) indicator analysis, psychological state analysis, video behavior analysis, voiceprint analysis, and TCM-Western Medicine consultation analysis—a weighted average is used to obtain the overall stroke risk score SS, expressed as follows: SS=W1×P1+W2×P2+W3×P3+W4×P4+W5×P5+W6×P6 Among them, W iThe stroke risk scores of the six modules are weighted and combined to obtain the total stroke risk score SS. Based on the total score SS, the stroke risk grading system, referring to the sensitivity requirements of clinical prediction models such as CHA2DS2-VASc and the FAST principle, sets the warning thresholds as follows: SS < 0.45 is considered low risk (normal), 0.45 ≤ SS < 0.60 is considered medium risk (triggered warning, observation and further testing recommended), and SS ≥ 0.60 is considered high risk (triggered warning and medical intervention recommended). This grading strategy, while ensuring screening sensitivity, follows the principle of "better to have a false positive than a missed diagnosis" in acute stroke identification. [Beneficial Effects]

[0041] Compared with the prior art, the present invention has the following beneficial effects:

[0042] 1. Adopting a multimodal fusion approach to cover multiple dimensions related to stroke risk;

[0043] 2. Introduce a modular weighted scoring mechanism to improve the sensitivity and interpretability of predictions;

[0044] 3. Consider emotional and psychological factors and strengthen the ability to identify prodromal states of stroke;

[0045] 4. The system has low data collection costs and can be deployed in communities, homes, and mobile devices to achieve widespread health screening. [Attached Image Description]

[0046] Other features, objects, and advantages of the present invention will become more apparent from the following detailed description of non-limiting embodiments with reference to the accompanying drawings:

[0047] Figure 1 A schematic diagram of a stroke early warning system that integrates TCM characteristics, Western medicine indicators, psychological state, video behavior, voiceprints, and TCM and Western medicine consultation information.

Detailed Implementation Methods

[0048] The present invention will now be described in detail with reference to specific embodiments. These embodiments will help those skilled in the art to further understand the present invention, but do not limit the invention in any way. It should be noted that those skilled in the art can make several changes and improvements without departing from the concept of the present invention. These all fall within the protection scope of the present invention. Example 1

[0049] According to the present invention, a stroke early warning system integrating traditional Chinese medicine characteristics, Western medicine indicators, psychological state, video behavior, voiceprint, and information from both traditional Chinese and Western medicine consultations is provided, such as... Figure 1 As shown, it includes:

[0050] The TCM feature analysis module, based on four types of features—facial features, eye diagnosis, tongue appearance, and pulse appearance—quantifies the stroke risk scores of each part and uses a weighted fusion algorithm to perform comprehensive calculations to generate a stroke risk assessment score based on TCM features.

[0051] The Western medicine indicator analysis module, based on three types of clinical data—electronic medical records, family medical history, and past medical history—quantitatively assesses stroke risk scores across various dimensions and uses a weighted fusion algorithm for comprehensive analysis to generate a stroke risk prediction score based on evidence-based medicine.

[0052] The psychological state analysis module is used to identify users' psychological states such as low mood, anxiety, and mood swings, and to calculate stroke risk scores.

[0053] The video behavior analysis module is used to analyze individual behavioral characteristics, such as falls, convulsions, and gait asymmetry, and to calculate stroke risk scores.

[0054] The voiceprint analysis module is used to analyze information such as unclear pronunciation, aphasia, and abnormal tone in speech, and to calculate stroke risk scores;

[0055] The integrated traditional Chinese and Western medicine consultation and analysis module is used to identify subjective discomfort descriptions from users, such as "dizziness" and calculate stroke risk scores.

[0056] The stroke risk comprehensive assessment module is used to assess stroke risk based on six modules: traditional Chinese medicine characteristics, Western medicine indicators, psychological state, video behavior, voiceprint, and traditional Chinese and Western medicine consultation. It comprehensively obtains the overall stroke risk score and realizes risk stratification and early warning.

[0057] Specifically, the TCM feature analysis module includes: filtering feature parameters with stroke warning characteristics based on collected data; and triggering a feature warning mechanism based on facial features with stroke warning characteristics, such as facial color features and facial symmetry features, when facial flushing or asymmetry features are detected, calculating the overall stroke risk score A1 of the facial features, as expressed below: Where, n 1 S represents the number of facial warning features. 1 This represents the set of all facial features that can trigger an alert. Indicates the weight of facial warning features. This indicates whether facial features triggered an alert; 1 indicates triggered, and 0 indicates not triggered.

[0058] Based on visual features that can trigger stroke warnings, such as eye color, a feature-based warning mechanism is triggered when red eye color is detected. The overall stroke risk score A2 of the visual assessment is calculated, as shown in the following expression:

[0059] Where, n 2 S represents the number of eye warning features. 2 This represents the set of all visual diagnostic features that can trigger an alert. Indicates the weight of eye warning features. This indicates whether the eye feature triggers a warning; 1 indicates a trigger and 0 indicates no trigger.

[0060] Based on tongue features that indicate stroke, such as tongue deviation and petechiae, a feature warning mechanism is triggered when petechiae are detected or tongue deviation is detected. The overall stroke risk score A3 of the tongue image is calculated, as shown in the following expression: Where, n 3 S represents the number of tongue appearance warning features. 3 This represents the set of all tongue features that can trigger an alert. Indicates the weight of tongue appearance warning features. This indicates whether the tongue appearance features trigger an alert; 1 indicates triggering and 0 indicates not triggering.

[0061] Based on pulse characteristics that indicate stroke, such as triggering a feature-based warning mechanism when a wiry-rapid or wiry-slippery pulse is detected, the overall stroke risk score A4 of the pulse is calculated, as shown in the following expression: Where, n 4 S represents the number of pulse warning features. 4 This represents the set of all pulse characteristics that can trigger an early warning. Indicates the weight of pulse warning features. This indicates whether the pulse characteristics trigger an alert; 1 indicates triggering, and 0 indicates not triggering.

[0062] Based on the stroke risk scores {A1, A2, A3, A4} from four aspects—facial appearance, eye diagnosis, tongue appearance, and pulse diagnosis—a weighted average is used to obtain the stroke risk score P1 based on the overall characteristics of traditional Chinese medicine. The expression is as follows: in, The weights of stroke risk scores are determined by four components: facial appearance, eye diagnosis, tongue appearance, and pulse diagnosis.

[0063] More specifically, in the stroke risk assessment of the TCM feature analysis module, the weights of each feature are allocated as follows: tongue appearance accounts for 40%, because its color and coating can comprehensively reflect the body's Qi and blood status and pathological evolution; facial appearance accounts for 30%, because changes in facial color and symmetry can directly reflect stroke precursors; pulse appearance accounts for 20%, because a wiry, hard, and forceful pulse suggests liver Yang transforming into wind, and a slippery pulse reflects phlegm and dampness accumulation; and eye diagnosis accounts for 10%, because red eyes mainly indicate internal heat rising, and their correlation with stroke is relatively weak.

[0064] Specifically, the Western medicine indicator analysis module includes: filtering indicator parameters with stroke early warning capabilities based on collected data. Based on electronic medical record indicators with stroke early warning capabilities, such as coagulation function and arachnoid status indicators, when abnormal coagulation function (e.g., elevated INR) or bloody cerebrospinal fluid is detected in subarachnoid hemorrhage, an indicator early warning mechanism is triggered, and the overall stroke risk score B1 contained in the electronic medical record is calculated, as shown in the following expression: Where, m 1 T represents the number of early warning indicators for electronic medical records. 1 This represents the set of all electronic medical record indicators that can trigger alerts. Indicates the weight of the early warning indicators for electronic medical records. This indicates whether an electronic medical record indicator has triggered a warning; 1 indicates a triggered warning, and 0 indicates no triggered warning.

[0065] Based on the family history indicator with stroke early warning, the indicator early warning mechanism is triggered when there is a stroke or cerebral thrombosis in the family history. The overall stroke risk score B2 of the family history is calculated, as shown in the following expression: Where, m 2 Indicates the number of warning indicators related to family medical history, T 2 This represents the set of all family medical history indicators that can trigger an alert. Indicates the weight of the family medical history warning indicator. This indicates whether the family medical history indicator has triggered a warning; 1 indicates a triggered warning, and 0 indicates no triggered warning.

[0066] Based on the pre-existing medical history indicator with stroke early warning, the indicator early warning mechanism is triggered when coronary heart disease or hypertension are present in the pre-existing medical history. The overall stroke risk score B3 of the pre-existing medical history is calculated, as shown in the following expression: Where, m 3 Indicates the number of warning indicators related to family medical history, T 3 This represents the set of all family medical history indicators that can trigger an alert. Indicates the weight of the family medical history warning indicator. This indicates whether the family medical history indicator has triggered a warning; 1 indicates a triggered warning, and 0 indicates no triggered warning.

[0067] Based on the stroke risk scores {B1, B2, B3} from electronic medical records, family medical history, and past medical history, a weighted average was used to obtain the overall stroke risk score P2 from Western medicine, as expressed below: in, The weights of stroke risk scores are determined by the three components: electronic medical records, family medical history, and past medical history.

[0068] More specifically, in the stroke risk assessment of the Western medicine indicator analysis module, the weights of each feature are allocated as follows: electronic medical records account for 50% of the weight because they contain real-time, objective clinical indicators and examination results, which can accurately reflect the patient's current physiological state; family medical history accounts for 30% of the weight, mainly considering the significant impact of genetic factors on stroke; and past medical history accounts for 20% of the weight, serving as a supplementary reference for historical data.

[0069] Specifically, the psychological state analysis module includes: based on the collected psychological state data, filtering out state parameters with stroke warning characteristics, such as whether the mood is depressed or whether the mood changes rapidly; when depressed mood or rapid mood changes are present in the video, an early warning mechanism is triggered; and the stroke risk score P3 of the overall psychological state is calculated, as shown in the following expression: Where k represents the number of psychological warning states, β represents the set of all psychological states that can trigger a warning, and W K K represents the weight of the psychological warning state. i This indicates whether a psychological state triggers a warning; 1 indicates a trigger and 0 indicates no trigger.

[0070] Specifically, the video behavior analysis module includes: filtering out behavioral parameters with stroke warning based on video behavior captured by a camera, such as whether the gait is symmetrical or whether a fall occurs; triggering an early warning mechanism when there is gait asymmetry or sudden fall behavior in the video; and calculating the overall behavioral stroke risk score P4 contained in the video, as shown in the following expression: Where z represents the number of warning behaviors in the video, R represents the set of all video behaviors that can trigger warnings, and W R Z represents the weight of the video alert behavior. i This indicates whether the behavior in the video triggered a warning; 1 indicates a warning and 0 indicates no warning.

[0071] Specifically, the voiceprint analysis module includes: based on the collected voiceprints, filtering out voiceprint features with stroke warning characteristics, such as whether there is slurred speech, difficulty in understanding, aphasia, or unclear pronunciation; triggering an early warning mechanism when there is slurred speech, difficulty in understanding, aphasia, or unclear pronunciation or substitute sounds in the video; and calculating the overall voiceprint stroke risk score P5 contained in the speech, as shown in the following expression: Where q represents the number of warning voiceprints in the speech, δ represents the set of all voiceprints that can trigger warnings, and W δ Indicates the voiceprint weight for voice warnings, Q i This indicates whether the voiceprint triggers an alert; 1 indicates triggering and 0 indicates not triggering.

[0072] Specifically, the TCM and Western medicine consultation analysis module includes: based on the collected TCM and Western medicine consultation data, screening out physiological characteristics with stroke warning, for example, triggering an early warning mechanism when dizziness and slurred speech are present in the consultation results, and calculating the stroke risk score P6 of the consultation results, as shown in the following expression: Where d represents the number of warning physiological features in TCM and Western medicine consultations, γ represents the set of all physiological features that can trigger warnings, and W γ D represents the weight of early warning physiological characteristics in traditional Chinese medicine and Western medicine consultations. i This indicates whether physiological characteristics during consultations between traditional Chinese medicine and Western medicine trigger an alert; 1 indicates triggering and 0 indicates not triggering.

[0073] Specifically, the stroke risk comprehensive assessment module includes: a weighted average of stroke risk scores {P1, P2, P3, P4, P5, P6} from six modules: Traditional Chinese Medicine characteristic analysis module, Western medicine indicator analysis module, psychological state analysis module, video behavior analysis module, voiceprint analysis module, and Traditional Chinese and Western medicine consultation analysis module, to obtain the overall stroke risk score SS, as expressed below: SS=W1×P1+W2×P2+W3×P3+W4×P4+W5×P5+W6×P6 Among them, W i The stroke risk scores for the six modules are weighted and combined to obtain the total stroke risk score SS.

[0074] More specifically, in the stroke risk comprehensive assessment module, the weights of each feature are allocated as follows: the Western medicine module accounts for 30%, as it directly reflects pathological changes through CT / MRI imaging and blood biochemical indicators, and has clear diagnostic value; the traditional Chinese medicine module accounts for 20%, which can effectively identify stroke precursor symptoms through a comprehensive assessment of tongue appearance, facial appearance, pulse appearance, and eye examination; the video module accounts for 20%, which intuitively reflects neurological abnormalities by monitoring motor dysfunction in real time; the voiceprint module accounts for 15%, which focuses on capturing language dysfunction, a typical stroke symptom; the psychological module accounts for 10%, which assesses the impact of emotional changes such as depression and anxiety on stroke risk; and the medical history module accounts for 5%, serving as a basic screening method.

[0075] The stroke risk grading system based on the total score (SS) sets the warning thresholds as follows, referencing clinical prediction models such as CHA2DS2-VASc and the sensitivity requirements of the FAST principle: SS < 0.45 is considered low risk (normal); 0.45 ≤ SS < 0.60 is considered medium risk (warning triggered, observation and further testing recommended); and SS ≥ 0.60 is considered high risk (warning triggered and medical intervention recommended). This grading strategy, while ensuring screening sensitivity, adheres to the principle of "better to have a false positive than a missed diagnosis" in acute stroke identification.

[0076] This invention also provides a stroke early warning system that integrates traditional Chinese medicine characteristics, Western medicine indicators, psychological state, video behavior, voiceprint, and information from both traditional Chinese and Western medicine consultations. This stroke early warning system can be implemented by executing the steps of the stroke early warning method that integrates these characteristics. In other words, those skilled in the art can understand the stroke early warning method that integrates these characteristics as a preferred embodiment of the stroke early warning system.

[0077] Those skilled in the art will understand that, besides implementing the system and its various devices, modules, and units provided by this invention in the form of purely computer-readable program code, the same functions can be achieved entirely through logical programming of the method steps, making the system and its various devices, modules, and units of this invention function in the form of logic gates, switches, application-specific integrated circuits, programmable logic controllers, and embedded microcontrollers. Therefore, the system and its various devices, modules, and units provided by this invention can be considered as a hardware component, and the devices, modules, and units included therein for implementing various functions can also be considered as structures within the hardware component; alternatively, the devices, modules, and units for implementing various functions can be considered as both software modules implementing the method and structures within the hardware component.

[0078] Specific embodiments of the present invention have been described above. It should be understood that the present invention is not limited to the specific embodiments described above, and those skilled in the art can make various changes or modifications within the scope of the claims, which do not affect the essence of the present invention. Unless otherwise specified, the embodiments and features described in this application can be arbitrarily combined with each other.

Claims

1. A stroke early warning system integrating traditional Chinese medicine characteristics, Western medicine indicators, psychological state, video behavior, voiceprint, and information from both traditional Chinese and Western medicine consultations, characterized in that, include: The TCM feature analysis module, based on four types of features—facial features, eye diagnosis, tongue appearance, and pulse appearance—quantifies the stroke risk scores of each part and uses a weighted fusion algorithm to perform comprehensive calculations to generate a stroke risk assessment score based on TCM features. The Western medicine indicator analysis module, based on three types of clinical data—electronic medical records, family medical history, and past medical history—quantitatively assesses stroke risk scores across various dimensions and uses a weighted fusion algorithm for comprehensive analysis to generate a stroke risk prediction score based on evidence-based medicine. The psychological state analysis module is used to identify users' psychological states such as low mood, anxiety, and mood swings, and to calculate stroke risk scores. The video behavior analysis module is used to analyze individual behavioral characteristics, such as falls, convulsions, and gait asymmetry, and to calculate stroke risk scores. The voiceprint analysis module is used to analyze information such as unclear pronunciation, aphasia, and abnormal tone in speech, and to calculate stroke risk scores; The integrated traditional Chinese and Western medicine consultation and analysis module is used to identify subjective discomfort descriptions from users, such as "dizziness" and calculate stroke risk scores. The stroke risk comprehensive assessment module is used to assess stroke risk based on six modules: traditional Chinese medicine characteristics, Western medicine indicators, psychological state, video behavior, voiceprint, and traditional Chinese and Western medicine consultation. It comprehensively obtains the overall stroke risk score and realizes risk stratification and early warning.

2. The stroke early warning system according to claim 1, which integrates traditional Chinese medicine characteristics, Western medicine indicators, psychological state, video behavior, voiceprint, and information from both traditional Chinese and Western medicine consultations, is characterized in that... The TCM feature analysis module includes: Based on facial image data captured by a camera, feature parameters with stroke warning characteristics are selected, such as facial color features and facial symmetry features. When facial flushing or asymmetry features are detected, a feature warning mechanism is triggered, and the overall stroke risk score A1 of the facial image is calculated, as shown in the following expression: Where, n 1 S represents the number of facial warning features. 1 This represents the set of all facial features that can trigger an alert. Indicates the weight of facial warning features. This indicates whether facial features triggered an alert; 1 indicates triggered, and 0 indicates not triggered. Based on eye image data captured by a camera, feature parameters with stroke warning characteristics are selected, such as eye color. When red eye color is detected, a feature warning mechanism is triggered, and the overall stroke risk score A2 of the eye examination is calculated, as shown in the following expression: Where, n 2 S represents the number of eye warning features. 2 This represents the set of all visual diagnostic features that can trigger an alert. Indicates the weight of eye warning features. This indicates whether the eye feature triggers a warning; 1 indicates triggering and 0 indicates not triggering. Based on tongue image data captured by a camera, feature parameters with stroke warning characteristics are selected, such as tongue deviation and tongue petechiae. When the tongue petechiae feature is detected as present or the tongue deviation feature is detected as deviation, the feature warning mechanism is triggered, and the overall stroke risk score A3 of the tongue image is calculated, as shown in the following expression: Where, n 3 S represents the number of tongue appearance warning features. 3 This represents the set of all tongue features that can trigger an alert. Indicates the weight of tongue appearance warning features. This indicates whether the tongue appearance features trigger an alert; 1 indicates triggering and 0 indicates not triggering. Based on pulse data collected by wearable devices, characteristic parameters with stroke warning features are selected. For example, when a wiry pulse or a wiry and slippery pulse is detected, a feature warning mechanism is triggered, and the overall stroke risk score A4 of the pulse is calculated, as shown in the following expression: Where, n 4 S represents the number of pulse warning features. 4 This represents the set of all pulse characteristics that can trigger an early warning. Indicates the weight of pulse diagnosis warning features. This indicates whether the pulse characteristics trigger an alert; 1 indicates triggering, and 0 indicates not triggering. Based on the stroke risk scores {A1, A2, A3, A4} from four aspects—facial appearance, eye diagnosis, tongue appearance, and pulse diagnosis—a weighted average is used to obtain the stroke risk score P1 based on the overall characteristics of traditional Chinese medicine. The expression is as follows: in, The weights of stroke risk scores are determined by four components: facial appearance, eye diagnosis, tongue appearance, and pulse diagnosis.

3. The stroke early warning system according to claim 1, which integrates traditional Chinese medicine characteristics, Western medicine indicators, psychological state, video behavior, voiceprint, and information from both traditional Chinese and Western medicine consultations, is characterized in that... The Western medicine indicator analysis module includes: Based on the collected electronic medical records, indicators with stroke early warning capabilities are selected, such as coagulation function and arachnoid status indicators. When abnormal coagulation function (e.g., elevated INR) or bloody cerebrospinal fluid is detected in subarachnoid hemorrhage, the indicator early warning mechanism is triggered. The overall stroke risk score B1 contained in the electronic medical records is calculated, as shown in the following expression: Where, m 1 T represents the number of early warning indicators for electronic medical records. 1 This represents the set of all electronic medical record indicators that can trigger alerts. Indicates the weight of the early warning indicators for electronic medical records. This indicates whether an electronic medical record indicator has triggered a warning; 1 indicates a triggered warning, and 0 indicates no triggered warning. Based on collected family medical history, indicators with stroke early warning capabilities are selected. For example, when there is a family history of stroke or cerebral thrombosis, an indicator early warning mechanism is triggered, and the overall stroke risk score B2 of the family medical history is calculated, as shown in the following expression: Where, m 2 Indicates the number of warning indicators related to family medical history, T 2 This represents the set of all family medical history indicators that can trigger an alert. Indicates the weight of the family medical history warning indicator. This indicates whether the family medical history indicator has triggered a warning; 1 indicates a triggered warning, and 0 indicates no triggered warning. Based on the collected medical history, indicators with stroke early warning capabilities are selected. For example, when the medical history includes coronary heart disease or hypertension, the indicator early warning mechanism is triggered, and the overall stroke risk score B3 of the medical history is calculated, as shown in the following expression: Where, m 3 Indicates the number of warning indicators related to family medical history, T 3 This represents the set of all family medical history indicators that can trigger an alert. Indicates the weight of the family medical history warning indicator. This indicates whether the family medical history indicator has triggered a warning; 1 indicates a triggered warning, and 0 indicates no triggered warning. Based on the stroke risk scores {B1, B2, B3} from electronic medical records, family medical history, and past medical history, a weighted average was used to obtain the overall stroke risk score P2 from Western medicine, as expressed below: in, The weights of stroke risk scores are determined by the three components: electronic medical records, family medical history, and past medical history.

4. The stroke early warning system according to claim 1, which integrates traditional Chinese medicine characteristics, Western medicine indicators, psychological state, video behavior, voiceprint, and information from both traditional Chinese and Western medicine consultations, is characterized in that... The psychological state analysis module includes: Based on the collected psychological state data, state parameters with stroke warning characteristics are selected, such as whether the mood is depressed or whether the mood changes rapidly. When a video shows a depressed mood or a rapid change in mood, the warning mechanism is triggered, and the stroke risk score P3 of the overall psychological state is calculated, as shown in the following expression: Where k represents the number of psychological warning states, β represents the set of all psychological states that can trigger a warning, and W K K represents the weight of the psychological warning state. i This indicates whether a psychological state triggers a warning; 1 indicates a trigger and 0 indicates no trigger.

5. The stroke early warning system according to claim 1, which integrates traditional Chinese medicine characteristics, Western medicine indicators, psychological state, video behavior, voiceprint, and information from both traditional Chinese and Western medicine consultations, is characterized in that... The video behavior analysis module includes: Based on video behavior captured by cameras, behavioral parameters with stroke warning characteristics are filtered out, such as whether the gait is symmetrical and whether a fall occurs. When gait asymmetry or sudden fall is observed in the video, the warning mechanism is triggered. The overall stroke risk score P4 of the behaviors contained in the video is calculated, as shown in the following expression: Where z represents the number of warning behaviors in the video, R represents the set of all video behaviors that can trigger warnings, and W R Z represents the weight of the video alert behavior. i This indicates whether the behavior in the video triggered a warning; 1 indicates a warning and 0 indicates no warning.

6. The stroke early warning system according to claim 1, which integrates traditional Chinese medicine characteristics, Western medicine indicators, psychological state, video behavior, voiceprint, and information from both traditional Chinese and Western medicine consultations, is characterized in that... The voiceprint analysis module includes: Based on the collected voiceprints, voiceprint features with stroke warning characteristics are selected, such as whether there is slurred speech, difficulty in understanding, aphasia, or unclear pronunciation. When slurred speech, difficulty in understanding, aphasia, or unclear pronunciation or substitute sounds are present in the video, the warning mechanism is triggered. The overall voiceprint stroke risk score P5 contained in the speech is calculated, as shown in the following expression: Where q represents the number of warning voiceprints in the speech, δ represents the set of all voiceprints that can trigger warnings, and W δ Indicates the voiceprint weight for voice warnings, Q i This indicates whether the voiceprint triggers an alert; 1 indicates triggering and 0 indicates not triggering.

7. The stroke early warning system according to claim 1, which integrates traditional Chinese medicine characteristics, Western medicine indicators, psychological state, video behavior, voiceprint, and information from both traditional Chinese and Western medicine consultations, is characterized in that... The integrated traditional Chinese and Western medicine consultation and analysis module includes: Based on the collected data from both traditional Chinese medicine and Western medicine consultations, physiological characteristics that could indicate stroke were identified. For example, when dizziness or slurred speech was present in the consultation results, an early warning mechanism was triggered. The stroke risk score P6 of the consultation results was calculated, as shown in the following expression: Where d represents the number of warning physiological features in TCM and Western medicine consultations, γ represents the set of all physiological features that can trigger warnings, and W γ D represents the weight of early warning physiological characteristics in traditional Chinese medicine and Western medicine consultations. i This indicates whether physiological characteristics during consultations between traditional Chinese medicine and Western medicine trigger an alert; 1 indicates triggering and 0 indicates not triggering.

8. The stroke early warning system according to claim 1, which integrates traditional Chinese medicine characteristics, Western medicine indicators, psychological state, video behavior, voiceprint, and information from both traditional Chinese and Western medicine consultations, is characterized in that... The comprehensive stroke risk assessment module includes: Based on the stroke risk scores {P1, P2, P3, P4, P5, P6} from six modules—Traditional Chinese Medicine (TCM) feature analysis, Western Medicine (WC) indicator analysis, psychological state analysis, video behavior analysis, voiceprint analysis, and TCM-Western Medicine consultation analysis—a weighted average is used to obtain the overall stroke risk score SS, expressed as follows: SS=W1×P1+W2×P2+W3×P3+W4×P4+W5×P5+W6×P6 Among them, W i The stroke risk scores of the six modules are weighted and combined to obtain the total stroke risk score SS. Based on the total score SS, the stroke risk grading system, referring to the sensitivity requirements of clinical prediction models such as CHA2DS2-VASc and the FAST principle, sets the warning thresholds as follows: SS < 0.45 is considered low risk (normal), 0.45 ≤ SS < 0.60 is considered medium risk (triggered warning, observation and further testing recommended), and SS ≥ 0.60 is considered high risk (triggered warning and medical intervention recommended). This grading strategy, while ensuring screening sensitivity, follows the principle of "better to have a false positive than a missed diagnosis" in acute stroke identification.

Images