Campus psychological assessment method and system based on non-contact detection

Abstract

The application discloses a kind of campus psychological evaluation method and system based on non-contact detection, belong to psychological evaluation technical field, comprising: S1.multisource data acquisition;S2.face multiscale time series signal decoupling;S3.physiological and behavioral characteristics construction;S4.individual dynamic baseline modeling;S5.visual dominant multisource fusion evaluation;S6.psychological state inference and anomaly detection.The application has strong innovativeness and practical value in the field of non-contact psychological evaluation, and is suitable for application scenarios such as campus mental health management.

Description

Technical Field

[0001] This invention relates to a campus psychological assessment method and system based on non-contact detection, belonging to the field of psychological assessment technology. Background Technology

[0002] With the increasing prominence of student mental health issues, traditional psychological assessment methods (such as questionnaires and interviews) have significant limitations, making it difficult to achieve continuous, objective, and dynamic monitoring of students' mental states. In recent years, with the development of computer vision and artificial intelligence technologies, methods for identifying emotions or mental states based on non-contact data have gradually become a research hotspot.

[0003] Existing technologies mainly include the following categories:

[0004] Psychological assessment methods based on questionnaires or scales rely on subjective completion and are prone to concealment, bias, and infrequent updates, making it difficult to reflect true and continuous psychological changes.

[0005] Methods based on single-vision facial expression recognition: These methods capture facial expressions through a camera and classify emotions, but they mostly remain at the macroscopic level of facial expressions and lack the ability to depict deep physiological and neural states.

[0006] Physiological signal detection methods based on wearable devices: Data is collected through devices such as heart rate monitors and blood pressure monitors. Although they have a certain degree of accuracy, they are highly invasive, rely on user cooperation, and are difficult to deploy on a large scale.

[0007] Multimodal fusion methods integrate various data such as speech, images, and behavior, but most methods use simple splicing or unified modeling methods, lack a primary and secondary hierarchical mechanism, and have high computational complexity and insufficient stability.

[0008] In summary, the existing technology has at least the following drawbacks:

[0009] Lacking a non-contact physiological signal deep mining mechanism centered on vision, it is difficult to obtain latent features that are highly correlated with psychological state;

[0010] The utilization of facial signals remains at the expression level, failing to separate multi-scale information such as blood flow changes and nerve vibrations;

[0011] The lack of individual difference modeling and the reliance on group threshold judgments lead to insufficient assessment accuracy.

[0012] The multi-source data fusion method is crude and has not established a hierarchical collaborative mechanism between the main source and auxiliary sources;

[0013] There is a lack of dedicated modeling methods for the relationship between behavioral rhythms and psychological states in campus settings.

[0014] Therefore, there is an urgent need for a psychological assessment method that is primarily based on non-contact visual signals and combined with behavioral auxiliary information, in order to achieve more accurate, continuous and scalable monitoring of campus mental health. Summary of the Invention

[0015] The main purpose of this invention is to address the shortcomings of existing technologies by providing a non-contact detection-based method and system for campus psychological assessment.

[0016] The objective of this invention can be achieved by adopting the following technical solution:

[0017] A campus psychological assessment method based on non-contact detection, characterized by comprising:

[0018] S1. Multi-source data acquisition: Non-contact data acquisition method is adopted, and the acquired data includes visual main source data and behavioral auxiliary data;

[0019] S2. Decoupling of multi-scale temporal signals of the face: Temporal analysis is performed on the acquired facial video to decompose it into signal components of different scales, so as to obtain multi-level psychologically relevant features from a single visual input;

[0020] S3. Construction of Physiological and Behavioral Features: Constructing multidimensional feature vectors based on decoupling results;

[0021] S4. Individual dynamic baseline modeling: Construct an individual historical baseline model for each student, model their long-term behavior and physiological characteristics through a sliding time window, and continuously adjust the baseline parameters through an adaptive update mechanism to reflect individual differences;

[0022] S5. Visual-led multi-source fusion assessment: Construct a psychological state assessment model and improve the system's robustness in complex environments by establishing a hierarchical fusion mechanism of main source and auxiliary source;

[0023] S6. Psychological state inference and anomaly detection: Calculate psychological state indicators based on the fusion results. At the same time, based on the multi-timescale analysis mechanism, jointly judge short-term fluctuations and long-term trends. When an abnormal deviation is detected, trigger an early warning.

[0024] Preferably, the main source of visual data is: acquiring video sequences of students' faces through camera equipment set up in classrooms or public spaces, and obtaining: facial nerve vibration images and facial micro-expression information from them;

[0025] Physiological signals based on temporal changes in the facial region, including heart rate, heart rate variability, and hemodynamic features;

[0026] Behavioral auxiliary data: Low-intrusion behavioral characteristics are obtained through campus information systems, including:

[0027] Time patterns for entering and exiting dormitories and teaching buildings;

[0028] Sparse trajectory characteristics of campus paths;

[0029] Classroom participation metrics include head-up rate and interaction frequency.

[0030] Preferably, the signal components include:

[0031] Low-frequency component: used to characterize blood flow changes and extract heart rate and related physiological parameters;

[0032] Mid-frequency component: used to characterize the micro-vibrational features of facial neuromuscular tissue;

[0033] High-level semantic components: used to extract micro-expression features based on action units.

[0034] Preferably, the construction of the multidimensional feature vector includes:

[0035] Physiological characteristics: heart rate, heart rate variability, blood flow fluctuation characteristics;

[0036] Nerve vibration characteristics: Facial micro-vibration spectrum characteristics;

[0037] Facial features: intensity and rate of change of motor units;

[0038] Behavioral characteristics: time rhythm, activity range, and classroom participation.

[0039] Preferably, the data obtained through modeling includes:

[0040] Physiological baseline range;

[0041] baseline for facial expression activity;

[0042] Behavioral rhythm baseline.

[0043] Preferably, the construction of the psychological state assessment model includes:

[0044] Visual features are used as the primary criterion in decision-making;

[0045] Behavioral characteristics serve as auxiliary correction factors;

[0046] The weights of visual and behavioral features are dynamically adjusted based on data quality.

[0047] Preferably, the psychological state indicators include:

[0048] Psychological stress index;

[0049] Emotional stability index;

[0050] Psychological risk level.

[0051] A system for a campus psychological assessment method based on non-contact detection, applicable to any one of the aforementioned campus psychological assessment methods based on non-contact detection, comprising:

[0052] Data acquisition module;

[0053] Signal decoupling module;

[0054] Feature building module;

[0055] Baseline modeling module;

[0056] Integration assessment module;

[0057] Early warning module.

[0058] Preferably, the modules work together through computing devices to achieve continuous assessment of the psychological state of the campus population.

[0059] Beneficial technical effects of the present invention: The campus psychological assessment method and system based on non-contact detection according to the present invention:

[0060] 1. Non-contact and non-perceptible data acquisition: This invention does not require wearing devices or human intervention. It can acquire physiological and psychological information through visual means, which significantly improves the applicability and scalability of the application.

[0061] 2. In-depth utilization of multi-scale facial information: By decoupling facial signals at multiple scales, not only can expression information be obtained, but also blood flow and nerve vibration characteristics can be extracted, thus reflecting psychological state more comprehensively;

[0062] 3. Vision-driven multi-source fusion mechanism: Unlike traditional uniform fusion methods, this invention adopts a main source-auxiliary source hierarchical structure to improve the stability and accuracy of evaluation;

[0063] 4. Adaptive modeling of individual differences: By introducing an individual dynamic baseline model, offset detection is performed with reference to individual historical data, which significantly reduces the false positive rate;

[0064] 5. Behavioral modeling for campus scenarios: By combining data from specific scenarios such as access control, path tracking, and classroom behavior, the evaluation results are made more valuable for practical application.

[0065] 6. Multi-timescale anomaly detection capability: Simultaneously considering short-term fluctuations and long-term trends, enabling earlier and more reliable identification of psychological risks;

[0066] In summary, this invention has strong innovation and practical value in the field of non-contact psychological assessment, and is applicable to application scenarios such as campus mental health management. Attached Figure Description

[0067] Figure 1This is a system overall structure block diagram according to a preferred embodiment of the present invention;

[0068] Figure 2 This is a flowchart of facial multi-scale temporal signal decoupling according to a preferred embodiment of the present invention;

[0069] Figure 3 This is a flowchart of a psychological assessment method according to a preferred embodiment of the present invention. Detailed Implementation

[0070] To enable those skilled in the art to understand the technical solution of the present invention more clearly, the present invention will be further described in detail below with reference to the embodiments and accompanying drawings, but the embodiments of the present invention are not limited thereto.

[0071] like Figures 1-3 As shown, this invention proposes a campus psychological assessment method and system based on non-contact detection, the core of which lies in constructing a psychological state assessment mechanism that is "visual-driven, multi-source fusion, and individual baseline driven".

[0072] 1. Multi-source data acquisition

[0073] This invention employs a non-contact data acquisition method, including primary visual data and secondary behavioral data:

[0074] Visual primary source data: This involves capturing video sequences of students' faces using cameras installed in classrooms or public spaces, and extracting the following data from them:

[0075] Vibrational imaging of facial nerves;

[0076] Facial micro-expression information;

[0077] Physiological signals based on temporal changes in the facial region, including heart rate, heart rate variability, and hemodynamic features.

[0078] (2) Behavioral Support Data: Low-intrusion behavioral characteristics are obtained through the campus information system, including:

[0079] Time patterns for entering and exiting dormitories and teaching buildings;

[0080] Sparse trajectory characteristics of campus paths;

[0081] Classroom participation metrics include head-up rate and interaction frequency.

[0082] The behavioral data mentioned above are presented in the form of statistical characteristics and do not involve specific content or precise location information.

[0083] 2. Decoupling of multi-scale temporal signals in the face

[0084] Temporal analysis was performed on the acquired facial videos, decomposing them into signal components of different scales, including:

[0085] Low-frequency component: used to characterize blood flow changes and extract heart rate and related physiological parameters;

[0086] Mid-frequency component: used to characterize the micro-vibrational features of facial neuromuscular tissue;

[0087] High-level semantic components: used to extract micro-expression features based on action units.

[0088] The above decoupling method enables the acquisition of multi-level psychologically relevant features from a single visual input.

[0089] 3. Construction of Physiological and Behavioral Characteristics

[0090] A multidimensional feature vector is constructed based on the decoupling results, including:

[0091] Physiological characteristics: heart rate, heart rate variability, blood flow fluctuation characteristics;

[0092] Nerve vibration characteristics: Facial micro-vibration spectrum characteristics;

[0093] Facial features: intensity and rate of change of motor units;

[0094] Behavioral characteristics: time rhythm, activity range, and classroom participation.

[0095] 4. Individual dynamic baseline modeling

[0096] For each student, an individual historical baseline model is constructed, and their long-term behavioral and physiological characteristics are modeled using a sliding time window, resulting in:

[0097] Physiological baseline range;

[0098] baseline for facial expression activity;

[0099] Behavioral rhythm baseline.

[0100] The baseline parameters are continuously adjusted through an adaptive update mechanism to reflect individual differences.

[0101] 5. Vision-led multi-source fusion assessment

[0102] Construct a psychological state assessment model, in which:

[0103] Visual features are used as the primary criterion in decision-making;

[0104] Behavioral characteristics serve as auxiliary correction factors;

[0105] The weights of visual and behavioral features are dynamically adjusted based on data quality.

[0106] By establishing a hierarchical fusion mechanism of main source and auxiliary source, the robustness of the system in complex environments can be improved.

[0107] 6. Psychological state inference and anomaly detection

[0108] Psychological state indicators are calculated based on the fusion results, including:

[0109] Psychological stress index;

[0110] Emotional stability index;

[0111] Psychological risk level.

[0112] Meanwhile, based on a multi-timescale analysis mechanism, short-term fluctuations and long-term trends are jointly judged, and an early warning is triggered when an abnormal deviation is detected.

[0113] 7. System Implementation

[0114] The present invention also provides a corresponding system, comprising:

[0115] Data acquisition module;

[0116] Signal decoupling module;

[0117] Feature building module;

[0118] Baseline modeling module;

[0119] Integration assessment module;

[0120] Early warning module.

[0121] The modules work together through computing devices to achieve continuous assessment of the psychological state of the campus population.

[0122] Example 1: System Operation Flow

[0123] like Figure 1 As shown in the figure, this embodiment provides a non-contact psychological assessment system in a campus environment, and its operation process is as follows:

[0124] 1. Facial signal acquisition

[0125] Ordinary video cameras are deployed in common areas of classrooms or dormitories to capture video of students' faces, obtaining a continuous sequence of images. No active student cooperation is required during the data acquisition process, achieving non-contact and seamless data acquisition.

[0126] 2. Decoupling of multi-scale temporal signals in the face (core step)

[0127] like Figure 2 As shown, the acquired video sequence is processed as follows:

[0128] (1) Detect and track the facial area;

[0129] The key lies in extracting the average color signal of the facial ROI. For each frame of the image, the average green channel value is calculated within the facial ROI:

[0130]

[0131] in:

[0132] For the first Each pixel at time grayscale value

[0133] Number of ROI pixels

[0134] (2) Extract the temporal series changes of pixels in the facial region;

[0135] (3) The signal is divided into: by bandpass filtering and time-series decomposition methods.

[0136] Low-frequency component: used to extract blood flow change signals (rPPG);

[0137] right Perform bandpass filtering:

[0138]

[0139] Heart rate calculation:

[0140]

[0141] in This represents the main peak frequency of the spectrum.

[0142] Mid-frequency component: used to extract facial muscle micro-vibrations (nerve vibrations);

[0143] Intermediate frequency band filtering:

[0144]

[0145] And calculate the vibration intensity:

[0146]

[0147] High-level semantic components: used to extract micro-expression action units (AUs).

[0148] Constructing action unit vectors:

[0149]

[0150] And calculate the rate of change in facial expressions:

[0151]

[0152] This results in three types of characteristics:

[0153] Physiological characteristics (heart rate, heart rate variability, blood flow characteristics);

[0154] Nerve vibration characteristics;

[0155] Micro-expression features.

[0156] 3. Physiological parameter estimation

[0157] Based on the aforementioned low-frequency signal, the remote photoplethysmography (rPPG) method is used to estimate:

[0158] Heart rate (HR);

[0159] Heart rate variability (HRV);

[0160] Hemodynamic characteristics (as an indirect representation of blood pressure).

[0161] 4. Acquisition of behavioral auxiliary data

[0162] Obtaining non-privacy-sensitive behavioral data through campus information systems, including:

[0163] Access control records: times of entering and exiting dormitories and teaching buildings;

[0164] Campus route characteristics: activity range, route redundancy (sparse representation);

[0165] Classroom participation: head-up rate, gaze direction, and interaction frequency.

[0166] 5. Individual dynamic baseline modeling

[0167] For each student, a historical data model is established, including:

[0168] Physiological baseline (mean heart rate, fluctuation range);

[0169] Facial baseline (facial activity level);

[0170] Behavioral rhythm (work-rest pattern).

[0171] Continuous updates are achieved through a sliding time window.

[0172] For any feature (e.g., HR, vibration amplitude, etc.):

[0173] Baseline mean and variance:

[0174]

[0175]

[0176] Offset degree (Z-score):

[0177]

[0178] 6. Psychological state assessment (visual-dominant fusion)

[0179] like Figure 3 As shown, the system adopts a layered fusion strategy of "visual primary source + behavioral secondary source":

[0180] Visual features are used as the primary criterion;

[0181] Behavioral characteristics as correction factors;

[0182] The weights are dynamically adjusted based on data quality.

[0183] Main source-secondary source fusion model

[0184] definition:

[0185] Visual main source features:

[0186] Behavioral auxiliary characteristics:

[0187] Confidence weight:

[0188]

[0189] in:

[0190] Visual signal quality (e.g., illumination and occlusion scores)

[0191] Completeness of behavioral data

[0192] Fusion characteristics:

[0193]

[0194] The evaluation output includes:

[0195] Psychological stress index;

[0196] Emotional stability index;

[0197] Psychological risk level.

[0198] Calculation of psychological assessment indicators

[0199] Psychological stress index :

[0200]

[0201] illustrate:

[0202] Increased heart rate → Increased stress

[0203] Abnormal vibration ↑ → Tension ↑

[0204] Reduced facial expression changes → Increased inhibition

[0205] Emotional stability index :

[0206]

[0207] Comprehensive Risk Score :

[0208]

[0209] 7. Anomaly Detection and Early Warning

[0210] The system performs analysis based on multiple time scales:

[0211] Short-term abnormalities: sudden changes in heart rate + facial expression inhibition;

[0212] Long-term trend: Reduced activity + decreased participation.

[0213] When the indicator exceeds the threshold, an alert is triggered and sent to the psychological management system.

[0214] Set threshold:

[0215]

[0216] Contents not described in detail in this specification are existing technologies known to those skilled in the art. Standard parts used in this invention can be purchased commercially, and irregularly shaped parts can be custom-made according to the description and drawings. The specific connection methods for each part all employ conventional methods such as bolts, rivets, and welding, which are already mature technologies. The machinery, parts, and equipment all use conventional models from the prior art, and the circuit connections also employ conventional connection methods from the prior art, which will not be detailed here.

[0217] Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A campus psychological assessment method based on non-contact detection, characterized in that, include: S1. Multi-source data acquisition: Non-contact data acquisition method is adopted, and the acquired data includes visual main source data and behavioral auxiliary data; S2. Decoupling of multi-scale temporal signals of the face: Temporal analysis is performed on the acquired facial video to decompose it into signal components of different scales, so as to obtain multi-level psychologically relevant features from a single visual input; S3. Construction of Physiological and Behavioral Features: Constructing multidimensional feature vectors based on decoupling results; S4. Individual dynamic baseline modeling: Construct an individual historical baseline model for each student, model their long-term behavior and physiological characteristics through a sliding time window, and continuously adjust the baseline parameters through an adaptive update mechanism to reflect individual differences; S5. Visual-led multi-source fusion assessment: Construct a psychological state assessment model and improve the system's robustness in complex environments by establishing a hierarchical fusion mechanism of main source and auxiliary source; S6. Psychological state inference and anomaly detection: Calculate psychological state indicators based on the fusion results. At the same time, based on the multi-timescale analysis mechanism, jointly judge short-term fluctuations and long-term trends. When an abnormal deviation is detected, trigger an early warning.

2. The campus psychological assessment method based on non-contact detection according to claim 1, characterized in that, Visual primary source data: Collecting video sequences of students' faces through camera equipment set up in classrooms or public spaces, and obtaining: facial nerve vibration images and facial micro-expression information from them; Physiological signals based on temporal changes in the facial region, including heart rate, heart rate variability, and hemodynamic features; Behavioral auxiliary data: Low-intrusion behavioral characteristics are obtained through campus information systems, including: Time patterns for entering and exiting dormitories and teaching buildings; Sparse trajectory characteristics of campus paths; Classroom participation metrics include head-up rate and interaction frequency.

3. The campus psychological assessment method based on non-contact detection according to claim 1, characterized in that, The signal components include: Low-frequency component: used to characterize blood flow changes and extract heart rate and related physiological parameters; Mid-frequency component: used to characterize the micro-vibrational features of facial neuromuscular tissue; High-level semantic components: used to extract micro-expression features based on action units.

4. The campus psychological assessment method based on non-contact detection according to claim 1, characterized in that, The constructed multidimensional feature vector includes: Physiological characteristics: heart rate, heart rate variability, blood flow fluctuation characteristics; Nerve vibration characteristics: Facial micro-vibration spectrum characteristics; Facial features: intensity and rate of change of motor units; Behavioral characteristics: time rhythm, activity range, and classroom participation.

5. The campus psychological assessment method based on non-contact detection according to claim 1, characterized in that, The data obtained through modeling include: Physiological baseline range; baseline for facial expression activity; Behavioral rhythm baseline.

6. The campus psychological assessment method based on non-contact detection according to claim 1, characterized in that, The psychological state assessment model is constructed, wherein: Visual features are used as the primary criterion in decision-making; Behavioral characteristics serve as auxiliary correction factors; The weights of visual and behavioral features are dynamically adjusted based on data quality.

7. The campus psychological assessment method based on non-contact detection according to claim 1, characterized in that, The psychological state indicators include: Psychological stress index; Emotional stability index; Psychological risk level.

8. A system for a campus psychological assessment method based on non-contact detection, applicable to the campus psychological assessment method based on non-contact detection as described in any one of claims 1 to 7, characterized in that, include: Data acquisition module; Signal decoupling module; Feature building module; Baseline modeling module; Integration assessment module; Early warning module.

9. The system of a campus psychological assessment method based on non-contact detection according to claim 8, characterized in that, The modules work together through computing devices to achieve continuous assessment of the psychological state of the campus population.

Images