Multi-mode and time sequence combined old people ability dynamic assessment method

Abstract

The invention provides a multi-mode and time sequence combined old people ability dynamic evaluation method, which comprises the following steps of: S1, data acquisition: aiming at any old people user wearing intelligent equipment, acquiring question and answer interaction data of the old people user and sensor data output by the intelligent equipment; s2, data pre-processing: performing data pre-processing on the question and answer interaction data and the sensor data to obtain pre-processed question and answer interaction data and pre-processed sensor data; s3, multi-modal data fusion: based on a deep learning algorithm, carrying out data fusion on the preprocessed question and answer interaction data and the preprocessed sensor data, and extracting key information; and S4, dynamic evaluation: based on a hierarchical analysis algorithm, carrying out dynamic weighting calculation on the key information to obtain ability scores of the elderly users in different dimensions, and generating an ability dynamic evaluation report according to each ability score. The method has the beneficial effects that the evaluation cost can be reduced, and the evaluation objectivity and accuracy can be improved.

Description

Technical Field

[0001] The present invention relates to the technical field of the ability assessment of the elderly. Specifically, it relates to a dynamic assessment method for the ability of the elderly combining multi-modal and time series. Background Art

[0002] With the increasingly serious trend of global population aging, the health assessment of the elderly has become an important social issue. The traditional health assessment methods for the elderly mainly rely on the experience and subjective judgment of doctors. This method is not only time-consuming and laborious, but also easily affected by the personal experience and subjective factors of doctors, resulting in the inaccuracy and incompleteness of the assessment results. With the development of data fusion technology and artificial intelligence technology, using multi-modal data for the health assessment of the elderly has become a new research direction.

[0003] The existing health assessment technologies for the elderly mainly include the following aspects: one is the data assessment method based on a single modality, such as only using image data or text data for assessment; the second is the static assessment method, that is, the assessment result at one time can only reflect the health status of the elderly at a certain time point, lacking the dynamic reflection of the change of the elderly's ability; the third is the assessment method relying on professional doctors, with high assessment costs and being easily affected by the personal experience and subjective factors of doctors.

[0004] Although the existing health assessment technologies for the elderly have solved the problems of the health assessment of the elderly to a certain extent, there are still some problems and limitations. First of all, the existing data assessment methods based on a single modality cannot make full use of the advantages brought by multi-modal data, resulting in the incompleteness and inaccuracy of the assessment results; secondly, the existing static assessment methods cannot timely reflect the change of the elderly's ability and cannot meet the health management needs of the elderly; thirdly, the assessment method relying on professional doctors has high assessment costs and is easily affected by the personal experience and subjective factors of doctors, restricting its popularization and application; therefore, how to use multi-modal data for dynamic health assessment of the elderly, reduce the assessment cost, and improve the objectivity and accuracy of the assessment is an urgent problem to be solved in the current field of the health assessment of the elderly. Summary of the Invention

[0005] The technical problem to be solved by the present invention is to reduce the assessment cost and improve the objectivity and accuracy of the assessment. To overcome the defects of the above prior art (or related art), the present invention provides a dynamic assessment method for the ability of the elderly combining multi-modal and time series.

[0006] The present invention provides a dynamic assessment method for the ability of the elderly combining multi-modal and time series, including: Step S1, Data Acquisition: For any elderly user wearing a smart device, obtain the Q&A interaction data of the elderly user and the sensor data output by the smart device; Step S2, Data Preprocessing: Perform data preprocessing on the Q&A interaction data and the sensor data respectively to obtain the corresponding preprocessed Q&A interaction data and preprocessed sensor data; Step S3, Multimodal Data Fusion: Based on deep learning algorithms, fuse the preprocessed Q&A interaction data and the preprocessed sensor data and extract key information; Step S4, Dynamic Evaluation: Based on hierarchical analysis algorithms, perform dynamic weighted calculations on the key information to obtain the ability scores of the elderly user in different dimensions, and generate corresponding ability dynamic evaluation reports according to each ability score.

[0007] Compared with the prior art, a method for dynamically evaluating the abilities of the elderly combining multimodality and time series in this application has the following advantages: In this application, by integrating Q&A interaction data, sensor data and applying artificial intelligence algorithms, multimodal data fusion is achieved, providing more comprehensive and accurate evaluation results. By adopting a dynamic ability evaluation service, it can reflect the ability changes of the elderly in real time, meet the health management needs of the elderly. Through an automated evaluation process, the dependence on professionals is reduced, thus reducing the evaluation cost and improving its popularization and practicability. By introducing sensor data with strong objectivity and artificial intelligence algorithms for data analysis, the influence of subjective factors on the ability evaluation results of the elderly is reduced, and the objectivity and accuracy of the evaluation are improved.

[0008] In a possible implementation manner, in the step S1, the Q&A interaction data is collected by conducting a voice Q&A test or a touch interaction questionnaire on the elderly user through the smart device.

[0009] In a possible implementation manner, in the step S1, the heart rate data, blood pressure data, blood oxygen saturation data, gait analysis data, resting energy consumption data, activity energy consumption data, behavior data, emotional state data, and cognitive state data of the elderly user are collected as the sensor data.

[0010] In a possible implementation, the preprocessing method adopted in step S2 includes missing data completion processing, abnormal data elimination processing, and normalization processing. The missing data completion processing is to fill the missing parts of the Q&A interaction data and the sensor data by using the interpolation method or the average value method. The abnormal data elimination processing is to filter out the abnormal data in the Q&A interaction data and the sensor data by setting a threshold. The normalization processing is to normalize the Q&A interaction data and the sensor data to the range of [0, 1].

[0011] In a possible implementation, in step S3, feature extraction is performed on the preprocessed Q&A interaction data and the preprocessed sensor data to obtain physiological index features including mean heart rate, blood pressure difference, and gait stability, behavioral features including mean activity level, daily average number of steps, and posture change frequency, and cognitive emotion features including voice emotion score, positive proportion of facial expression recognition, and Q&A response duration. The physiological index features, the behavioral features, and the cognitive emotion features are included as feature values in the key information.

[0012] In a possible implementation, step S4 includes: Step S41, single ability scoring: Perform single-dimensional scoring on each of the feature values in the key information according to three different ability dimensions of living ability, motor ability, and cognitive ability to obtain corresponding single ability scores; Step S42, comprehensive ability scoring: Introduce a weight factor and perform dynamic weighted calculation on each of the single ability scores to generate a comprehensive ability score; Step S43, divide the elderly users into ability levels according to the comprehensive ability score to generate the ability dynamic assessment report.

[0013] In a possible implementation, in step S41, the single ability score is obtained through the following calculation formula: where, represents the single ability score; represents the weight corresponding to the nth feature value; represents the nth feature value.

[0014] In a possible implementation, in step S42, the comprehensive ability score is obtained through the following calculation formula: where, Represents the comprehensive ability score; Represents the m-th single ability score; Represents the weight factor corresponding to the m-th single ability score.

[0015] In a possible implementation manner, in step S42, it further includes dynamically adjusting the weight factors of the respective eigenvalue through the following calculation formula: Wherein, Represents the data acquisition time point corresponding to the eigenvalue; Represents the weight factor corresponding to the eigenvalue; Represents the current time point; Represents the time weight adjustment factor.

[0016] In a possible implementation manner, in step S3, the ability level division includes five categories of levels, namely, category A level representing completely independent living, category B level representing mild dependence and requiring partial assistance, category C level representing moderate dependence and requiring more assistance, category D level representing severe dependence and requiring full care, and category E level representing complete disability and requiring professional care. Brief Description of the Drawings

[0017] Figure 1 Is the step flow chart of the present invention. Detailed Implementation Manner

[0018] First of all, those skilled in the art should understand that these implementation manners are only used to explain the technical principles of the embodiments of the present application, and are not intended to limit the protection scope of the embodiments of the present application. Those skilled in the art can make adjustments according to needs to adapt to specific application scenarios.

[0019] The following further describes the present application in detail with reference to the drawings and specific embodiments.

[0020] See Figure 1 , The embodiments of the present application disclose a dynamic assessment method for the abilities of the elderly combining multi-modal and time series, including: Step S1, data acquisition: For any elderly user wearing a smart device, obtain the Q&A interaction data of the elderly user and the sensor data output by the smart device; Step S2, data preprocessing: Perform data preprocessing on the Q&A interaction data and the sensor data respectively to obtain the corresponding preprocessed Q&A interaction data and preprocessed sensor data; Step S3, multimodal data fusion: Based on deep learning algorithms, fuse the preprocessed Q&A interaction data and the preprocessed sensor data and extract key information; Step S4, dynamic assessment: Based on hierarchical analysis algorithms, perform dynamic weighted calculations on the key information to obtain the ability scores of the elderly users in different dimensions, and generate corresponding ability dynamic assessment reports according to each ability score.

[0021] In the embodiments of the present application, hierarchical analysis algorithms are adopted to collect multi-dimensional data including physiological data, behavioral data, and psychological data. Among them, physiological data includes heart rate, blood pressure, blood oxygen saturation (collected through wearable devices or medical sensors), gait analysis (stride length, walking speed, gait variation coefficient), resting energy consumption and activity energy consumption (basal metabolic rate and activity consumption); behavioral data includes self-care ability and basic motor ability; psychological data includes emotional state: evaluate the emotional stability and depression risk of the elderly through voice interaction or facial expression recognition, and perception and social participation: obtain cognitive state scores through simple Q&A (such as memory and attention tests); subsequently, perform dynamic weighted calculations on the above data to generate an ability dynamic assessment report (the provided results include ability assessment levels, sub-item ability scores, safety hazard warnings, and personalized suggestions).

[0022] The method of the present application includes a hardware layer and a software layer. Among them, the intelligent devices in the hardware layer include sensors mainly composed of wearable devices (such as smart bracelets) and environmental sensors (such as gait recognition cameras), and electronic devices mainly composed of mobile phones, tablets, or specific intelligent terminals; the software layer includes a data collection module, an evaluation algorithm module, and a data storage and sharing module. The data collection module accesses the sensors through API interfaces to obtain Q&A interaction data. The evaluation algorithm module dynamically calculates ability scores and generates ability dynamic assessment reports (the calculation logic of ability indicators refers to the addition of national standard specification documents). The data storage and sharing module encrypts and stores data in the cloud and supports multi-device synchronization; the working principle during specific operations is as follows: users complete voice or touch interaction questionnaires through intelligent devices, and at the same time collect physiological and behavioral data (physiological and behavioral data refers to various objective data collected through sensors and other devices that can reflect the physical state and behavioral characteristics of the elderly). Subsequently, the data is uploaded to the cloud, and the evaluation algorithm performs analysis and processing. Finally, the system generates an ability dynamic assessment report and pushes relevant service suggestions.

[0023] The specific implementation steps of the method of the present application include: 1. User initialization: Log in to the device through identity authentication and select ability assessment items; 2. Multimodal information collection: Complete voice Q&A and wear sensors for data collection; 3. Evaluation and analysis: The system performs hierarchical weighted analysis on the collected data according to the algorithm; 3.1 Data Preprocessing Before performing weighted analysis, the system cleans and normalizes the collected multimodal data to ensure data quality and consistency; Completing Missing Data: Use interpolation or the mean method to fill in the missing parts of the data; Removing Abnormal Data: Filter out obviously abnormal data (such as extreme values of heart rate data) by setting thresholds; Normalization Processing: Normalize data with different units and dimensions (such as heart rate and number of steps) to the range of [0, 1] for unified calculation; 3.2 Data Stratification Processing According to the characteristics of multimodal data, the dynamic weighted algorithm divides the data into different dimensions for stratified analysis, mainly including the following three levels: - Basic Layer Analysis (Extracting Original Data Features) Physiological Index Features: Extract key features such as mean heart rate, blood pressure difference, and gait stability; Behavioral Features: Extract behavioral data such as mean activity level, daily average number of steps, and frequency of posture changes; Cognitive and Emotional Features: Extract psychological features such as speech emotion scores, positive proportion of facial expression recognition, and question-and-answer response duration; - Intermediate Layer Analysis (Single Ability Scoring) According to different ability dimensions (such as living ability, motor ability, and cognitive ability), perform single-dimensional scoring on the stratified features, and use linear regression or decision tree algorithms to map the features to scoring values; - High-level Analysis (Comprehensive Ability Scoring) Perform dynamic weighted calculation on the scores of each dimension to generate an ability dynamic assessment report, and the weight factor is adjusted in real time according to individual circumstances (for example, older elderly people attach more importance to living ability, while younger elderly people are more concerned about motor ability); 3.3 Dynamic Weighting Mechanism Dynamic weighting is the core of this algorithm. The weight factor is adjusted in combination with the specific characteristics of the elderly and the real-time state of the data. The specific adjustment strategies include: I. Adjustment in the Time Dimension Assign higher weights to time-sensitive indicators (such as gait and heart rate), and give priority to recent data; II. Individualized Adjustment Dynamically adjust the weights according to the user's age, health status, and historical assessment records. For example: For older elderly people with mobility difficulties, reduce the weight of motor ability and increase the weight of living ability; 3.4 Generation of Comprehensive Report The system divides users into the following levels according to the final comprehensive ability score and generates an ability dynamic assessment report; Ability level classification: Category A: Completely independent in daily life; Category B: Slightly dependent and requires partial assistance; Category C: Moderately dependent and requires more assistance; Category D: Severely dependent and requires comprehensive care; Category E: Completely disabled and requires professional care; Report content: Detailed ability scores and dimensional analysis, health improvement suggestions, personalized service recommendations (such as suitable community activities or rehabilitation courses); 4. Report generation and recommendation Generate a detailed dynamic ability assessment report and push relevant health management services; 5. Dynamic update Regularly remind users to conduct re - evaluations and provide trend analysis.

[0024] The following is an example of the operation process for an elderly user, including: User login: An elderly person logs in to the assessment system through a mobile phone and enters basic information; Q&A module: Answer questions about daily living ability (such as dressing independently, taking a bath) through voice interaction; Sensor collection: Wear a smart bracelet to record real - time gait, heart rate and other data; Report generation: The system combines subjective Q&A and objective sensor data to generate a dynamic ability assessment report, including ability level (same as the above ability classification), health suggestions, etc.; Service recommendation: After the dynamic ability assessment report is generated, recommend community meal assistance services and rehabilitation course links.

[0025] In summary, it can be seen that the technical solution of this application solves: 1. The problem of multi - modal data fusion: Existing single - modal data assessment methods cannot fully utilize the advantages brought by multi - modal data, resulting in the incompleteness and inaccuracy of assessment results. However, the technical solution of this application realizes multi - modal data fusion by integrating Q&A interaction data, sensor data collection and artificial intelligence algorithms, providing more comprehensive and accurate assessment results; 2. The problem of dynamic assessment: Existing static assessment methods cannot timely reflect the ability changes of the elderly and cannot meet the health management needs of the elderly. However, the technical solution of this application provides a dynamic ability assessment service that can reflect the ability changes of the elderly in real - time and meet the health management needs of the elderly; 3. Cost reduction issue: The evaluation method relying on professional doctors is costly and vulnerable to the influence of doctors' personal experience and subjective factors, which restricts its promotion and application. However, the technical solution of this application reduces the dependence on professionals through an automated evaluation process, thereby reducing the evaluation cost and improving its promotability and practicability. 4. Influence of subjective factors issue: The existing elderly health evaluation technology is easily affected by doctors' personal experience and subjective factors, resulting in inaccurate evaluation results. However, the technical solution of this application reduces the influence of subjective factors on the evaluation results of the elderly's abilities by introducing sensor data with strong objectivity and artificial intelligence algorithms for data analysis, improving the objectivity and accuracy of the evaluation. 5. Single evaluation dimension issue: The existing elderly health evaluation technology can only reflect some of the abilities of the elderly and lacks a comprehensive reflection of the elderly's ability status. However, the technical solution of this application establishes a comprehensive elderly ability evaluation system including multiple dimensions such as physical function, cognitive function, and social participation to comprehensively reflect the elderly's ability status and solve the problem of single evaluation dimension in the existing technology.

[0026] The technical solutions adopted in this application mainly include: 1. Multimodal data fusion technology: By integrating question-and-answer interaction data, sensor data collection, and artificial intelligence algorithms, multimodal data fusion is achieved to provide more comprehensive and accurate evaluation results. Specifically, the question-and-answer interaction module collects the basic information and daily behavior data of the elderly through natural language processing technology; the sensor data collection module collects the physiological and behavioral data of the elderly through various sensors such as gait sensors and sphygmomanometers; the artificial intelligence algorithm module uses technologies such as deep learning and time series analysis to analyze and process the collected data, extract key information, and provide accurate evaluation results. 2. Dynamic evaluation technology: Provide a dynamic ability evaluation service that can reflect the ability changes of the elderly in real time. Specifically, the system regularly collects the data of the elderly and conducts real-time analysis and processing through artificial intelligence algorithms to generate a dynamic ability evaluation report reflecting the ability change trend of the elderly. 3. Automated evaluation process technology: By automating the evaluation process, the dependence on professionals is reduced, thereby reducing the evaluation cost. Specifically, the system automatically collects, analyzes, and processes data to generate a dynamic ability evaluation report without the need for professional doctors to conduct manual evaluations, greatly reducing the evaluation cost. 4. Sensor data technology with strong objectivity: The introduction of sensor data with strong objectivity and artificial intelligence algorithms for data analysis reduces the impact of subjective factors on the evaluation results of the elderly's abilities and improves the objectivity and accuracy of the evaluation. Specifically, the sensor data acquisition module collects the elderly's physiological data and behavioral data such as gait, blood pressure, heart rate, etc. These data are objective and not easily affected by subjective factors; the artificial intelligence algorithm module uses deep learning, time series analysis and other technologies to analyze and process these data, extract key information, and provide accurate evaluation results; 5. Multi-dimensional comprehensive assessment technology: Establish a comprehensive elderly ability assessment system including physical function, cognitive function, social participation and other dimensions to comprehensively reflect the ability status of the elderly and solve the problem of single assessment dimension in existing technologies. Specifically, the system will collect data on the elderly from multiple dimensions, such as physical function data, cognitive function data, social participation data, etc., and conduct comprehensive analysis and processing through artificial intelligence algorithms to generate a multi-dimensional dynamic ability assessment report to comprehensively reflect the ability status of the elderly.

[0027] Due to the advanced nature of the technical solution of this application, it can be widely used in application fields such as elderly health management, medical services, and smart health equipment; first, in the field of elderly health management, the dynamic, multimodal, and intelligent elderly ability assessment method provided by the technical solution of this application can reflect the health status and ability changes of the elderly in real time, which is helpful to timely discover the health problems of the elderly, formulate scientific health management plans, and improve the health level and quality of life of the elderly.

[0028] In addition, since the evaluation method of the technical solution of the present application is simple, easy to implement, low-cost, and easy to promote and apply, it has broad application prospects in places where the elderly gather, such as communities and nursing homes.

[0029] Secondly, in the field of medical services, the technical solution of this application can assist doctors in conducting health assessments of the elderly, improve the objectivity and accuracy of the assessments, reduce the workload of doctors, and improve the efficiency and quality of medical services.

[0030] At the same time, since the application of this technical solution can reflect the health status and ability changes of the elderly in real time, it can provide doctors with more comprehensive clinical information, which is helpful for the early diagnosis and treatment of diseases.

[0031] Finally, in the field of smart health devices, the technical solution of this application can be integrated into various smart health devices, such as smart bracelets, smart home devices, etc., to achieve continuous health monitoring and management of the elderly, which will greatly improve the convenience and safety of the lives of the elderly, and also provide new technical support for the development of smart health devices.

[0032] Generally speaking, with the increasingly serious trend of global population aging, the demand for the health assessment of the elderly is growing. Due to the advantages of the technical solution of this application in aspects such as multi-modal data fusion, dynamic assessment, cost reduction, user experience, and intelligence level, it will have broad application prospects and market demands in the fields of elderly health management, medical services, and intelligent health devices.

[0033] In the description of this application, the description with reference to terms such as "one embodiment", "some embodiments", "in this embodiment", "specific examples", or "some examples" means that the specific features, mechanisms, materials, or characteristics described in connection with the embodiment or example are included in at least one embodiment or example of this application. In this specification, the schematic expressions of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, mechanisms, materials, or characteristics described can be combined in a suitable manner in any one or more embodiments or examples. In addition, without contradiction, those skilled in the art can combine and combine the different embodiments or examples described in this specification and the features of different embodiments or examples.

[0034] As described above, it is only the specific implementation manner of this application, but the protection scope of this application is not limited thereto. Any changes or substitutions that can be easily thought of by those skilled in the art within the technical scope disclosed in this application should be covered by the protection scope of this application. Therefore, the protection scope of this application should be subject to the protection scope of the claims.

Claims

1. A multi-modal and time series combined dynamic assessment method for the ability of the elderly, characterized by: The following steps are involved: Step S1, data collection: for any elderly user wearing a smart device, obtaining the question-and-answer interaction data of the elderly user and the sensor data output by the smart device; Step S2, data preprocessing: performing data preprocessing on the question-answer interaction data and the sensor data to obtain corresponding preprocessed question-answer interaction data and preprocessed sensor data; Step S3, multimodal data fusion: based on a deep learning algorithm, the pre-processed question-answer interaction data and the pre-processed sensor data are fused and key information is extracted; Step S4, dynamic evaluation: Based on the hierarchical analysis algorithm, the key information is dynamically weighted to obtain the ability scores of the elderly users in different dimensions, and a corresponding ability dynamic evaluation report is generated according to each of the ability scores.

2. The method for dynamic assessment of the ability of the elderly according to claim 1, characterized in that: In the step S1, the question-and-answer interaction data is collected by conducting a voice question-and-answer test or a touch interaction questionnaire on the elderly user through the smart device.

3. The method for dynamic assessment of the ability of the elderly according to claim 1, characterized in that: In the step S1, the heart rate data, blood pressure data, blood oxygen saturation data, gait analysis data, resting energy consumption data, activity energy consumption data, behavior data, emotional state data, and cognitive state data of the elderly user are collected as the sensor data.

4. The method for dynamic assessment of the ability of the elderly according to claim 1, characterized in that: The preprocessing method adopted in the step S2 includes missing data completion processing, abnormal data elimination processing and normalization processing. The missing data completion processing is to use interpolation method or average method to fill the missing data parts of the question and answer interaction data and the sensor data. The abnormal data elimination processing is to filter out the abnormal data in the question and answer interaction data and the sensor data by setting a threshold. The normalization processing is to normalize the question and answer interaction data and the sensor data to the range of [0,1].

5. The method for dynamic assessment of the ability of the elderly according to claim 3 is characterized in that: In the step S3, feature extraction is performed on the preprocessed question-and-answer interaction data and the preprocessed sensor data to obtain physiological indicator features including heart rate mean, blood pressure difference, and gait stability, behavioral features including activity mean, average daily steps, and posture change frequency, and cognitive emotion features including voice emotion score, positive expression recognition ratio, and question-and-answer response time. The physiological indicator features, the behavioral features, and the cognitive emotion features are included in the key information as feature values.

6. The method for dynamic assessment of the ability of the elderly according to claim 5 is characterized in that: The step S4 comprises: Step S41, single ability scoring: performing single dimension scoring on each of the characteristic values ​​in the key information according to the three different ability dimensions of life ability, motor ability, and cognitive ability to obtain a corresponding single ability score; Step S42, comprehensive ability score: introducing a weight factor, and dynamically weighting the scores of the single abilities to generate a comprehensive ability score; Step S43, classifying the elderly users into ability levels according to the comprehensive ability scores to generate the dynamic ability assessment report.

7. The method for dynamic assessment of the ability of the elderly according to claim 6, characterized in that: In step S41, the single ability score is obtained by the following calculation formula: in, represents said single competency score; represents the weight corresponding to the nth eigenvalue; represents the nth eigenvalue.

8. The method for dynamic assessment of the ability of the elderly according to claim 6, characterized in that: In step S42, the comprehensive ability score is obtained by the following calculation formula: in, represents the comprehensive ability score; represents the mth single ability score; Represents the weight factor corresponding to the mth single ability score.

9. The method for dynamic assessment of the ability of the elderly according to claim 6, characterized in that: The step S42 further includes dynamically adjusting the weight factor of each characteristic value by the following calculation formula: in, Indicates the data collection time point corresponding to the characteristic value; represents the weight factor corresponding to the eigenvalue; Indicates the current time point; Represents the time weight adjustment factor.

10. The method for dynamic assessment of the ability of the elderly according to claim 6, characterized in that: In step S3, the ability level classification includes five levels, namely, Class A indicating completely independent living, Class B indicating mild dependence requiring partial help, Class C indicating moderate dependence requiring more help, Class D indicating severe dependence requiring comprehensive care, and Class E indicating complete disability requiring professional care.

Images