A dual-scale fusion method for age-specific dual-track psychological assessment of adolescents

Abstract

This invention discloses a dual-scale fusion method for age-appropriate dual-track psychological assessment of adolescents, belonging to the field of adolescent psychological assessment technology. The method includes the following steps: S1, anchoring the age-appropriate dual-track system; S2, decomposing the scale kernel; S3, constructing a weighted algorithm for dual-scale fusion; S4, dynamically adapting age-appropriate norms; S5, fully embedding assessment factors into the script; and S6, seamlessly collecting and calculating multimodal data. This invention constructs an age-appropriate dual-track assessment system for adolescents aged 6-16, adapting single-scale and dual-scale fusion assessment modes with 12 years old as the dividing line. This achieves accurate assessment adaptation across all age groups and full coverage of scale dimensions, effectively solving the problems of distortion, poor adaptability, and low student participation in traditional assessments. It significantly improves the reliability and validity of the assessment and student participation, while forming a closed loop of assessment, early warning, and intervention, fully aligning with domestic educational compliance requirements and the psychological development patterns of adolescents.

Description

Technical Field

[0001] This invention relates to the field of adolescent psychological assessment technology, specifically a dual-scale fusion method for age-specific dual-track psychological assessment of adolescents. Background Technology

[0002] With the implementation of the "Special Action Plan for Comprehensively Strengthening and Improving Students' Mental Health Work in the New Era (2023-2025)" issued by the Ministry of Education and 17 other departments, the normalization of mental health screening in primary and secondary schools has become a mandatory policy requirement, and digital psychological assessment has become a core tool for campus mental health work. However, the industry generally adopts a universal model that uses a single scale to cover the entire age range of 6-16 years old, completely ignoring the essential differences in cognitive level, psychological characteristics, and behavioral expression patterns between children and adolescents before and after the 12-year-old adolescence dividing point. For children aged 6-12 in the primary school stage, the core characteristics of their cognitive level being insufficient to complete self-assessment questionnaires and the high comorbidity rate of psychological and developmental problems are not taken into account, resulting in seriously distorted assessment results that cannot truly reflect the mental health status of children. Summary of the Invention

[0003] To address the shortcomings of existing technologies, this invention provides a dual-scale fusion method for age-appropriate dual-track psychological assessment of adolescents, solving the problems mentioned in the background section.

[0004] To achieve the above objectives, the present invention provides the following technical solution: a dual-scale fusion method for age-appropriate dual-track psychological assessment of adolescents, comprising the following steps: S1, Age-Segmented Dual-Track System Anchoring: Using 12 years old as the core dividing point, we construct a dual-track assessment framework for adolescents aged 6-16, a standardized assessment system with a single scale for adolescents aged 12-16 and above, and a dual-scale integrated assessment system for children aged 6-12 and below. S2, Scale Core Disassembly: For the dual-track system, the core of the scales was standardized and decomposed. For adolescents aged 12 and above, the original core dimensions and corresponding clinical behavior factors of the "Middle School Students' Mental Health Scale" were extracted. For children under 12 years old, the core dimensions and clinical behavior factors of the main scale "Primary School Students' Mental Health Assessment Scale" and the auxiliary scale "Gesell Developmental Diagnostic Scale" were extracted respectively, and the dual-dimensional factor library of mental health and developmental level was constructed. S3, Construction of a dual-scale fusion weighted algorithm: For children under 12 years old, a dual-scale fusion weighting algorithm with fixed weights was constructed, with the main scale weight set at 60%, the auxiliary scale weight at 30%, and the age-specific norm adaptation weight at 10%, so as to achieve simultaneous assessment of the two scales and calculation of the comprehensive score in one data collection. S4, Age-Specific Norm Dynamic Adaptation: For children under 12 years old, the curriculum is further divided into two sub-levels: lower grades 1-3 and upper grades 4-6, with the third grade of primary school as the dividing line. Different norm thresholds, task difficulty, collection indicators and factor weights are matched for each sub-level to achieve dynamic adaptation of age-specific norms. S5, Evaluation Factor Fully Embedded Script: Based on the age-based dual-track clinical behavior factor library, a single-scene fixed main line and a 5-minute continuous closed loop of non-sensory assessment interaction scripts are designed for each age group. The collection indicators corresponding to each clinical behavior factor are 100% embedded in the storyline tasks of the script. There are no assessment, answering questions, or sensitive expressions for detection throughout the process, realizing that the interactive process is also the assessment process. S6. Seamless acquisition and computation of multimodal data: The script guides users through an immersive interactive experience, simultaneously collecting multimodal behavioral data. Based on the algorithm model corresponding to the age-based dual-track system, it completes standardized scale score calculation, intelligent risk level determination, and outputs professional assessment reports and intervention plans for the corresponding age group.

[0005] Furthermore, in step S2, the "Middle School Students' Mental Health Scale" is abbreviated as MSSMHS, the "Primary School Students' Mental Health Assessment Scale" is abbreviated as MHRSP, and the "Developmental Diagnostic Scale" is abbreviated as Gesell; The four core dimensions of MSSMHS and their corresponding clinical behavioral factors are as follows: Dimensions of emotional adaptation: anxiety, depression, impulsivity, and emotional stability; Learning adaptation dimensions: learning anxiety, concentration, procrastination, and learning efficacy; Interpersonal adaptation dimensions: social avoidance, parent-child communication, peer conflict, teacher-student adaptation; Behavioral adaptation dimensions: delinquent behavior, aggressive behavior, and internet addiction tendencies; The six core dimensions of the MHRSP and their corresponding clinical behavioral factors are as follows: Learning Disabilities Dimensions: Inattention, procrastination on assignments, reading and writing difficulties, fear of learning, hyperactivity in class, and low learning efficiency; Dimensions of emotional disorders: separation anxiety, fear, irritability, depression, emotional sensitivity, and impulsiveness. Personality flaws: low self-esteem and social isolation, dependency and willfulness, stubbornness and rebellion, timidity and cowardice, egocentrism, and low stress tolerance; Dimensions of social adaptation disorders: social withdrawal, peer rejection, poor group integration, weak rule awareness, slow environmental adaptation, and communication barriers; Moral defects: lying and deceiving, damaging public property, bullying peers, and willful and reckless behavior; Unhealthy habits include: nail biting, picky eating, sleep disorders, and frequent fidgeting. Gesell's five developmental domains and their corresponding clinical behavioral factors are as follows: Adaptive behavior domain: perception and judgment, problem-solving, environmental adaptation, attention allocation; Gross motor behavior domain: limb balance, running and jumping coordination, postural control, and movement strength; Fine motor skills domain: hand-eye coordination, object manipulation, pen holding, and finger dexterity; Language behavior domain: language expression, language comprehension, speech intelligibility, and dialogue response; Personal-social behavior domain: self-care, social interaction, emotion expression, rule compliance; Each of the above clinical behavioral factors is determined using its corresponding discrimination criteria.

[0006] Furthermore, in step S3, the formula for the dual-scale fusion weighting algorithm is: in, The standardized score of the main scale MHRSP; Standardized scores of the Gesell supplementary scale; The algorithm provides age-appropriate norm-matched scores for lower and upper elementary grades; it also outputs independent scores and risk levels for each scale simultaneously, preserving complete clinical reference value.

[0007] Furthermore, in step S4, the age-specific norm adaptation is constructed based on 5,000+ clinical samples of Chinese children aged 6-12. A simplified version of task difficulty and collection indicators is used for the lower grades of primary school, while an advanced version of task difficulty and collection indicators is used for the upper grades of primary school, respectively matching the cognitive level, motor development characteristics and psychological development patterns of the corresponding age groups. For lower elementary grades, the Gesell Developmental Diagnostic Scale (DDDS) universal percentile norms are used, with P50 as the median to classify four levels of risk: normal P50-P84, watchful P16-P49, warning P5-P15, and high warning <P5, which is suitable for the developmental assessment characteristics of young children. For upper elementary grades, the Elementary School Students' Mental Health Assessment Scale (MHRSP) standard T-score norms are used, with T50 as the mean and T10 as the standard deviation to classify four levels of risk: normal T55-T70, watchful T40-T54, warning T30-T39, and high warning <T30, which is suitable for the psychological assessment standards of older children, achieving accurate and dynamic adaptation of age-specific norms. The specific difficulty and data collection indicators for the age-based task are as follows: For lower elementary grades: a single instruction, single step, and single task ≤1 minute was used; the collected indicators were the discrimination criteria of the corresponding clinical behavioral factors; the factor weights focused on emotion and concentration, totaling 25%, and gross motor and fine motor skills, totaling 20%; the remaining 55% weights were distributed equally among secondary factors including but not limited to social, learning, and adaptation factors. Upper elementary school: Multiple instructions and steps, single task ≤1.5min. The data collection indicators are the discrimination conditions of corresponding clinical behavioral factors. The factor weights focus on social and learning, totaling 30%; adaptation and language, totaling 15%; the remaining 55% weights are distributed among factors including but not limited to emotion, behavior, development, personality or habits, using a differentiated allocation.

[0008] Furthermore, in step S5, for children aged 12-16, the program includes, but is not limited to, "Campus Growth Radio Station". Interactive scripts for "Special Guest Host Recruitment" and "Campus Growth Radio Station" The interactive script for "Special Guest Host Recruitment" includes four tasks: script reading, impromptu expression, teamwork, and emergency response, and embeds clinical behavioral factors corresponding to MSSMHS. For ages 6-12, the script will be used, but is not limited to, the script from the "Campus Starlight Botanical Garden: Growth Tree Protection Project". The "Growth Tree Guardian Project" script uses the protection and care of seedlings as a gamified medium throughout, including four tasks: heartwarming watering, scaffold building, seed classification, and protection from wind and rain, and embeds clinical behavioral factors corresponding to dual scales.

[0009] Furthermore, in step S6, the risk level is intelligently determined based on domestic norm thresholds and divided into four levels: high warning, warning, attention, and normal. For different risk levels, corresponding targeted intervention courses and home-school collaborative guidance programs are automatically matched. The intervention plan corresponding to a level 4 risk is as follows: High alert level: One-on-one clinical psychological counseling and family-specific intervention; Warning: Group psychological courses and home-school collaborative guidance; Focus: Classroom psychological literacy courses and daily follow-up; Normal: Growth incentives, routine mental health education.

[0010] Furthermore, in step S6, the multimodal data acquisition includes the user's 0.5cm-level limb micro-movements and 108 facial micro-expression key points; The MSSMHS single-scale score calculation model is used for adolescents aged 12-16, and the steps are as follows: The algorithm model converts the collected unstructured behavioral data into continuous quantified values ​​of 0–4 points according to a preset mapping rule; the single factor score is the arithmetic mean of the quantified values ​​of the corresponding collected indicators. The formula for calculating the clinical behavioral factor score is as follows: in, For the first Item factor score; For the first Quantitative values ​​of the collected indicators; The number of indicators bound to this factor; The dimension scores of the four core dimensions of MSSMHS are the weighted sum of all factor scores under each dimension according to their original weights. MSSMHS raw total score calculation: in, to For the dimension scores, the original summation rules of the scale were strictly followed without adding any weights; then, according to the official MSSMHS norms, the raw scores were converted into standard T-scores with a mean of 50 and a standard deviation of 10. in, The mean of the norm. The T-score, calculated using the norm standard deviation, is used to determine the risk level. The algorithm automatically outputs the risk level based on the T-score norm threshold: Normal: T55–T70; Watch: T40–T54; Warning: T30–T39; High Warning: <T30. The MHRSP+Gesell dual-scale fusion score calculation model was used for children aged 6–12 years. The steps are as follows: Calculation of the main scale MHRSP score: in, The quantitative value of the indicator ranges from 0 to 4 points; The number of indicators; the scores for the six dimensions are equal to the arithmetic mean of the corresponding factor scores under each dimension: in, , , , , , In These are the serial numbers used to distinguish different factors within the same scale dimension; Convert raw scores for lower elementary grades to percentile scores. Convert raw scores for upper elementary grades to T-scores (standardized scores). ; The Gesell score calculation for the supplementary scale uses factor scores equal to the mean of the corresponding behavioral indicator quantification values, and the scores for the five domains are equal to the arithmetic mean of the factor scores within each domain. The formula is consistent with the formula used to calculate the MHRSP score for the main scale. Then, the raw scores for lower elementary grades are converted to percentile standard scores. Convert raw scores for upper elementary grades to T-scores (standardized scores). ; Age-specific norm fit score calculation: Matching score = deviation rate between actual score and mean of age-matching norm. The smaller the deviation, the higher the fit score; the larger the deviation, the lower the fit score. Calculate the final overall score: Lower elementary grades: Upper elementary school: The algorithm automatically matches the age-classification norm threshold to output the level: Lower elementary grades: Normal: P50–P84; Attention: P16–P49; Warning: P5–P15; High warning: <P5; Upper elementary grades: Normal: T55–T70; Attention: T40–T54; Warning: T30–T39; High warning: <T30.

[0011] Furthermore, the following steps are also included: S7. Privacy Data Processing: Data involving user privacy is encrypted using the national cryptographic SM4 symmetric encryption algorithm in segments. Core privacy data is encrypted separately, while basic data is encrypted in groups. Core privacy data includes the overall assessment score, factor scores for each dimension, risk level, clinical behavioral characteristics, and targeted intervention plan. Basic data includes age, grade, assessment time, and interactive behavior data. After encryption, only the ciphertext is stored, and plaintext data is not stored on the ground, cached, or transmitted to the cloud. When a user undergoes their first assessment or registration, dynamic facial recognition liveness detection is initiated, instructing the user to complete three standard liveness actions: blinking, opening their mouth, and turning their head horizontally. During the data acquisition process, complete facial image pixel data is discarded, and only the following is extracted: Coordinates of 108 sparse facial feature points, including but not limited to geometric key points of the eye socket, bridge of the nose, corner of the mouth, and cheek; The temporal motion trajectories of three sets of live body movements, namely, the displacement, velocity, and angle changes of feature points during the movement process; Both feature points and motion trajectories are non-image-based geometric data, lacking essential information for face reconstruction such as texture, contour, and skin color. The coordinates of 108 feature points and dynamic motion trajectory data are merged, and a unique, irreversible, and collision-free identity hash index is generated using the SHA-256 one-way hash algorithm. This hash index serves as the unique matching identifier for the encrypted privacy data and is bound one-to-one with the encrypted privacy data. The system does not store the original feature points or the original motion trajectory, but only retains the hash index, completely severing the association between the face image and the privacy data. When a user requests to view their private data, the system restarts dynamic liveness detection, instructing the user to perform the same three sets of actions. At this time, the system extracts the current feature point coordinates and movement trajectory in real time, simultaneously generating a temporary hash value. This temporary hash value is then precisely compared with the stored identity hash index. Matching results: The user is confirmed to have performed the operation; proceed to the decryption process. Inconsistent comparison: If the data is determined to be from someone else, data access will be immediately rejected, and a security alert will be triggered. After verification that the user is the actual person, only the encrypted private data of the current user will be decrypted.

[0012] Furthermore, in step S7, the feature points, motion trajectories, and hash indexes are all one-way irreversible data, making it impossible to reverse-engineer the original feature information, let alone reconstruct a complete face image. Even if the hash index and encrypted ciphertext are illegally stolen, the attacker can only obtain discrete geometric data and ciphertext, and cannot deduce the face image from the feature points / trajectories, nor can they associate it with the identity of a specific minor.

[0013] This invention provides a dual-scale fusion method for age-appropriate dual-track psychological assessment of adolescents, which has the following beneficial effects: 1. This dual-scale fusion method for age-appropriate dual-track psychological assessment of adolescents constructs an age-appropriate dual-track assessment system for adolescents aged 6-16. It adapts single-scale and dual-scale fusion assessment models with 12 years old as the dividing line. Combining standardized scale decomposition, fixed-weight fusion algorithms, dynamic adaptation of age-appropriate norms, factor-embedded seamless scripts, and multimodal data calculation, it achieves accurate assessment adaptation across all age groups and full coverage of scale dimensions. This effectively solves the problems of distortion, poor adaptability, and low student participation in traditional assessments, significantly improving the reliability and validity of the assessment and student participation. Simultaneously, it forms a complete closed loop of assessment, early warning, and intervention, fully aligning with domestic educational compliance requirements and the psychological development patterns of adolescents.

[0014] 2. This dual-scale fusion method for age-appropriate dual-track psychological assessment of adolescents employs national cryptographic encryption algorithms to encrypt and store assessment privacy data in segments. It relies on dynamic face recognition to extract sparse facial feature points and action trajectories to generate an irreversible hash index, abandoning deep feature vectors that can be used to reverse-engineer faces. Only the individual can verify and access their own data, fundamentally blocking the association between facial images and privacy data and the risk of malicious reverse engineering. This comprehensively protects the privacy data security of adolescents, complies with the data protection regulations for minors, and completely eliminates the privacy concerns of students and parents. Attached Figure Description

[0015] Figure 1 This is a flowchart illustrating the steps of a dual-scale fusion method for age-appropriate dual-track psychological assessment of adolescents according to the present invention. Detailed Implementation

[0016] The embodiments of the present invention will be described in further detail below with reference to the accompanying drawings and examples. The following examples are for illustrative purposes only and should not be construed as limiting the scope of the invention.

[0017] like Figure 1 As shown, the present invention provides a technical solution: a dual-scale fusion method for age-appropriate dual-track psychological assessment of adolescents, comprising the following steps: S1, Age-Segmented Dual-Track System Anchoring: Using 12 years old as the core dividing point, we construct a dual-track assessment framework for adolescents aged 6-16, a standardized assessment system with a single scale for adolescents aged 12-16 and above, and a dual-scale integrated assessment system for children aged 6-12 and below. S2, Scale Core Disassembly: For the dual-track system, the core of the scales was standardized and decomposed. For adolescents aged 12 and above, the original core dimensions and corresponding clinical behavior factors of the "Middle School Students' Mental Health Scale" were extracted. For children under 12 years old, the core dimensions and clinical behavior factors of the main scale "Primary School Students' Mental Health Assessment Scale" and the auxiliary scale "Gesell Developmental Diagnostic Scale" were extracted respectively, and the dual-dimensional factor library of mental health and developmental level was constructed. The Mental Health Scale for Middle School Students (MSSMHS), the Mental Health Assessment Scale for Primary School Students (MHRSP), and the Developmental Diagnostic Scale (Gesell) are all abbreviated as Gesell. The four core dimensions of MSSMHS and their corresponding clinical behavioral factors are as follows: Dimensions of emotional adaptation: anxiety, depression, impulsivity, and emotional stability; Learning adaptation dimensions: learning anxiety, concentration, procrastination, and learning efficacy; Interpersonal adaptation dimensions: social avoidance, parent-child communication, peer conflict, teacher-student adaptation; Behavioral adaptation dimensions: delinquent behavior, aggressive behavior, and internet addiction tendencies; The six core dimensions of the MHRSP and their corresponding clinical behavioral factors are as follows: Learning Disabilities Dimensions: Inattention, procrastination on assignments, reading and writing difficulties, fear of learning, hyperactivity in class, and low learning efficiency; Dimensions of emotional disorders: separation anxiety, fear, irritability, depression, emotional sensitivity, and impulsiveness. Personality flaws: low self-esteem and social isolation, dependency and willfulness, stubbornness and rebellion, timidity and cowardice, egocentrism, and low stress tolerance; Dimensions of social adaptation disorders: social withdrawal, peer rejection, poor group integration, weak rule awareness, slow environmental adaptation, and communication barriers; Moral defects: lying and deceiving, damaging public property, bullying peers, and willful and reckless behavior; Unhealthy habits include: nail biting, picky eating, sleep disorders, and frequent fidgeting. Gesell's five developmental domains and their corresponding clinical behavioral factors are as follows: Adaptive behavior domain: perception and judgment, problem-solving, environmental adaptation, attention allocation; Gross motor behavior domain: limb balance, running and jumping coordination, postural control, and movement strength; Fine motor skills domain: hand-eye coordination, object manipulation, pen holding, and finger dexterity; Language behavior domain: language expression, language comprehension, speech intelligibility, and dialogue response; Personal-social behavior domain: self-care, social interaction, emotion expression, rule compliance; The above-mentioned clinical behavioral factors are determined by their corresponding discrimination criteria; S3, Construction of a dual-scale fusion weighted algorithm: For children under 12 years old, a dual-scale fusion weighting algorithm with fixed weights was constructed, with the main scale weight set at 60%, the auxiliary scale weight at 30%, and the age-specific norm adaptation weight at 10%, so as to achieve simultaneous assessment of the two scales and calculation of the comprehensive score in one data collection. The formula for the weighted algorithm for fusion of two scales is: in, The standardized score of the main scale MHRSP; Standardized scores of the Gesell supplementary scale; The algorithm provides age-appropriate norm-fit scores for lower and upper elementary grades; it simultaneously outputs independent scores and risk levels for each scale, preserving complete clinical reference value. S4, Age-Specific Norm Dynamic Adaptation: For children under 12 years old, the curriculum is further divided into two sub-levels: lower grades 1-3 and upper grades 4-6, with the third grade of primary school as the dividing line. Different norm thresholds, task difficulty, collection indicators and factor weights are matched for each sub-level to achieve dynamic adaptation of age-specific norms. Age-specific norms were constructed based on clinical samples of over 5,000 Chinese children aged 6-12. Simplified task difficulty and data collection indicators were used for lower elementary grades, while advanced task difficulty and data collection indicators were used for upper elementary grades, respectively matching the cognitive level, motor development characteristics, and psychological development patterns of the corresponding age groups. For lower elementary grades, the Gesell Developmental Diagnostic Scale (DDDS) universal percentile norms are used, with P50 as the median to classify four levels of risk: normal P50-P84, watchful P16-P49, warning P5-P15, and high warning <P5, which is suitable for the developmental assessment characteristics of young children. For upper elementary grades, the Elementary School Students' Mental Health Assessment Scale (MHRSP) standard T-score norms are used, with T50 as the mean and T10 as the standard deviation to classify four levels of risk: normal T55-T70, watchful T40-T54, warning T30-T39, and high warning <T30, which is suitable for the psychological assessment standards of older children, achieving accurate and dynamic adaptation of age-specific norms. The specific difficulty and data collection indicators for the age-based task are as follows: For lower elementary grades: a single instruction, single step, and single task ≤1 minute was used; the collected indicators were the discrimination criteria of the corresponding clinical behavioral factors; the factor weights focused on emotion and concentration, totaling 25%, and gross motor and fine motor skills, totaling 20%; the remaining 55% weights were distributed equally among secondary factors including but not limited to social, learning, and adaptation factors. Upper elementary school: multiple instructions and steps, single task ≤1.5min, the collection indicators are the discrimination conditions of corresponding clinical behavioral factors, the factor weight focuses on social and learning, totaling 30%; adaptation and language, totaling 15%; the remaining 55% weight is distributed by factors including but not limited to emotion, behavior, development, personality or habit, and a differentiated allocation is adopted. S5, Evaluation Factor Fully Embedded Script: Based on the age-based dual-track clinical behavior factor library, a single-scene fixed main line and a 5-minute continuous closed loop of non-sensory assessment interaction scripts are designed for each age group. The collection indicators corresponding to each clinical behavior factor are 100% embedded in the storyline tasks of the script. There are no assessment, answering questions, or sensitive expressions for detection throughout the process, realizing that the interactive process is also the assessment process. For children aged 12-16, the program includes, but is not limited to, "Campus Growth Radio Station". Interactive scripts for "Special Guest Host Recruitment" and "Campus Growth Radio Station" The interactive script for "Special Guest Host Recruitment" includes four tasks: script reading, impromptu expression, teamwork, and emergency response, and embeds clinical behavioral factors corresponding to MSSMHS. For ages 6-12, the script will be used, but is not limited to, the script from the "Campus Starlight Botanical Garden: Growth Tree Protection Project". The "Growth Tree Guardian Project" script uses the protection and care of seedlings as a gamified medium throughout, including four tasks: heartwarming watering, scaffold building, seed classification, and protection from wind and rain, and embeds clinical behavioral factors corresponding to dual scales. S6. Seamless acquisition and computation of multimodal data: The script guides users to complete immersive somatosensory interaction, simultaneously collects users' multimodal behavioral data, and, based on the algorithm model corresponding to the age-based dual-track system, completes the standardized score calculation of the scale, intelligently determines the risk level, and outputs professional assessment reports and intervention plans for the corresponding age group. The risk level is intelligently determined based on domestic norm thresholds, and is divided into four levels: high warning, warning, attention, and normal. The system automatically matches corresponding targeted intervention courses and home-school collaborative guidance programs for different risk levels. The intervention plan corresponding to a level 4 risk is as follows: High alert level: One-on-one clinical psychological counseling and family-specific intervention; Warning: Group psychological courses and home-school collaborative guidance; Focus: Classroom psychological literacy courses and daily follow-up; Normal: Growth motivation, routine mental health education; Multimodal data acquisition includes user micro-movements at the 0.5cm level and 108 key points of facial micro-expressions; The MSSMHS single-scale score calculation model is used for adolescents aged 12-16, and the steps are as follows: The algorithm model converts the collected unstructured behavioral data into continuous quantified values ​​of 0–4 points according to a preset mapping rule; the single factor score is the arithmetic mean of the quantified values ​​of the corresponding collected indicators. The formula for calculating the clinical behavioral factor score is as follows: in, For the first Item factor score; For the first Quantitative values ​​of the collected indicators; The number of indicators bound to this factor; The dimension scores of the four core dimensions of MSSMHS are the weighted sum of all factor scores under each dimension according to their original weights. MSSMHS raw total score calculation: in, to For the dimension scores, the original summation rules of the scale were strictly followed without adding any weights; then, according to the official MSSMHS norms, the raw scores were converted into standard T-scores with a mean of 50 and a standard deviation of 10. in, The mean of the norm. The T-score, calculated using the norm standard deviation, is used to determine the risk level. The algorithm automatically outputs the risk level based on the T-score norm threshold: Normal: T55–T70; Watch: T40–T54; Warning: T30–T39; High Warning: <T30. The MHRSP+Gesell dual-scale fusion score calculation model was used for children aged 6–12 years. The steps are as follows: Calculation of the main scale MHRSP score: in, The quantitative value of the indicator ranges from 0 to 4 points; The number of indicators; the scores for the six dimensions are equal to the arithmetic mean of the corresponding factor scores under each dimension: in, , , , , , In These are the serial numbers used to distinguish different factors within the same scale dimension; Convert raw scores for lower elementary grades to percentile scores. Convert raw scores for upper elementary grades to T-scores (standardized scores). ; The Gesell score calculation for the supplementary scale uses factor scores equal to the mean of the corresponding behavioral indicator quantification values, and the scores for the five domains are equal to the arithmetic mean of the factor scores within each domain. The formula is consistent with the formula used to calculate the MHRSP score for the main scale. Then, the raw scores for lower elementary grades are converted to percentile standard scores. Convert raw scores for upper elementary grades to T-scores (standardized scores). ; Age-specific norm fit score calculation: Matching score = deviation rate between actual score and mean of age-matching norm. The smaller the deviation, the higher the fit score; the larger the deviation, the lower the fit score. Calculate the final overall score: Lower elementary grades: Upper elementary school: The algorithm automatically matches the age-classification norm threshold to output the level: Lower elementary grades: Normal: P50–P84; Attention: P16–P49; Warning: P5–P15; High warning: <P5; Upper elementary grades: Normal: T55–T70; Attention: T40–T54; Warning: T30–T39; High warning: <T30; Example 1: Implementation of a dual-scale integrated assessment method for children under 12 years old This embodiment targets primary school students aged 6-12, and fully implements the dual-scale integrated assessment method of this invention. The specific steps are as follows: Age-based dual-track system anchors and matches users: Users complete facial recognition through a vertical integrated terminal. The system simultaneously retrieves student registration information, double-verifies the age to ≤12 years old, and automatically matches the dual-scale integrated assessment system for children under 12 years old. It further reads grade information: grades 1-3 match the parameters for lower elementary grades, and grades 4-6 match the parameters for upper elementary grades, completing the automatic binding of norms, tasks, indicators, and weights. Standardized decomposition of scale kernel and construction of factor library: The system invokes the scale standardization and decomposition module to complete the full-dimensional decomposition of the dual-scale kernel: The main scale MHRSP breaks down learning disabilities, emotional disorders, personality defects, social adaptation disorders, moral defects, and bad habits into six dimensions, and extracts clinical behavioral factors such as inattention, separation anxiety, low self-esteem and social withdrawal. Gesell Supplemental Scale: Deconstructs five major domains: adaptive behavior, gross motor skills, fine motor skills, language, and personal-social skills, and extracts clinical behavioral factors such as perception and judgment, balance, hand-eye coordination, and language expression. A two-dimensional factor library is generated by merging the data, and two multimodal data collection indicators are bound to each factor to form a standardized mapping rule library.

[0018] Initialization of the dual-scale fusion weighting algorithm: The system loads the fixed-weight algorithm formula and simultaneously retrieves the corresponding age-specific norm thresholds, indicator standard values, and expert calibration coefficients; it first performs a standardized transformation on the MHRSP and Gesell raw scores to eliminate scale dimension differences and ensure that the calculation results are scientifically comparable. Age-based norm dynamic adaptation execution: Lower elementary school (grades 1-3): Percentile norms are used (normal P50-P84, attentive P16-P49, warning P5-P15, high warning <P5). Tasks are single-instruction, single-step, with a single task duration of ≤1 minute. The collected indicators are simplified to 20 core indicators, with factor weights emphasizing emotion 25%, concentration 25%, and gross / fine motor skills 20%. Upper elementary grades (grades 4-6): T-score norms are used (normal T55-T70, attention T40-T54, warning T30-T39, high warning <T30). Tasks are multi-instruction and multi-step, with a single task duration of ≤1.5min. 82 full-dimensional indicators are used for data collection, with factor weights emphasizing social (30%), learning (30%), and adaptation / language (15%).

[0019] 5-minute contactless assessment script execution and factor full embedding data collection: The system launched "Campus Starlight Botanical Garden" The script for "Growth Tree Guardian Project" incorporates clinical behavioral factors into all four story missions: Heartwarming Nurturing Task: Corresponds to positive emotional expression and emotional stability, and collects facial smile frequency and micro-expression stability; Scaffolding collaborative tasks: corresponding to social initiative, cooperation level, gross motor ability, and data collection of active interaction frequency and movement synchronization. Seed classification task: corresponds to concentration, fine motor control, and data collection of gaze duration and object manipulation accuracy; Storm Protection Mission: Corresponding to environmental adaptation and emotional regulation, collecting data on response speed to environmental changes and the amplitude of emotional fluctuations; The entire process is free of sensitive descriptions, features a soft-light and low-noise design, and allows students to complete tasks through motion-sensing interaction, while the system simultaneously collects multimodal data. Multimodal data calculation and result output: The system collects data through 60 frames of motion capture, extracts features, and then inputs them into a fusion algorithm to calculate the MHRSP, Gesell independent scores, and comprehensive scores to determine the four-level risk level. The student end displays a positive evaluation of "Guardian Ability Star Rating", while the teacher and parent end receive a professional report containing factor scores, weakness analysis, and targeted intervention plans. Example 2: Implementation of a single-scale assessment method for adolescents aged 12 and above This embodiment targets junior and senior high school students aged 12-16, and uses a dual-track system with a single-scale assessment: For facial recognition and student registration verification, the age is ≥12 years old. The system automatically matches the MSSMHS single scale system, breaks down the four dimensions of emotional adaptation, learning adaptation, interpersonal adaptation, and behavioral adaptation, extracts clinical behavioral factors such as anxiety, learning anxiety, social avoidance, and disciplinary behavior, and binds them to the corresponding collection indicators. Launching the "Campus Growth Radio Station" The script for "Special Guest Host Recruitment" incorporates all clinical behavioral factors through four tasks: script reading, impromptu expression, teamwork, and emergency response, with seamless interaction throughout. The MSSMHS-specific weighted algorithm is used to calculate scores, determine four levels of risk, and output standardized reports and intervention plans. It shares the same system architecture, data management, and intervention loop with the children under 12 years old.

[0020] Current facial recognition methods extract 128 or 512-dimensional dense deep feature vectors from facial images and use these vectors as search indexes. Each dimension of the dense deep feature vector corresponds to a global feature of the face (such as contour, facial proportions, texture distribution, and facial muscle direction), which is the "digital fingerprint" of the face. Facial models can be reconstructed through these vectors. However, existing facial recognition reverse engineering techniques (such as GAN, StyleGAN2, and Vec2Face) can accurately reconstruct complete facial images using only facial feature vectors. This means that even if existing solutions are encrypted, as long as the feature vectors are leaked, the face and its associated private data can be reverse-engineered, which can easily lead to the risk of privacy data leakage. To address this problem, the method of the present invention further includes the following steps: S7. Privacy Data Processing: Data involving user privacy is encrypted using the national cryptographic SM4 symmetric encryption algorithm in segments. Core privacy data is encrypted separately, while basic data is encrypted in groups. Core privacy data includes the overall assessment score, factor scores for each dimension, risk level, clinical behavioral characteristics, and targeted intervention plan. Basic data includes age, grade, assessment time, and interactive behavior data. After encryption, only the ciphertext is stored, and plaintext data is not stored on the ground, cached, or transmitted to the cloud. When a user undergoes their first assessment or registration, dynamic facial recognition liveness detection is initiated, instructing the user to complete three standard liveness actions: blinking, opening their mouth, and turning their head horizontally. During the data acquisition process, complete facial image pixel data is discarded, and only the following is extracted: Coordinates of 108 sparse facial feature points, including but not limited to geometric key points of the eye socket, bridge of the nose, corner of the mouth, and cheek; The temporal motion trajectories of three sets of live body movements, namely, the displacement, velocity, and angle changes of feature points during the movement process; Both feature points and motion trajectories are non-image-based geometric data, lacking essential information for face reconstruction such as texture, contour, and skin color. The coordinates of 108 feature points and dynamic motion trajectory data are merged, and a unique, irreversible, and collision-free identity hash index is generated using the SHA-256 one-way hash algorithm. This hash index serves as the unique matching identifier for the encrypted privacy data and is bound one-to-one with the encrypted privacy data. The system does not store the original feature points or the original motion trajectory, but only retains the hash index, completely severing the association between the face image and the privacy data. When a user requests to view their private data, the system restarts dynamic liveness detection, instructing the user to perform the same three sets of actions. At this time, the system extracts the current feature point coordinates and movement trajectory in real time, simultaneously generating a temporary hash value. This temporary hash value is then precisely compared with the stored identity hash index. Matching results: The user is confirmed to have performed the operation; proceed to the decryption process. Inconsistent comparison: If the data is determined to be from someone else, data access will be immediately rejected, and a security alert will be triggered. After verification that the user is the actual person, only the encrypted private data of the current user is decrypted; Feature points, motion trajectories, and hash indices are all one-way irreversible data, making it impossible to reverse-engineer the original feature information, let alone reconstruct a complete face image. Even if the hash index and encrypted ciphertext are illegally stolen, attackers can only obtain discrete geometric data and ciphertext, and cannot deduce the face image from feature points / trajectories, nor can they associate it with the identity of a specific minor.

[0021] In summary, this dual-scale fusion method for age-appropriate dual-track psychological assessment of adolescents constructs an age-appropriate dual-track assessment system for adolescents aged 6-16. It adapts single-scale and dual-scale fusion assessment models with 12 years old as the dividing line. Combining standardized scale decomposition, fixed-weight fusion algorithms, dynamic adaptation of age-appropriate norms, factor-embedded seamless scripts, and multimodal data calculation, it achieves accurate assessment adaptation across all age groups and full coverage of scale dimensions. This effectively solves the problems of distortion, poor adaptability, and low student participation in traditional assessments, significantly improving the reliability and validity of the assessment and student participation. Simultaneously, it forms a complete closed loop of assessment, early warning, and intervention, fully aligning with domestic educational compliance requirements and the psychological development patterns of adolescents. The system employs national cryptographic encryption algorithms to encrypt and store evaluation privacy data in segments. It relies on dynamic face recognition to extract sparse facial feature points and action trajectories to generate an irreversible hash index, abandoning deep feature vectors that can be used to reverse-engineer faces. Only the individual can verify and access their own data, thus fundamentally blocking the association between facial images and privacy data and preventing the risk of malicious reverse engineering. This comprehensively protects the privacy data security of teenagers, complies with the data protection regulations for minors, and completely eliminates the privacy concerns of students and parents.

[0022] The embodiments of the present invention are given for illustrative and descriptive purposes only, and are not intended to be exhaustive or to limit the invention to the forms disclosed. Many modifications and variations will be apparent to those skilled in the art. The embodiments were chosen and described in order to better illustrate the principles and practical application of the invention, and to enable those skilled in the art to understand the invention and to design various embodiments with various modifications suitable for a particular purpose.

Claims

1. A dual-scale fusion method for age-appropriate dual-track psychological assessment of adolescents, characterized by: Includes the following steps: S1, Age-Segmented Dual-Track System Anchoring: Using 12 years old as the core dividing point, we construct a dual-track assessment framework for adolescents aged 6-16, a standardized assessment system with a single scale for adolescents aged 12-16 and above, and a dual-scale integrated assessment system for children aged 6-12 and below. S2, Scale Core Disassembly: For the dual-track system, the core of the scale was standardized and decomposed separately. The original core dimensions and corresponding clinical behavioral factors of the "Middle School Students' Mental Health Scale" were extracted for adolescents aged 12 and above. For children under 12 years old, the core dimensions and clinical behavior factors of the main scale "Primary School Students' Mental Health Assessment Scale" and the auxiliary scale "Gesell Developmental Diagnostic Scale" were extracted to complete the construction of a dual-dimensional factor library for mental health and developmental level. S3, Construction of a dual-scale fusion weighted algorithm: For children under 12 years old, a dual-scale fusion weighting algorithm with fixed weights was constructed, with the main scale weight set at 60%, the auxiliary scale weight at 30%, and the age-specific norm adaptation weight at 10%, so as to achieve simultaneous assessment of the two scales and calculation of the comprehensive score in one data collection. S4, Age-Specific Norm Dynamic Adaptation: For children under 12 years old, the curriculum is further divided into two sub-levels: lower grades 1-3 and upper grades 4-6, with the third grade of primary school as the dividing line. Different norm thresholds, task difficulty, collection indicators and factor weights are matched for each sub-level to achieve dynamic adaptation of age-specific norms. S5, Evaluation Factor Fully Embedded Script: Based on the age-based dual-track clinical behavior factor library, a single-scene fixed main line and a 5-minute continuous closed loop of non-sensory assessment interaction scripts are designed for each age group. The collection indicators corresponding to each clinical behavior factor are 100% embedded in the storyline tasks of the script. There are no assessment, answering questions, or sensitive expressions for detection throughout the process, realizing that the interactive process is also the assessment process. S6. Seamless acquisition and computation of multimodal data: The script guides users through an immersive interactive experience, simultaneously collecting multimodal behavioral data. Based on the algorithm model corresponding to the age-based dual-track system, it completes standardized scale score calculation, intelligent risk level determination, and outputs professional assessment reports and intervention plans for the corresponding age group.

2. The dual-scale fusion method for age-appropriate dual-track psychological assessment of adolescents according to claim 1, characterized in that: In step S2, the "Middle School Students' Mental Health Scale" is abbreviated as MSSMHS, the "Primary School Students' Mental Health Assessment Scale" is abbreviated as MHRSP, and the "Developmental Diagnostic Scale" is abbreviated as Gesell; The four core dimensions of MSSMHS and their corresponding clinical behavioral factors are as follows: Dimensions of emotional adaptation: anxiety, depression, impulsivity, and emotional stability; Learning adaptation dimensions: learning anxiety, concentration, procrastination, and learning efficacy; Interpersonal adaptation dimensions: social avoidance, parent-child communication, peer conflict, teacher-student adaptation; Behavioral adaptation dimensions: delinquent behavior, aggressive behavior, and internet addiction tendencies; The six core dimensions of the MHRSP and their corresponding clinical behavioral factors are as follows: Learning Disabilities Dimensions: Inattention, procrastination on assignments, reading and writing difficulties, fear of learning, hyperactivity in class, and low learning efficiency; Dimensions of emotional disorders: separation anxiety, fear, irritability, depression, emotional sensitivity, and impulsiveness. Personality flaws: low self-esteem and social isolation, dependency and willfulness, stubbornness and rebellion, timidity and cowardice, egocentrism, and low stress tolerance; Dimensions of social adaptation disorders: social withdrawal, peer rejection, poor group integration, weak rule awareness, slow environmental adaptation, and communication barriers; Moral defects: lying and deceiving, damaging public property, bullying peers, and willful and reckless behavior; Unhealthy habits include: nail biting, picky eating, sleep disorders, and frequent fidgeting. Gesell's five developmental domains and their corresponding clinical behavioral factors are as follows: Adaptive behavior domain: perception and judgment, problem-solving, environmental adaptation, attention allocation; Gross motor behavior domain: limb balance, running and jumping coordination, postural control, and movement strength; Fine motor skills domain: hand-eye coordination, object manipulation, pen holding, and finger dexterity; Language behavior domain: language expression, language comprehension, speech intelligibility, and dialogue response; Personal-social behavior domain: self-care, social interaction, emotion expression, rule compliance; Each of the above clinical behavioral factors is determined using its corresponding discrimination criteria.

3. The dual-scale fusion method for age-appropriate dual-track psychological assessment of adolescents according to claim 1, characterized in that: In step S3, the formula for the dual-scale fusion weighting algorithm is: in, The standardized score of the main scale MHRSP; Standardized scores of the Gesell secondary scale; The algorithm provides age-appropriate norm-matched scores for lower and upper elementary grades; it also outputs independent scores and risk levels for each scale simultaneously, preserving complete clinical reference value.

4. The dual-scale fusion method for age-appropriate dual-track psychological assessment of adolescents according to claim 1, characterized in that: In step S4, the age-specific norms are adapted based on clinical samples of more than 5,000 Chinese children aged 6-12. The simplified version of task difficulty and collection indicators is used for the lower grades of primary school, and the advanced version of task difficulty and collection indicators is used for the upper grades of primary school, respectively matching the cognitive level, motor development characteristics and psychological development patterns of the corresponding age groups. For lower elementary grades, the Gesell Developmental Diagnostic Scale general percentile norms are used, with P50 as the median to divide the risk into four levels: normal P50-P84, attention P16-P49, warning P5-P15, and high warning <P5, which is suitable for the developmental assessment characteristics of young children. For upper elementary grades, the MHRSP standard T-score norm is used, with T50 as the mean and T10 as the standard deviation to divide the risk into four levels: normal T55-T70, attention T40-T54, warning T30-T39, and high warning <T30. This is adapted to the psychological assessment standards for older children and achieves accurate and dynamic adaptation of age-specific norms. The specific difficulty and data collection indicators for the age-based task are as follows: For lower elementary grades: a single instruction, single step, and single task ≤1 minute was used; the collected indicators were the discrimination criteria of the corresponding clinical behavioral factors; the factor weights focused on emotion and concentration, totaling 25%, and gross motor and fine motor skills, totaling 20%; the remaining 55% weights were distributed equally among secondary factors including but not limited to social, learning, and adaptation factors. Upper elementary school: Multiple instructions and steps, single task ≤1.5min. The data collection indicators are the discrimination conditions of corresponding clinical behavioral factors. The factor weights focus on social and learning, totaling 30%; adaptation and language, totaling 15%; the remaining 55% weights are distributed among factors including but not limited to emotion, behavior, development, personality or habits, using a differentiated allocation.

5. The dual-scale fusion method for age-appropriate dual-track psychological assessment of adolescents according to claim 1, characterized in that: In step S5, for children aged 12-16, the program includes, but is not limited to, "Campus Growth Broadcasting Station". Interactive scripts for "Special Guest Host Recruitment" and "Campus Growth Radio Station" The interactive script for "Special Guest Host Recruitment" includes four tasks: script reading, impromptu expression, teamwork, and emergency response, and embeds clinical behavioral factors corresponding to MSSMHS. For ages 6-12, the script will be used, but is not limited to, the script from the "Campus Starlight Botanical Garden: Growth Tree Protection Project". The "Growth Tree Guardian Project" script uses the protection and care of seedlings as a gamified medium throughout, including four tasks: heartwarming watering, scaffold building, seed classification, and protection from wind and rain, and embeds clinical behavioral factors corresponding to dual scales.

6. The dual-scale fusion method for age-appropriate dual-track psychological assessment of adolescents according to claim 1, characterized in that: In step S6, the risk level is intelligently determined by benchmarking against domestic norm thresholds and divided into four levels: high warning, warning, attention, and normal. For different risk levels, corresponding targeted intervention courses and home-school collaborative guidance programs are automatically matched. The intervention plan corresponding to a level 4 risk is as follows: High alert level: One-on-one clinical psychological counseling and family-specific intervention; Warning: Group psychological courses and home-school collaborative guidance; Focus: Classroom psychological literacy courses and daily follow-up; Normal: Growth incentives, routine mental health education.

7. The dual-scale fusion method for age-appropriate dual-track psychological assessment of adolescents according to claim 1, characterized in that: In step S6, the multimodal data acquisition includes the user's 0.5cm-level limb micro-movements and 108 facial micro-expression key points; The MSSMHS single-scale score calculation model is used for adolescents aged 12-16, and the steps are as follows: The algorithm model converts the collected unstructured behavioral data into continuous quantified values ​​of 0–4 points according to a preset mapping rule; the single factor score is the arithmetic mean of the quantified values ​​of the corresponding collected indicators. The formula for calculating the clinical behavioral factor score is as follows: in, For the first Item factor score; For the first Quantitative values ​​of the collected indicators; The number of indicators bound to this factor; The dimension scores of the four core dimensions of MSSMHS are the weighted sum of all factor scores under each dimension according to their original weights. MSSMHS raw total score calculation: in, to For the dimension scores, the original summation rules of the scale were strictly followed without adding any weights; then, according to the official MSSMHS norms, the raw scores were converted into standard T-scores with a mean of 50 and a standard deviation of 10. in, The mean of the norm. The T-score, calculated using the norm standard deviation, is used to determine the risk level. The algorithm automatically outputs the risk level based on the T-score norm threshold: Normal: T55–T70; Watch: T40–T54; Warning: T30–T39; High Warning: <T30. The MHRSP+Gesell dual-scale fusion score calculation model was used for children aged 6–12 years. The steps are as follows: Calculation of the main scale MHRSP score: in, The indicator is a quantitative value, ranging from 0 to 4 points; The number of indicators; the scores for the six dimensions are equal to the arithmetic mean of the corresponding factor scores under each dimension: in, , , , , , In These are the serial numbers used to distinguish different factors within the same scale dimension; Convert raw scores for lower elementary grades to percentile scores. Convert raw scores for upper elementary grades to T-scores (standardized scores). ; The Gesell score calculation for the supplementary scale uses factor scores equal to the mean of the corresponding behavioral indicator quantification values, and the scores for the five domains are equal to the arithmetic mean of the factor scores within each domain. The formula is consistent with the formula used to calculate the MHRSP score for the main scale. Then, the raw scores for lower elementary grades are converted to percentile standard scores. Convert raw scores for upper elementary grades to T-scores (standardized scores). ; Age-specific norm fit score calculation: Matching score = deviation rate between actual score and mean of age-matching norm. The smaller the deviation, the higher the fit score; the larger the deviation, the lower the fit score. Calculate the final overall score: Lower elementary grades: Elementary school upper rank: The algorithm automatically matches the age-classification norm threshold to output the level: Lower elementary grades: Normal: P50–P84; Attention: P16–P49; Warning: P5–P15; High warning: <P5; Upper elementary grades: Normal: T55–T70; Attention: T40–T54; Warning: T30–T39; High warning: <T30.

8. The dual-scale fusion method for age-appropriate dual-track psychological assessment of adolescents according to claim 1, characterized in that: It also includes the following steps: S7. Privacy Data Processing: Data involving user privacy is encrypted using the national cryptographic SM4 symmetric encryption algorithm in segments. Core privacy data is encrypted separately, while basic data is encrypted in groups. Core privacy data includes the overall assessment score, factor scores for each dimension, risk level, clinical behavioral characteristics, and targeted intervention plan. Basic data includes age, grade, assessment time, and interactive behavior data. After encryption, only the ciphertext is stored, and plaintext data is not stored on the ground, cached, or transmitted to the cloud. When a user undergoes their first assessment or registration, dynamic facial recognition liveness detection is initiated, instructing the user to complete three standard liveness actions: blinking, opening their mouth, and turning their head horizontally. During the data acquisition process, complete facial image pixel data is discarded, and only the following is extracted: Coordinates of 108 sparse facial feature points, including but not limited to geometric key points of the eye socket, bridge of the nose, corner of the mouth, and cheek; The temporal motion trajectories of three sets of live body movements, namely, the displacement, velocity, and angle changes of feature points during the movement process; Both feature points and motion trajectories are non-image-based geometric data, lacking essential information for face reconstruction such as texture, contour, and skin color. The coordinates of 108 feature points and dynamic motion trajectory data are merged, and a unique, irreversible, and collision-free identity hash index is generated using the SHA-256 one-way hash algorithm. This hash index serves as the unique matching identifier for the encrypted privacy data and is bound one-to-one with the encrypted privacy data. The system does not store the original feature points or the original motion trajectory, but only retains the hash index, completely severing the association between the face image and the privacy data. When a user requests to view their private data, the system restarts dynamic liveness detection, instructing the user to perform the same three sets of actions. At this time, the system extracts the current feature point coordinates and movement trajectory in real time, simultaneously generating a temporary hash value. This temporary hash value is then precisely compared with the stored identity hash index. Matching results: The user is confirmed to have performed the operation; proceed to the decryption process. Inconsistent comparison: If the data is determined to be from someone else, data access will be immediately rejected, and a security alert will be triggered. After verification that the user is the actual person, only the encrypted private data of the current user will be decrypted.

9. The dual-scale fusion method for age-appropriate dual-track psychological assessment of adolescents according to claim 8, characterized in that: In step S7, the feature points, motion trajectories, and hash indexes are all one-way irreversible data, making it impossible to reverse-engineer the original feature information or restore the complete face image. Even if the hash index and encrypted ciphertext are illegally stolen, the attacker can only obtain discrete geometric data and ciphertext, and cannot deduce the face image from the feature points / trajectories, nor can they associate it with the identity of a specific minor.

Images