Child development dynamic evaluation method and system based on multi-source data fusion

Abstract

The present application belongs to the technical field of children's growth and development evaluation, and particularly relates to a children's development dynamic evaluation method and system based on multi-source data fusion. The present application can preliminarily judge the growth and development state of children by fusing the generated growth and development sample curve, obtain the second index data of children by using a non-invasive acquisition device, adjust the acquisition frequency based on the initial state type, ensure the comprehensiveness and accuracy of the data, and avoid unnecessary frequent acquisition, process the second index data by using a pre-trained convolutional neural network model to generate an individual growth fingerprint, filter out historical fingerprint data with high similarity to the individual growth fingerprint from a cloud database, and construct a personalized growth corridor, so as to provide a personalized reference trajectory range for the growth and development of children, and judge whether the growth and development of children is abnormal based on the same, so as to realize the precise evaluation of the growth and development of children.

Description

Technical Field

[0001] The present invention belongs to the technical field of child growth and development assessment, and specifically relates to a method and system for dynamic child development assessment based on multi-source data fusion. Background Art

[0002] With the development of social economy and the improvement of people's living standards, parents are paying more and more attention to children's growth and development. Traditional methods of evaluating children's growth and development mainly rely on regular height and weight measurements and evaluation against standard growth curves. However, this method has limitations such as long evaluation cycle, single data, and inability to monitor in real time. With the advancement of medical testing technology, the current evaluation of children's development is no longer limited to traditional static indicators such as height and weight, but is developing towards multi-source data fusion and real-time monitoring.

[0003] In the prior art, although there are some child development assessment methods based on multi-source data, there are still some problems. For example, they still focus on monitoring children's static indicators and lack a comprehensive assessment of children's dynamic development process. Accidental deviations may occur in the child's development process. At this time, static assessments are often difficult to accurately reflect the child's true developmental status, resulting in inaccurate assessment results. In addition, due to differences in individual development, the use of standard growth curves for assessment often ignores the differences between individual children, making the assessment results lack of personalization. Based on this, the present invention aims to provide a dynamic assessment method for child development based on multi-source data fusion to solve the above problems. Summary of the Invention

[0004] The purpose of the present invention is to provide a dynamic assessment method and system for child development based on multi-source data fusion, which can dynamically and comprehensively assess the child's growth and development status based on the child's multi-source indicator data, while taking into account individual differences and generating personalized growth and development assessment reports.

[0005] The technical solutions adopted by the present invention are as follows:

[0006] A dynamic assessment method for child development based on multi-source data fusion, including:

[0007] Obtaining the child's first indicator data and matching it with a corresponding standard assessment curve, the first indicator data including the child's age, gender, height, and weight;

[0008] Obtain the ideal assessment curve and combine it with the standard assessment curve to form a growth and development sample curve;

[0009] Based on the growth and development sample curve, output the initial state of growth and development, and record the first indicator data under the initial state;

[0010] Using non-invasive data collection equipment, collect the child's secondary indicator data, and adjust the collection frequency based on the initial state type. The secondary indicator data includes bone contour point cloud data, blood oxygen saturation thermal map, and nighttime limb activity data;

[0011] The second indicator data is input into the pre-trained convolutional neural network model to generate an individual growth fingerprint;

[0012] Historical fingerprint data with a matching similarity greater than a preset value is obtained from the cloud database to generate a personalized growth corridor. When the frequency or duration of the second indicator data collected in real time breaking through the growth corridor boundary exceeds the corresponding threshold, an early warning is triggered and a personalized growth and development assessment report is generated.

[0013] In a preferred embodiment, the step of obtaining the first indicator data of the child and matching the corresponding standard assessment curve includes:

[0014] Classify the first indicator data according to the age groups of the children to obtain multiple classification subsets;

[0015] Calculating the average value of the first indicator data in each of the classification subsets to form average growth data for each age group, and recording the average growth data as an evaluation condition parameter;

[0016] Performing bidirectional offset processing on the evaluation condition parameters to obtain an initial offset boundary;

[0017] Collect the quantity of the first indicator data under the initial offset boundary and record it as the screening condition parameter;

[0018] Obtaining a screening threshold and comparing the screening condition parameter with the screening threshold;

[0019] When the screening condition parameter is greater than the screening threshold, it indicates that the initial offset boundary corresponding to the screening condition parameter does not meet the offset requirement, and the offset operation continues until the screening condition parameter corresponding to the initial offset boundary is equal to the screening threshold, and the corresponding initial offset boundary is recorded as the standard boundary;

[0020] When the screening condition parameter is equal to the screening threshold, it indicates that the initial offset boundary corresponding to the screening condition parameter meets the offset requirement, and the corresponding initial offset boundary is recorded as the standard boundary;

[0021] When the screening condition parameter is less than the screening threshold, it indicates that the initial offset boundary corresponding to the screening condition parameter is over-shifted, and the initial offset boundary is reversed until the initial offset boundary after reverse shifting is greater than or equal to the screening threshold for the first time, and the standard boundary is obtained;

[0022] The standard range for distinguishing children's growth and development is determined based on the standard boundaries, and the standard assessment curve for children is constructed based on the standard range.

[0023] In a preferred embodiment, the step of obtaining an ideal assessment curve and combining it with a standard assessment curve to form a growth and development sample curve includes:

[0024] Obtain ideal evaluation curve and standard evaluation curve;

[0025] Setting a plurality of equidistant nodes on the ideal evaluation curve and the standard evaluation curve, wherein the equidistant nodes on the ideal evaluation curve and the standard evaluation curve are aligned with each other on the time axis;

[0026] The first indicator data corresponding to the equidistant nodes are collected and fused and calculated to obtain the fused development index, which is then sorted according to the time axis to form a growth and development sample curve.

[0027] In a preferred embodiment, the step of outputting the initial state of growth and development based on the growth and development sample curve includes:

[0028] Collecting first indicator data of the child to be evaluated, aligning the data with the growth and development sample curve on the time axis, and determining the deviation value between the first indicator data of the child to be evaluated and the corresponding node on the growth and development sample curve;

[0029] The assessment phase of the children to be assessed is segmented to obtain multiple assessment sub-phases;

[0030] Collect the average of the deviation values ​​between the first indicator data of the child to be evaluated and the corresponding node on the growth and development sample curve in each evaluation sub-stage to form the sub-stage deviation mean;

[0031] When the mean value of the sub-stage deviation is positive, it means that the growth and development of the child to be evaluated in the corresponding evaluation sub-stage is better than the growth and development sample curve;

[0032] When the mean value of the sub-stage deviation is negative, it means that the growth and development of the child to be evaluated in the corresponding evaluation sub-stage is worse than the growth and development sample curve;

[0033] The weighted summation of the mean deviations of the sub-stages in each assessment sub-stage is performed to obtain the initial growth and development status score of the child to be assessed;

[0034] The initial growth and development status score is compared with the preset growth and development status threshold, and when the initial growth and development status score is lower than the preset growth and development status threshold, the child to be evaluated is judged to be in a state of growth retardation; otherwise, the child to be evaluated is judged to be in a good growth and development state.

[0035] In a preferred embodiment, collecting the second indicator data by a non-invasive collection device and adjusting the collection frequency based on the initial state type include:

[0036] Use non-invasive acquisition equipment to continuously collect bone growth plate micro-vibration signals and simultaneously obtain multispectral imaging data and blood oxygen saturation parameters;

[0037] Adjust the collection frequency of the non-invasive collection device according to the initial state;

[0038] Among them, for children who are in a state of growth retardation, the collection frequency is increased to monitor the growth and development of the children to be collected. For children who are in a state of good growth and development, the current collection frequency is maintained to continue monitoring, and when the monitoring results continuously show that the growth and development are in a good state, the collection frequency of the non-invasive collection equipment is reduced.

[0039] In a preferred embodiment, the non-invasive acquisition equipment includes a non-contact millimeter-wave radar array and a multispectral imaging cabin, wherein the millimeter-wave radar array is used to acquire bone contour point cloud data, and the multispectral imaging cabin is used to acquire blood oxygen saturation thermal maps and nighttime limb activity data.

[0040] In the preferred solution, the non-invasive acquisition device includes a mobile phone TOF camera without infrared fill light, as well as a mobile phone multi-camera fusion, a smartphone time-of-flight camera, and a disabled infrared fill light module;

[0041] If the ambient light intensity is greater than 200 lux, the TOF camera will actively transmit a pulse signal.

[0042] If the ambient light intensity is ≤200 lux, switch to the passive optical flow tracking mode of the multi-camera fusion module.

[0043] In a preferred embodiment, the step of inputting the second indicator data into a pre-trained convolutional neural network model to generate an individual growth fingerprint includes:

[0044] Preprocessing the second indicator data includes denoising, normalization, and missing data filling;

[0045] Inputting the preprocessed second indicator data into a pre-trained convolutional neural network model for feature extraction, the pre-trained convolutional neural network model includes a skeletal feature branch, a metabolic feature branch, and a behavioral feature branch;

[0046] Among them, the skeleton feature branch is used to extract the features of the skeleton contour point cloud data, the metabolic feature branch is used to extract the features of the blood oxygen saturation thermal map and skin metabolic imaging data, and the behavioral feature branch is used to extract the features of the nighttime limb activity data;

[0047] The features extracted from the skeletal feature branch, metabolic feature branch, and behavioral feature branch are integrated to generate an individual growth fingerprint that includes skeletal growth, metabolic status, and behavioral characteristics.

[0048] In a preferred embodiment, the steps of obtaining historical fingerprint data with a matching similarity greater than a preset value from a cloud database, generating a personalized growth corridor, triggering an early warning when the frequency or duration of the second indicator data collected in real time breaking through the growth corridor boundary exceeds a corresponding threshold, and generating a personalized growth and development assessment report include:

[0049] Extract known historical fingerprint data from the cloud database;

[0050] Calculate the similarity between historical fingerprint data and individual growth fingerprint;

[0051] Filter out historical fingerprint data with similarity greater than a preset value and sort them according to the time axis to form a reference growth and development trajectory;

[0052] Based on the reference growth and development trajectory, a personalized growth corridor with similar growth and development characteristics to the child to be assessed is constructed;

[0053] Collect the second indicator data of the child to be evaluated in real time and compare it with the upper and lower boundaries of the personalized growth corridor;

[0054] Establish a sliding time window mechanism and count the frequency or duration of the second indicator data breaking through the personalized growth corridor boundary within the preset time window;

[0055] When the frequency or duration of the second indicator data breaking through the boundary of the personalized growth corridor exceeds the corresponding threshold, an early warning signal is triggered, indicating that the growth and development of the child to be evaluated is abnormal, and a personalized growth and development assessment report is generated simultaneously.

[0056] Otherwise, the current monitoring status will be maintained, and a personalized growth and development assessment report will be output simultaneously to show that the growth and development of the child to be assessed is within the normal range.

[0057] The present invention also provides a child development dynamic assessment system based on multi-source data fusion, using the above-mentioned child development dynamic assessment method based on multi-source data fusion, including:

[0058] A first acquisition module is used to obtain first indicator data of the child and match it with a corresponding standard assessment curve, the first indicator data including the child's age, gender, height and weight;

[0059] The curve fitting module is used to obtain the ideal evaluation curve and combine it with the standard evaluation curve to form a growth and development sample curve;

[0060] A first evaluation module is used to output the initial state of growth and development based on the growth and development sample curve, and record the first indicator data in the initial state;

[0061] A second acquisition module is configured to collect second indicator data of the child using a non-invasive acquisition device and adjust the acquisition frequency based on the initial state type, wherein the second indicator data includes bone contour point cloud data, blood oxygen saturation thermal map, and nighttime limb activity data;

[0062] A fingerprint construction module is used to input the second indicator data into a pre-trained convolutional neural network model to generate an individual growth fingerprint;

[0063] The second evaluation module is used to obtain historical fingerprint data with matching similarity greater than a preset value from the cloud database, generate a personalized growth corridor, and trigger an early warning when the frequency or duration of the second indicator data collected in real time breaking through the growth corridor boundary exceeds the corresponding threshold, and generate a personalized growth and development evaluation report.

[0064] The technical effects achieved by the present invention are:

[0065] The present invention collects the first indicator data of the child and matches it with the standard evaluation curve, combines the ideal evaluation curve with the standard evaluation curve, and forms a more accurate growth and development sample curve, which can preliminarily judge the growth and development status of the child. The second indicator data of the child is obtained by using non-invasive collection equipment, and the collection frequency is adjusted based on the strategy of the initial state type, which not only ensures the comprehensiveness and accuracy of the data, but also avoids unnecessary frequent collection and improves the evaluation efficiency. The second indicator data is processed by a pre-trained convolutional neural network model to generate an individual growth fingerprint, and historical fingerprint data with high similarity to the individual growth fingerprint is screened out from the cloud database to construct a personalized growth corridor, which provides a personalized reference trajectory range for the child's growth and development. Finally, the frequency or duration of boundary breakthroughs is counted through a sliding time window mechanism to determine whether the child's growth and development is abnormal, thereby achieving an accurate assessment of the child's growth and development. BRIEF DESCRIPTION OF THE DRAWINGS

[0066] Figure 1 It is a schematic flow chart of the method of the present invention;

[0067] Figure 2 It is a schematic diagram of the system module of the present invention. DETAILED DESCRIPTION

[0068] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0069] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways different from those described herein. Those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0070] Secondly, the term "one embodiment" or "embodiment" herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in a preferred embodiment" appearing in various places throughout this specification does not necessarily refer to the same embodiment, nor does it constitute a separate or selective embodiment that is mutually exclusive of other embodiments.

[0071] With the development of society and the improvement of living standards, the healthy growth of children is increasingly valued by parents and society. However, the assessment and monitoring of children's growth and development still face many challenges, such as the accuracy of the assessment methods, the convenience of the assessment process, and the timeliness of the assessment results. Traditional assessment methods often rely on the doctor's experience and judgment, lack objective quantitative standards, and have a long assessment cycle, which makes it difficult to meet the modern family's needs for the immediacy and continuity of children's growth and development monitoring.

[0072] See also Figure 1 As shown, the present invention provides a method for dynamic assessment of child development based on multi-source data fusion, comprising:

[0073] S1. Obtaining first indicator data of the child and matching it with a corresponding standard assessment curve, the first indicator data including the child's age, gender, height, and weight;

[0074] In step S1, the first indicator data of the child is first obtained. The first indicator data includes but is not limited to key information such as the child's age, gender, height, and weight, and is matched with a pre-set standard evaluation curve to provide a benchmark reference for subsequent evaluation. The steps of obtaining the first indicator data of the child and matching it with the corresponding standard evaluation curve include:

[0075] Classify the first indicator data according to the age groups of the children to obtain multiple classification subsets;

[0076] Calculate the average value of the first indicator data in each classification subset to form the average growth data of each age group and record it as the evaluation condition parameter;

[0077] Perform bidirectional offset processing on the evaluation condition parameters to obtain the initial offset boundary;

[0078] Collect the quantity of the first indicator data under the initial offset boundary and record it as the screening condition parameter;

[0079] Obtaining a screening threshold and comparing the screening condition parameter with the screening threshold;

[0080] When the screening condition parameter is greater than the screening threshold, it indicates that the initial offset boundary corresponding to the screening condition parameter does not meet the offset requirement, and the offset operation continues until the screening condition parameter corresponding to the initial offset boundary is equal to the screening threshold, and the corresponding initial offset boundary is recorded as the standard boundary;

[0081] When the screening condition parameter is equal to the screening threshold, it indicates that the initial offset boundary corresponding to the screening condition parameter meets the offset requirement, and the corresponding initial offset boundary is recorded as the standard boundary;

[0082] When the screening condition parameter is less than the screening threshold, it indicates that the initial offset boundary corresponding to the screening condition parameter is over-shifted, and the initial offset boundary is reversed until the initial offset boundary after reverse shifting is greater than or equal to the screening threshold for the first time, and the standard boundary is obtained;

[0083] Determine the standard range for differentiating children's growth and development based on the standard boundaries, and construct a standard assessment curve for children based on the standard range;

[0084] Specifically, when matching the first indicator data with the corresponding standard evaluation curve, the first indicator data obtained are first classified according to the different age groups of the children, so as to obtain multiple classification subsets. Here, the first indicator data of the children are preferably children of the same region, gender and age group, so as to ensure the accuracy of the data and the fairness of the evaluation. Then, the average value is calculated for the first indicator data in each classification subset to form the average growth data of each age group, and the average growth data is recorded as the evaluation condition parameter required for subsequent evaluation. Then, the recorded evaluation condition parameter is bidirectionally offset to determine the initial offset boundary to ensure the flexibility and adaptability of the data. On this basis, the number of first indicator data under the initial offset boundary conditions is collected and recorded as the screening condition parameter. Then, a pre-set screening threshold is introduced, and the recorded screening condition parameter is compared with the screening threshold. The thresholds are compared and analyzed. When the screening condition parameter is greater than the screening threshold, it indicates that the current initial offset boundary fails to meet the offset requirement, and the offset operation needs to be continued until the screening condition parameter corresponding to the initial offset boundary is exactly equal to the screening threshold, and the offset is stopped. The initial offset boundary at this time is recorded as the standard boundary. When the screening condition parameter is equal to the screening threshold, it indicates that the current initial offset boundary has fully met the offset requirement, and the corresponding initial offset boundary is directly recorded as the standard boundary. When the screening condition parameter is less than the screening threshold, it indicates that the initial offset boundary has over-offset, and the initial offset boundary needs to be reversed and adjusted until the initial offset boundary after the reverse offset is greater than or equal to the screening threshold for the first time, and the adjustment is stopped, thereby determining the standard boundary. Finally, based on the determined standard boundary, the standard range for distinguishing children's growth and development status is clarified, and a standard evaluation curve is constructed on this basis.

[0085] S2. Obtain the ideal assessment curve and combine it with the standard assessment curve to form a growth and development sample curve;

[0086] In step S2, after the standard assessment curve is output, an ideal assessment curve for children of the same period is also introduced. The ideal assessment curve represents the growth and development trajectory of the child under an ideal state. The ideal assessment curve is then organically combined with the previously matched standard assessment curve to form a comprehensive growth and development sample curve, which serves as an important basis for evaluating the growth and development status of the child. The step of obtaining the ideal assessment curve and combining it with the standard assessment curve to form the growth and development sample curve includes:

[0087] Obtain ideal evaluation curve and standard evaluation curve;

[0088] Setting a plurality of equidistant nodes on the ideal evaluation curve and the standard evaluation curve, wherein the equidistant nodes on the ideal evaluation curve and the standard evaluation curve are aligned with each other on the time axis;

[0089] Collecting the first indicator data corresponding to the equidistant nodes and performing fusion calculation to obtain the fused development index, and then sorting the fused development index according to the time axis to form a growth and development sample curve;

[0090] Specifically, when determining the growth and development sample curve, it is first necessary to obtain the ideal evaluation curve and the standard evaluation curve respectively. Next, multiple equidistant nodes are set on the ideal evaluation curve and the standard evaluation curve respectively. It is particularly important to note that the equidistant nodes on the ideal evaluation curve and the standard evaluation curve must be aligned with each other on the time axis to ensure the consistency and comparability of the data. Then the first indicator data corresponding to the equidistant nodes are collected. After the collection is completed, the first indicator data is fused and calculated to obtain a fused development index. The fusion calculation process is carried out through weighted averaging, data normalization, etc. The purpose is to integrate the ideal and standard information to form a more comprehensive fused development index. Finally, the fused development index is arranged in order according to the time axis to form a complete growth and development sample curve, thereby providing a scientific basis for the growth and development assessment of individuals or groups.

[0091] S3. Based on the growth and development sample curve, output the initial state of growth and development, and record the first indicator data in the initial state;

[0092] In step S3, after the growth and development sample curve is output, first indicator data of the child to be evaluated is collected, and based on the generated growth and development sample curve, the initial state of growth and development is preliminarily determined and output. The step of outputting the initial state of growth and development based on the growth and development sample curve includes:

[0093] Collecting first indicator data of the child to be evaluated, aligning the data with the growth and development sample curve on the time axis, and determining the deviation value between the first indicator data of the child to be evaluated and the corresponding node on the growth and development sample curve;

[0094] The assessment phase of the children to be assessed is segmented to obtain multiple assessment sub-phases;

[0095] Collect the average of the deviation values ​​between the first indicator data of the child to be evaluated and the corresponding node on the growth and development sample curve in each evaluation sub-stage to form the sub-stage deviation mean;

[0096] When the mean value of the sub-stage deviation is positive, it means that the growth and development of the child to be evaluated in the corresponding evaluation sub-stage is better than the growth and development sample curve;

[0097] When the mean value of the sub-stage deviation is negative, it means that the growth and development of the child to be evaluated in the corresponding evaluation sub-stage is worse than the growth and development sample curve;

[0098] The weighted summation of the mean deviations of the sub-stages in each assessment sub-stage is performed to obtain the initial growth and development status score of the child to be assessed;

[0099] Comparing the initial growth and development status score with the preset growth and development status threshold, and when the initial growth and development status score is lower than the preset growth and development status threshold, determining that the child to be evaluated is in a state of growth retardation; otherwise, determining that the child to be evaluated is in a state of good growth and development;

[0100] Specifically, when outputting the initial state of growth and development of the child to be evaluated, the first indicator data of the child to be evaluated will be collected first, and the collected first indicator data of the child to be evaluated will be aligned with the growth and development sample curve on the time axis. Through comparative analysis, the deviation value between the first indicator data of the child to be evaluated and the corresponding node on the growth and development sample curve is determined, thereby reflecting the difference between the growth and development of the child to be evaluated at a certain point in time and the standard sample. Then, the evaluation stage of the child to be evaluated is segmented and divided into multiple evaluation sub-stages, so as to analyze the growth and development in different time periods. The purpose is to accurately capture the changing trend of growth and development. In each evaluation sub-stage, the first indicator data of the child to be evaluated is collected again and compared with the corresponding node on the growth and development sample curve. The average value of the deviation value between the first indicator data of the child to be evaluated and the corresponding node on the growth and development sample curve in each evaluation sub-stage is calculated to form the sub-stage deviation mean ( , where Indicates the The mean deviation of the evaluation sub-stages, represents the number of data points within the evaluation sub-stage, Indicates the The time interval of each evaluation sub-stage, Indicates a time point Deviation value of the sub-stage deviation), if the value of the sub-stage deviation mean is positive, it means that in the corresponding assessment sub-stage, the growth and development of the child to be assessed is better than the growth and development sample curve, that is, its growth and development speed or level exceeds the standard sample. On the contrary, if the value of the sub-stage deviation mean is negative, it means that the growth and development of the child to be assessed in the corresponding assessment sub-stage is worse than the growth and development sample curve, that is, its growth and development speed or level does not meet the requirements of the standard sample. In order to comprehensively evaluate the initial growth and development status of the child to be assessed and avoid the influence of accidental factors, it is necessary to weightedly summarize the sub-stage deviation means in each assessment sub-stage, and finally obtain a comprehensive initial growth and development status score of the child to be assessed ( , where Indicates the initial growth and development status score, Indicates the The basic weight of each evaluation sub-stage, Represents the time attenuation coefficient, which is set according to expert experience. Indicates the The time interval from the first assessment sub-stage to the current time) is calculated, and finally the calculated initial growth and development status score is compared with the preset growth and development status threshold. If the initial growth and development status score is lower than the preset growth and development status threshold, the child to be assessed is judged to be in a state of growth retardation and needs to take timely intervention measures. Conversely, if the initial growth and development status score is higher than or equal to the preset growth and development status threshold, the child to be assessed is judged to be in a good growth and development state, and his or her growth and development process meets or exceeds the normal standard. At this time, the existing nutrition and life guidance plan can be maintained without additional intervention.

[0101] S4. Collecting the child's second indicator data using a non-invasive collection device, and adjusting the collection frequency based on the initial state type, wherein the second indicator data includes skeletal contour point cloud data, blood oxygen saturation thermal map, and nighttime limb activity data;

[0102] In step S4, after the initial state assessment of the child to be assessed is completed, the child's second indicator data will be continuously collected through a non-invasive collection device. The second indicator data includes more comprehensive physiological information such as bone contour point cloud data, blood oxygen saturation thermal map, and nighttime limb activity data. According to the type of initial state, the frequency of data collection will be intelligently adjusted to ensure the accuracy and real-time nature of the data. The second indicator data is collected through the non-invasive collection device, and the collection frequency is adjusted based on the initial state type, including:

[0103] Use non-invasive acquisition equipment to continuously collect bone growth plate micro-vibration signals and simultaneously obtain multispectral imaging data and blood oxygen saturation parameters;

[0104] Adjust the collection frequency of the non-invasive collection device according to the initial state;

[0105] For children with growth retardation, the collection frequency will be increased to monitor the growth and development of the children to be collected. For children with good growth and development, the current collection frequency will be maintained and the collection frequency of the non-invasive collection device will be reduced when the monitoring results show that the growth and development are in good condition.

[0106] Specifically, first, non-invasive collection equipment is used to continuously collect micro-vibration signals generated by the bone growth plate. At the same time, multispectral imaging data is synchronously acquired to ensure the comprehensiveness and accuracy of the data. According to the evaluation results of the initial state, the collection frequency of the non-invasive collection equipment is adjusted accordingly. Specifically, for children with delayed growth and development, in order to more closely monitor their growth and development, it is necessary to increase the collection frequency so as to timely discover and intervene in potential problems. For children in a good growth and development state, the current collection frequency is maintained and routine monitoring continues. In addition, when the monitoring results continuously show a good growth and development state, in order to reduce unnecessary intervention and resource consumption, the collection frequency of the non-invasive collection equipment can be appropriately reduced to achieve more efficient and accurate monitoring and management.

[0107] In addition, in medical institutions, non-invasive collection equipment includes non-contact millimeter-wave radar arrays and multispectral imaging cabins. The millimeter-wave radar array is connected and collected in the 60-64GHz frequency band to obtain bone contour point cloud data, and the multispectral imaging cabin is used for imaging in the 400-1000nm wavelength range to obtain blood oxygen saturation thermal maps and nighttime limb activity data.

[0108] In a home environment, non-invasive acquisition devices include mobile phone TOF cameras without infrared fill light, as well as mobile phone multi-camera fusion and smartphone time-of-flight cameras, which collect bone contour point cloud data at a frame rate of 30fps and disable the infrared fill light module;

[0109] The multi-camera fusion module of the smartphone uses an RGB camera and a monocular depth camera to collaboratively acquire skin metabolism imaging data;

[0110] If the ambient light intensity is greater than 200 lux, the TOF camera will actively transmit a pulse signal.

[0111] If the ambient light intensity is ≤200 lux, switch to the passive optical flow tracking mode of the multi-camera fusion module;

[0112] In a home environment, if the ambient light intensity is high, exceeding 200 lux, the smartphone's time-of-flight (TOF) camera is used to actively emit pulse signals to capture subtle changes in bone contours. This active acquisition method can maintain high acquisition accuracy even in strong light environments. When the ambient light intensity is low, not exceeding 200 lux, it switches to the passive optical flow tracking mode of the smartphone's multi-camera fusion module, using the collaborative work of the RGB camera and the monocular depth camera to obtain skin metabolism imaging data. This mode can still operate effectively in low-light environments, ensuring data continuity and accuracy, allowing non-invasive acquisition equipment to maintain stable acquisition performance under different lighting conditions, providing more reliable data support for children's growth and development monitoring.

[0113] S5. Inputting the second indicator data into the pre-trained convolutional neural network model to generate an individual growth fingerprint;

[0114] In step S5, after the second indicator data is output, the collected second indicator data will be input into a pre-trained convolutional neural network model. The convolutional neural network can be a classic network structure such as VGGNet, ResNet or Inception, or a network structure customized according to actual needs. The second indicator data such as bone contour point cloud data, blood oxygen saturation thermal map and nighttime limb activity data are subjected to feature extraction and pattern recognition through a deep learning algorithm, and finally a unique individual growth fingerprint is generated to comprehensively reflect the growth and development characteristics of the child. The step of inputting the second indicator data into the pre-trained convolutional neural network model to generate the individual growth fingerprint includes:

[0115] Preprocessing the second indicator data includes denoising, normalization, and missing data filling;

[0116] Inputting the preprocessed second indicator data into a pre-trained convolutional neural network model for feature extraction, the pre-trained convolutional neural network model includes a skeletal feature branch, a metabolic feature branch, and a behavioral feature branch;

[0117] Among them, the skeleton feature branch is used to extract the features of the skeleton contour point cloud data, the metabolic feature branch is used to extract the features of the blood oxygen saturation thermal map and skin metabolic imaging data, and the behavioral feature branch is used to extract the features of the nighttime limb activity data;

[0118] The features extracted from the skeletal feature branch, metabolic feature branch, and behavioral feature branch are fused to generate an individual growth fingerprint that includes skeletal growth, metabolic status, and behavioral characteristics.

[0119] Specifically, when outputting individual growth fingerprints, the second indicator data is first preprocessed accordingly. The preprocessing process includes several key links: the first is denoising, which aims to eliminate noise interference in the data and ensure the purity of the data; the second is normalization, which normalizes the data according to certain standards so that it can be compared and analyzed on a unified scale; the third is missing data filling, which adaptively fills the missing parts in the data to ensure the integrity and continuity of the data. The preprocessed second indicator data is then input into the pre-trained convolutional neural network model for feature extraction. The convolutional neural network model contains multiple feature branches, including bone feature branches, metabolic feature branches, and so on. The skeletal feature branch is specifically used to extract the features of the skeletal contour point cloud data. By analyzing the morphology and structure of the bones, it captures the key information of individual bone growth. The metabolic feature branch is responsible for extracting the features of the blood oxygen saturation thermogram and skin metabolic imaging data to reveal the individual's metabolic status. The behavioral feature branch is used to extract the features of the nighttime limb activity data. By monitoring and analyzing nighttime behavioral activities, it understands the individual's behavioral patterns and habits. Finally, the features extracted from the skeletal feature branch, metabolic feature branch, and behavioral feature branch are organically integrated to generate a comprehensive, multi-dimensional individual growth fingerprint that includes skeletal growth, metabolic status, and behavioral characteristics.

[0120] S6. Obtain historical fingerprint data with a matching similarity greater than a preset value from the cloud database, generate a personalized growth corridor, and trigger an alarm when the frequency or duration of the second indicator data collected in real time exceeding the growth corridor boundary exceeds the corresponding threshold, and generate a personalized growth and development assessment report;

[0121] In step S6, after the individual growth fingerprint of the child to be evaluated is output, historical fingerprint data with a similarity greater than a preset value to the growth fingerprint of the current child to be evaluated is retrieved and obtained from a cloud database (the cloud database stores a large amount of historical fingerprint data, which comes from children of different ages, genders, regions, and other backgrounds and has broad representativeness and diversity). A personalized growth corridor is generated based on the extracted historical fingerprint data. After the personalized growth corridor is constructed, an early warning mechanism is automatically triggered when the second indicator data collected in real time frequently breaks through the growth corridor boundary, or the duration of the breakthrough exceeds a preset threshold, and a personalized growth and development assessment report is generated to provide a corresponding reference for parents and doctors. The steps of obtaining historical fingerprint data with a matching similarity greater than a preset value from the cloud database, generating a personalized growth corridor, and triggering an early warning when the frequency or duration of the second indicator data collected in real time breaking through the growth corridor boundary exceeds the corresponding threshold, and generating a personalized growth and development assessment report include:

[0122] Extract known historical fingerprint data from the cloud database;

[0123] Calculate the similarity between historical fingerprint data and individual growth fingerprint;

[0124] Filter out historical fingerprint data with similarity greater than a preset value and sort them according to the time axis to form a reference growth and development trajectory;

[0125] Based on the reference growth and development trajectory, a personalized growth corridor with similar growth and development characteristics to the child to be assessed is constructed;

[0126] Collect the second indicator data of the child to be evaluated in real time and compare it with the upper and lower boundaries of the personalized growth corridor;

[0127] Establish a sliding time window mechanism and count the frequency or duration of the second indicator data breaking through the personalized growth corridor boundary within the preset time window;

[0128] When the frequency or duration of the second indicator data breaking through the boundary of the personalized growth corridor exceeds the corresponding threshold, an early warning signal is triggered, indicating that the growth and development of the child to be evaluated is abnormal, and a personalized growth and development assessment report is generated simultaneously.

[0129] Otherwise, the current monitoring status will be maintained, and a personalized growth and development assessment report will be output simultaneously, indicating that the growth and development of the child to be assessed is within the normal range;

[0130] Specifically, it is necessary to first extract known historical fingerprint data from the cloud database. The historical fingerprint data contains a large amount of growth and development information of different individuals, and then calculate the similarity between these historical fingerprint data and the growth fingerprint of the individual to be evaluated (specifically, cosine similarity, Euclidean distance or Manhattan distance can be used as the similarity measurement standard). Then, filter out the historical fingerprint data with a similarity greater than the preset value, and sort them according to the time axis to form a reference growth and development trajectory, which provides a reference basis for the subsequent construction of a personalized growth corridor. Then, based on the reference growth and development trajectory, construct a personalized growth corridor with similar growth and development characteristics to the child to be evaluated. The personalized growth corridor can reflect the expected growth and development trend of the child to be evaluated in the future. During the real-time monitoring process, the real-time collected data will be used to generate a reference for the growth and development of the child. The second indicator data of the child to be evaluated is compared with the upper and lower boundaries of the personalized growth corridor, and the frequency or duration of the second indicator data breaking through the personalized growth corridor boundary within the preset time window is counted. When the frequency or duration of the second indicator data breaking through the personalized growth corridor boundary exceeds the corresponding threshold, an early warning signal will be triggered, indicating that the growth and development of the child to be evaluated may be abnormal, and a detailed personalized growth and development assessment report will be generated simultaneously so that parents and doctors can take intervention measures in time. On the contrary, if the frequency or duration of the second indicator data not breaking through the growth corridor boundary does not exceed the corresponding threshold, the current monitoring status will be maintained, and a personalized growth and development assessment report indicating that the growth and development of the child to be evaluated is within the normal range will be output simultaneously to provide a reference for parents and doctors to ensure the healthy growth of children.

[0131] See also Figure 2 The child development dynamic assessment system based on multi-source data fusion uses the child development dynamic assessment method based on multi-source data fusion described above, including:

[0132] A first acquisition module is used to obtain first indicator data of the child and match it with a corresponding standard assessment curve, the first indicator data including the child's age, gender, height and weight;

[0133] The curve fitting module is used to obtain the ideal evaluation curve and combine it with the standard evaluation curve to form a growth and development sample curve;

[0134] A first evaluation module is used to output the initial state of growth and development based on the growth and development sample curve, and record the first indicator data in the initial state;

[0135] A second acquisition module is configured to collect second indicator data of the child using a non-invasive acquisition device and adjust the acquisition frequency based on the initial state type, wherein the second indicator data includes bone contour point cloud data, blood oxygen saturation thermal map, and nighttime limb activity data;

[0136] A fingerprint construction module is used to input the second indicator data into a pre-trained convolutional neural network model to generate an individual growth fingerprint;

[0137] The second evaluation module is used to obtain historical fingerprint data with matching similarity greater than a preset value from the cloud database, generate a personalized growth corridor, and trigger an early warning when the frequency or duration of the second indicator data collected in real time breaking through the growth corridor boundary exceeds the corresponding threshold, and generate a personalized growth and development evaluation report.

[0138] Specifically, the execution process of the child development dynamic assessment system based on multi-source data fusion is consistent with the above-mentioned child development dynamic assessment method based on multi-source data fusion, and will not be repeated here.

[0139] It should be noted that, in this document, the terms "comprises," "includes," or any other variations thereof are intended to encompass non-exclusive inclusion, such that a process, apparatus, article, or method comprising a series of elements includes not only those elements but also other elements not explicitly listed, or elements inherent to such process, apparatus, article, or method. In the absence of further limitations, an element defined by the phrase "comprising a ..." does not exclude the presence of other identical elements in the process, apparatus, article, or method comprising the element.

[0140] The foregoing is merely a preferred embodiment of the present invention. It should be noted that those skilled in the art may make various improvements and modifications without departing from the principles of the present invention, and such improvements and modifications are also within the scope of protection of the present invention. Structures, devices, and operating methods not specifically described or explained herein shall, unless otherwise specified or limited, be implemented in accordance with conventional means in the art.

Claims

1. A dynamic assessment method for child development based on multi-source data fusion, characterized by: include: Obtaining the child's first indicator data and matching it with a corresponding standard assessment curve, the first indicator data including the child's age, gender, height, and weight; Obtain the ideal assessment curve and combine it with the standard assessment curve to form a growth and development sample curve; Based on the growth and development sample curve, output the initial state of growth and development, and record the first indicator data under the initial state; Using non-invasive data collection equipment, collect the child's secondary indicator data, and adjust the collection frequency based on the initial state type. The secondary indicator data includes bone contour point cloud data, blood oxygen saturation thermal map, and nighttime limb activity data; The second indicator data is input into the pre-trained convolutional neural network model to generate an individual growth fingerprint; Obtain historical fingerprint data with a matching similarity greater than a preset value from the cloud database to generate a personalized growth corridor. When the frequency or duration of the second indicator data collected in real time exceeding the growth corridor boundary exceeds the corresponding threshold, an early warning is triggered and a personalized growth and development assessment report is generated. The step of obtaining the first indicator data of the child and matching the corresponding standard assessment curve includes: Classify the first indicator data according to the age groups of the children to obtain multiple classification subsets; Calculating the average value of the first indicator data in each of the classification subsets to form average growth data for each age group, and recording the average growth data as an evaluation condition parameter; Performing bidirectional offset processing on the evaluation condition parameters to obtain an initial offset boundary; Collect the quantity of the first indicator data under the initial offset boundary and record it as the screening condition parameter; Obtaining a screening threshold and comparing the screening condition parameter with the screening threshold; When the screening condition parameter is greater than the screening threshold, it indicates that the initial offset boundary corresponding to the screening condition parameter does not meet the offset requirement, and the offset operation continues until the screening condition parameter corresponding to the initial offset boundary is equal to the screening threshold, and the corresponding initial offset boundary is recorded as the standard boundary; When the screening condition parameter is equal to the screening threshold, it indicates that the initial offset boundary corresponding to the screening condition parameter meets the offset requirement, and the corresponding initial offset boundary is recorded as the standard boundary; When the screening condition parameter is less than the screening threshold, it indicates that the initial offset boundary corresponding to the screening condition parameter is over-shifted, and the initial offset boundary is reversed until the initial offset boundary after reverse shifting is greater than or equal to the screening threshold for the first time, and the standard boundary is obtained; Determine the standard range for differentiating children's growth and development based on the standard boundaries, and construct a standard assessment curve for children based on the standard range; The step of obtaining the ideal assessment curve and combining it with the standard assessment curve to form a growth and development sample curve includes: Obtain ideal evaluation curve and standard evaluation curve; Setting a plurality of equidistant nodes on the ideal evaluation curve and the standard evaluation curve, wherein the equidistant nodes on the ideal evaluation curve and the standard evaluation curve are aligned with each other on the time axis; The first indicator data corresponding to the equidistant nodes are collected and fused and calculated to obtain the fused development index, which is then sorted according to the time axis to form a growth and development sample curve.

2. The method for dynamic assessment of child development based on multi-source data fusion according to claim 1, characterized in that: The step of outputting the initial state of growth and development based on the growth and development sample curve includes: Collecting first indicator data of the child to be evaluated, aligning the data with the growth and development sample curve on the time axis, and determining the deviation value between the first indicator data of the child to be evaluated and the corresponding node on the growth and development sample curve; The assessment phase of the children to be assessed is segmented to obtain multiple assessment sub-phases; Collect the average of the deviation values ​​between the first indicator data of the child to be evaluated and the corresponding node on the growth and development sample curve in each evaluation sub-stage to form the sub-stage deviation mean; When the mean value of the sub-stage deviation is positive, it means that the growth and development of the child to be evaluated in the corresponding evaluation sub-stage is better than the growth and development sample curve; When the mean value of the sub-stage deviation is negative, it means that the growth and development of the child to be evaluated in the corresponding evaluation sub-stage is worse than the growth and development sample curve; The weighted summation of the mean deviations of the sub-stages in each assessment sub-stage is performed to obtain the initial growth and development status score of the child to be assessed; The initial growth and development status score is compared with the preset growth and development status threshold, and when the initial growth and development status score is lower than the preset growth and development status threshold, the child to be evaluated is judged to be in a state of growth retardation; otherwise, the child to be evaluated is judged to be in a good growth and development state.

3. The method for dynamic assessment of child development based on multi-source data fusion according to claim 1, characterized in that: The collecting of the second indicator data by the non-invasive collection device and adjusting the collection frequency based on the initial state type include: Use non-invasive acquisition equipment to continuously collect bone growth plate micro-vibration signals and simultaneously obtain multispectral imaging data and blood oxygen saturation parameters; Adjust the collection frequency of the non-invasive collection device according to the initial state; Among them, for children who are in a state of growth retardation, the collection frequency is increased to monitor the growth and development of the children to be collected. For children who are in a state of good growth and development, the current collection frequency is maintained to continue monitoring, and when the monitoring results continuously show that the growth and development are in a good state, the collection frequency of the non-invasive collection equipment is reduced.

4. The method for dynamic assessment of child development based on multi-source data fusion according to claim 1, characterized in that: The non-invasive acquisition equipment includes a non-contact millimeter-wave radar array and a multispectral imaging cabin, wherein the millimeter-wave radar array is used to collect and obtain bone contour point cloud data, and the multispectral imaging cabin is used to obtain blood oxygen saturation thermal maps and nighttime limb activity data.

5. The method for dynamic assessment of child development based on multi-source data fusion according to claim 1, characterized in that: The non-invasive acquisition equipment includes a mobile phone TOF camera without infrared fill light, as well as a mobile phone multi-camera fusion, a smartphone time-of-flight camera, and a disabled infrared fill light module; If the ambient light intensity is greater than 200 lux, the TOF camera will actively transmit a pulse signal. If the ambient light intensity is ≤200 lux, switch to the passive optical flow tracking mode of the multi-camera fusion module.

6. The method for dynamic assessment of child development based on multi-source data fusion according to claim 1, characterized in that: The step of inputting the second indicator data into the pre-trained convolutional neural network model to generate an individual growth fingerprint includes: Preprocessing the second indicator data includes denoising, normalization, and missing data filling; Inputting the preprocessed second indicator data into a pre-trained convolutional neural network model for feature extraction, the pre-trained convolutional neural network model includes a skeletal feature branch, a metabolic feature branch, and a behavioral feature branch; Among them, the skeleton feature branch is used to extract the features of the skeleton contour point cloud data, the metabolic feature branch is used to extract the features of the blood oxygen saturation thermal map and skin metabolic imaging data, and the behavioral feature branch is used to extract the features of the nighttime limb activity data; The features extracted from the skeletal feature branch, metabolic feature branch, and behavioral feature branch are fused to generate an individual growth fingerprint that includes skeletal growth, metabolic status, and behavioral characteristics.

7. The method for dynamic assessment of child development based on multi-source data fusion according to claim 1, characterized in that: The steps of obtaining historical fingerprint data with a matching similarity greater than a preset value from a cloud database, generating a personalized growth corridor, triggering an early warning when the frequency or duration of the second indicator data collected in real time breaking through the growth corridor boundary exceeds a corresponding threshold, and generating a personalized growth and development assessment report include: Extract known historical fingerprint data from the cloud database; Calculate the similarity between historical fingerprint data and individual growth fingerprint; Filter out historical fingerprint data with similarity greater than a preset value and sort them according to the time axis to form a reference growth and development trajectory; Based on the reference growth and development trajectory, a personalized growth corridor with similar growth and development characteristics to the child to be assessed is constructed; Collect the second indicator data of the child to be evaluated in real time and compare it with the upper and lower boundaries of the personalized growth corridor; Establish a sliding time window mechanism and count the frequency or duration of the second indicator data breaking through the personalized growth corridor boundary within the preset time window; When the frequency or duration of the second indicator data exceeding the personalized growth corridor boundary exceeds the corresponding threshold, an early warning signal is triggered, indicating that the growth and development of the child to be evaluated is abnormal, and a personalized growth and development assessment report is generated simultaneously; Otherwise, the current monitoring status will be maintained, and a personalized growth and development assessment report will be output simultaneously to show that the growth and development of the child to be assessed is within the normal range.

8. A dynamic assessment system for child development based on multi-source data fusion, characterized by: The method for dynamic assessment of child development based on multi-source data fusion according to any one of claims 1 to 7 comprises: A first acquisition module is used to obtain first indicator data of the child and match it with a corresponding standard assessment curve, the first indicator data including the child's age, gender, height and weight; The curve fitting module is used to obtain the ideal evaluation curve and combine it with the standard evaluation curve to form a growth and development sample curve; A first evaluation module is used to output the initial state of growth and development based on the growth and development sample curve, and record the first indicator data in the initial state; A second acquisition module is configured to collect second indicator data of the child using a non-invasive acquisition device and adjust the acquisition frequency based on the initial state type, wherein the second indicator data includes bone contour point cloud data, blood oxygen saturation thermal map, and nighttime limb activity data; A fingerprint construction module is used to input the second indicator data into a pre-trained convolutional neural network model to generate an individual growth fingerprint; The second evaluation module is used to obtain historical fingerprint data with matching similarity greater than a preset value from the cloud database, generate a personalized growth corridor, and trigger an early warning when the frequency or duration of the second indicator data collected in real time breaking through the growth corridor boundary exceeds the corresponding threshold, and generate a personalized growth and development evaluation report.

Images