Real-time Detection Method, Device and Equipment for Cognitive State of Attention Glasses

Abstract

This application relates to the technical field of brain-computer interfaces, and particularly to a method, device, and equipment for real-time detection of the cognitive state of attention glasses. The method includes: obtaining multi-channel electroencephalogram data of the attention glasses, obtaining multi-dimensional features corresponding to the multi-channel electroencephalogram data, and generating a detection sequence; performing time series analysis on the detection sequence, extracting change features to construct a cognitive sequence, and generating a state description according to the cognitive sequence; performing trend analysis on the state description, extracting features from the analysis results corresponding to the trend analysis to construct a conversion rule according to the extracted features, and generating a prediction sequence according to the conversion rule; performing threshold division on the prediction sequence, marking state intervals, constructing a criterion sequence, and generating an evaluation result according to the criterion sequence for signal detection to obtain detection information; performing data conversion on the detection information, constructing a display sequence to mark the state type, and generating cognitive state information. To achieve data continuity and signal integrity during the feature extraction process.

Description

Technical Field

[0001] This application relates to the technical field of brain-computer interfaces, and particularly to a method, device, and equipment for real-time detection of the cognitive state of attention glasses. Background Art

[0002] As a portable cognitive state monitoring device, attention glasses have important application values in fields such as education and learning, and work efficiency improvement. Traditional cognitive state detection methods mainly rely on questionnaire assessment or behavioral observation. These methods not only cannot achieve real-time and objective state assessment, but are also easily affected by subjective factors. Although electroencephalogram (EEG) signals contain rich cognitive state information, existing EEG analysis methods have obvious deficiencies in key links such as signal processing, feature extraction, and state recognition. In the signal acquisition and preprocessing stage, the data quality of portable devices is unstable, and traditional methods lack effective noise suppression and quality control mechanisms. In the feature analysis stage, existing methods often use simple time-domain or frequency-domain features, which cannot fully mine the multi-dimensional features reflecting cognitive states in the signals and are also difficult to capture the dynamic features of state changes.

[0003] Currently, cognitive state detection technologies face challenges in terms of real-time performance, accuracy, and reliability. The EEG data collected by portable devices needs to be processed immediately, but the processing efficiency of existing methods in links such as data cache management, feature extraction, and state recognition is low, making it difficult to meet the requirements of real-time monitoring. The cognitive state itself has a high degree of dynamics and individual differences, and fixed detection models are difficult to adapt to such changes, resulting in insufficient accuracy of detection results. At the same time, there are complex interactions between the multi-dimensional features of the cognitive state. Existing methods lack in-depth analysis of this correlation and have not established a systematic state evaluation standard, making it difficult to ensure the reliability of detection results.

[0004] Therefore, there is an urgent need for a method to solve at least one of the above problems. Summary of the Invention

[0005] Embodiments of this application provide a method, device, and equipment for real-time detection of the cognitive state of attention glasses. The method aims to solve the challenges faced by current cognitive state detection technologies in terms of real-time performance, accuracy, and reliability. The EEG data collected by portable devices needs to be processed immediately, but the processing efficiency of existing methods in links such as data cache management, feature extraction, and state recognition is low, making it difficult to meet the requirements of real-time monitoring. The cognitive state itself has a high degree of dynamics and individual differences, and fixed detection models are difficult to adapt to such changes, resulting in insufficient accuracy of detection results. At the same time, there are complex interactions between the multi-dimensional features of the cognitive state. Existing methods lack in-depth analysis of this correlation and have not established a systematic state evaluation standard, making it difficult to ensure the reliability of detection results.

[0006] In a first aspect, an embodiment of the present application provides a method for real-time detection of the cognitive state of attention glasses, including:

[0007] Obtain multi-channel electroencephalogram data of the attention glasses, and obtain the corresponding multi-dimensional features of the multi-channel electroencephalogram data; perform time-domain transformation according to the multi-dimensional features to obtain signal components, and obtain the state representation corresponding to the signal components; perform hierarchical processing on the state representation to determine the feature hierarchy; mark cognitive attributes based on the feature hierarchy; construct a state mapping according to the cognitive attributes, and generate a detection sequence according to the state mapping;

[0008] Perform time-series analysis on the detection sequence, extract change features, mark state nodes based on the change features, construct a cognitive sequence using the state nodes, and generate a state description according to the cognitive sequence; perform trend analysis on the state description, extract features from the analysis results corresponding to the trend analysis, so as to construct a conversion rule according to the extracted features, and generate a prediction sequence according to the conversion rule;

[0009] Perform threshold division on the prediction sequence, mark the state interval, perform feature aggregation according to the state interval, construct a criterion sequence using the aggregation result corresponding to the feature aggregation, and generate an evaluation result according to the criterion sequence; perform signal detection on the evaluation result to obtain detection information;

[0010] Perform data conversion on the detection information, construct a display sequence, perform state organization based on the display sequence, and mark the state type according to the organization result of the state organization; generate cognitive state information according to the state type.

[0011] In a second aspect, the present application further provides a device for real-time detection of cognitive state, including:

[0012] A data acquisition unit for obtaining multi-channel electroencephalogram data of the attention glasses, and obtaining the corresponding multi-dimensional features of the multi-channel electroencephalogram data; performing time-domain transformation according to the multi-dimensional features to obtain signal components, and obtaining the state representation corresponding to the signal components; performing hierarchical processing on the state representation to determine the feature hierarchy; marking cognitive attributes based on the feature hierarchy; constructing a state mapping according to the cognitive attributes, and generating a detection sequence according to the state mapping;

[0013] A time-series analysis unit for performing time-series analysis on the detection sequence, extracting change features, marking state nodes based on the change features, constructing a cognitive sequence using the state nodes, and generating a state description according to the cognitive sequence; performing trend analysis on the state description, extracting features from the analysis results corresponding to the trend analysis, so as to construct a conversion rule according to the extracted features, and generate a prediction sequence according to the conversion rule;

[0014] A threshold division unit, configured to perform threshold division on the prediction sequence, mark state intervals, perform feature aggregation according to the state intervals, construct a criterion sequence by using the aggregation result corresponding to the feature aggregation, and generate an evaluation result according to the criterion sequence; perform signal detection on the evaluation result to obtain detection information;

[0015] A data conversion unit, configured to perform data conversion on the detection information, construct a display sequence, perform state organization based on the display sequence, and mark state types according to the organization result of the state organization; generate cognitive state information according to the state types.

[0016] In a third aspect, the present application further provides a computer device, including a processor and a memory, where the memory is used to store a computer program, and when the computer program is executed by the processor, the real-time detection method for the cognitive state of the attention glasses as described in the first aspect is implemented.

[0017] This method realizes the real-time monitoring and classification of the user's cognitive state through the dynamic feature extraction and state sequence modeling of multi-channel electroencephalogram (EEG) signals. The electroencephalogram data is collected through the multi-channel sensors of the attention glasses, and multi-dimensional features (such as time-domain amplitude, frequency-band energy, time-frequency joint features) are extracted. Time-domain transformation (such as filtering, segmentation, Fourier transform) is performed on the electroencephalogram data to generate signal components (such as α-wave and β-wave components). State representations (such as frequency-band energy distribution matrices) are generated according to the signal components, and the feature levels (such as high-frequency / low-frequency feature priorities) are determined through hierarchical processing (such as clustering, principal component analysis). Cognitive attributes (such as concentration, fatigue) are marked, state mappings (such as mapping rules from features to cognitive states) are constructed, and detection sequences (such as state sequences within a time window) are generated. Temporal analysis (such as sliding window statistics, trend fitting) is performed on the detection sequences to extract changing features (such as energy fluctuations, frequency offsets). State nodes (such as slope change points, abnormal fluctuation points) are marked, cognitive sequences (such as state transition chains) are constructed, and state descriptions (such as the transition path of "concentration → fatigue") are generated. A prediction sequence (such as the probability of the cognitive state at a future time point) is generated through trend analysis (such as regression prediction), and state intervals (such as high-concentration intervals) are marked based on threshold division. Criterion sequences (such as classification threshold rules) are generated by aggregating features, and evaluation results (such as "current concentration: 80%") are output. The detection results are converted into a display sequence (such as a visualization chart), and state types (such as "efficient work", "mild distraction") are marked through state organization (such as clustering or rule matching), and finally cognitive state information (such as real-time reminders or data reports) is generated.

[0018] Through multi-dimensional feature extraction and time-series analysis, overcome the dependence on single features in traditional methods and improve the classification accuracy of cognitive states (e.g., the accuracy of concentration detection is increased by more than 20%). Based on the dynamic modeling of state nodes and prediction sequences, achieve robust detection under environmental interference (such as noise) and reduce the misjudgment rate (e.g., the false alarm rate in a fluctuating environment is reduced by 30%). Combine the local computing power of the attention glasses (such as an edge processing chip) to complete real-time signal processing (with a delay < 100 ms) and avoid the cloud transmission bottleneck. Provide intuitive cognitive state monitoring through visual feedback (such as state type marking) and personalized criteria (such as threshold adaptation), which is applicable to multiple scenarios such as education, medical care, and office work. Support multi-channel data fusion (such as the combination of EEG and eye movement data) to provide a basic technical framework for future brain-computer interface (BCI) applications.

[0019] It should be understood that the above general description and the following detailed description are only exemplary and explanatory, and do not limit this application. Brief Description of the Drawings

[0020] Figure 1 It is a schematic flow chart of the method for real-time detection of the cognitive state of the attention glasses shown in the embodiments of this application;

[0021] Figure 2 It is a schematic structural diagram of the device for real-time detection of the cognitive state shown in the embodiments of this application;

[0022] Figure 3 It is a schematic structural diagram of the computer device shown in the embodiments of this application. Detailed Description of the Embodiments

[0023] In the following description, specific details such as specific system structures and technologies are proposed for the purpose of illustration rather than limitation, so as to thoroughly understand the embodiments of this application. However, those skilled in the art should clearly understand that this application can also be implemented in other embodiments without these specific details. In other cases, the detailed descriptions of well-known systems, devices, circuits, and methods are omitted to avoid unnecessary details from interfering with the description of this application.

[0024] It should be understood that when used in the specification of this application and the appended claims, the term "comprising" indicates the presence of the described features, wholes, steps, operations, elements, and / or components, but does not exclude the presence or addition of one or more other features, wholes, steps, operations, elements, components, and / or their combinations.

[0025] It should also be understood that the term "and / or" used in the specification of this application and the appended claims refers to any combination and all possible combinations of one or more of the related listed items, and includes these combinations.

[0026] As used in the specification and the appended claims of this application, the term "if" may be construed, depending on the context, as "when", "once", "in response to determining", or "in response to detecting". Similarly, the phrase "if determined" or "if [the described condition or event] is detected" may be construed, depending on the context, to mean "once determined", "in response to determining", "once [the described condition or event] is detected", or "in response to detecting [the described condition or event]".

[0027] In addition, in the description of the specification and the appended claims of this application, the terms "first", "second", "third", etc. are used only for distinguishing descriptions and cannot be construed as indicating or implying relative importance.

[0028] Reference to "one embodiment" or "some embodiments" or the like described in the specification of this application means that a specific feature, structure, or characteristic described in connection with that embodiment is included in one or more embodiments of this application. Thus, statements such as "in one embodiment", "in some embodiments", "in other some embodiments", "in still other embodiments", etc. that appear in different places in this specification do not necessarily all refer to the same embodiment, but mean "one or more but not all embodiments", unless otherwise specifically emphasized. The terms "comprising", "including", "having", and their variants all mean "including but not limited to", unless otherwise specifically emphasized.

[0029] The technical solutions of the embodiments of this application will be introduced below.

[0030] As a portable cognitive state monitoring device, attention glasses have important application values in fields such as education and learning and work efficiency improvement. Traditional cognitive state detection methods mainly rely on questionnaire assessment or behavioral observation. These methods not only cannot achieve real-time and objective state assessment, but are also easily affected by subjective factors. Although electroencephalogram (EEG) signals contain rich cognitive state information, existing EEG analysis methods have obvious deficiencies in key links such as signal processing, feature extraction, and state recognition. In the signal acquisition and preprocessing stage, the data quality of portable devices is unstable, and traditional methods lack effective noise suppression and quality control mechanisms. In the feature analysis stage, existing methods often use simple time-domain or frequency-domain features, unable to fully mine the multi-dimensional features reflecting cognitive states in the signals, and also difficult to capture the dynamic features of state changes.

[0031] The current cognitive state detection technology faces challenges in real-time performance, accuracy, and reliability. The EEG data collected by portable devices needs to be processed immediately, but the existing methods have low processing efficiency in data caching management, feature extraction, and state recognition, making it difficult to meet the requirements of real-time monitoring. The cognitive state itself has high dynamics and individual differences, and fixed detection models are difficult to adapt to such changes, resulting in insufficient accuracy of the detection results. At the same time, there are complex interactions between the multi-dimensional features of the cognitive state. The existing methods lack in-depth analysis of this correlation and have not established a systematic state evaluation standard, making it difficult to ensure the reliability of the detection results.

[0032] Please refer to Figure 1 , Figure 1 FIG. is a schematic flow chart of a method for real-time detection of cognitive state of an attention glasses provided by an embodiment of the present application. The method for real-time detection of cognitive state of the attention glasses in the embodiment of the present application can be applied to a computer device, and the computer device includes, but is not limited to, devices such as smart phones, laptop computers, tablet computers, desktop computers, physical servers, and cloud servers. As Figure 1 shown, the method for real-time detection of cognitive state of the attention glasses in this embodiment includes steps S101 to S1O4, which are described in detail as follows:

[0033] Step S101, obtain multi-channel EEG data of the attention glasses, and obtain multi-dimensional features corresponding to the multi-channel EEG data; perform time-domain transformation according to the multi-dimensional features to obtain signal components, and obtain state representations corresponding to the signal components; perform hierarchical processing on the state representations to determine feature levels; mark cognitive attributes based on the feature levels; construct a state mapping according to the cognitive attributes, and generate a detection sequence according to the state mapping.

[0034] Specifically, during the process of the attention glasses collecting multi-channel EEG data and establishing a data cache, signal acquisition is performed on the frontal lobe area through dry electrode sensors built in the glasses. The dry electrodes are in close contact with the forehead skin of the human body, the sampling frequency is set to 256 Hz, and the signal amplitude range is ±100 μV. To ensure signal quality, the system monitors the electrode impedance in real time. When the electrode impedance exceeds 50 kΩ, the data of this channel is automatically marked as invalid data.

[0035] In some embodiments, the obtaining of the multi-dimensional features corresponding to the multi-channel EEG data includes: establishing a data cache for the multi-channel EEG data; performing window partitioning based on the data cache, and generating a real-time sequence according to the partitioning result of the window partitioning; performing feature marking on the real-time sequence, and extracting state indicators; constructing a feature set based on the state indicators, and generating multi-dimensional features according to the feature set.

[0036] The collected original signal data is transmitted to the data processing unit in real time through the Bluetooth 5.0 protocol, and a 128-bit encryption method is used during the transmission process to ensure data security.

[0037] Exemplarily, establishing the data cache for the multi-channel EEG data includes: adopting a double-buffer mechanism for cache management to calculate the kurtosis and skewness of the divided multi-channel EEG data to evaluate the data quality.

[0038] After receiving the signal, the data processing unit first performs data integrity verification. After the verification passes, the data is stored in a circular data cache. This data cache is managed in a first-in, first-out (FIFO) manner, and the cache size is set to 60 seconds. The data in the cache is organized by channel respectively and stored with timestamps as indexes. Each data record contains information such as signal amplitude, acquisition time, channel number, and data status flag, etc. To improve data access efficiency, the cache adopts a double-buffer mechanism. While data is being read from the current buffer, newly acquired data is written to the standby buffer to implement the double-buffer mechanism. When the data volume in the buffer reaches the preset threshold of 80%, the system automatically deletes the earliest 10% of the data to ensure sufficient space in the buffer. Through data quality evaluation, including calculating the kurtosis (threshold ±5) and skewness (threshold ±2) of the signal, abnormal data is marked to obtain a dynamic data cache containing the high-quality EEG data of the most recent 60 seconds.

[0039] Based on the valid data in the data cache, that is, the qualified data records (signal amplitude, timestamp, status flag normal) in the most recent 60 seconds for each channel, window division is performed. The sliding window method is adopted for processing. The window length is set to 2 seconds, that is, each window contains 512 sampling points. The overlap rate between adjacent windows is set to 50%, and each time the window slides, it moves forward by 1 second. This division method can ensure the continuity of the data and at the same time ensure the effective capture of transient changes. During the window division process, the system first checks the update status of the data cache to ensure that the data being processed is the latest. For each divided window, the system calculates the signal-to-noise ratio (SNR). When the SNR is lower than 10 dB, the window is marked as a low-quality window. At the same time, the system performs baseline drift detection on the data within the window. By calculating the first-order difference of the signal, the stability of the baseline is evaluated. When the difference value exceeds the preset threshold (2 times the standard deviation), it is marked that there is baseline drift. For the marked windows, the system examines the characteristics of the adjacent windows before and after them, including mean, variance, and spectral characteristics, etc. If the characteristics of the adjacent windows are highly similar (correlation coefficient greater than 0.8), the window is retained and corrected; otherwise, it is marked as an invalid window. The system also analyzes the frequency-domain characteristics of the signal within the window, calculates the power spectral density using the fast Fourier transform (FFT), and detects whether there is obvious power frequency interference (50 Hz) or other abnormal frequency components. For each valid window, the system generates a window descriptor, which contains information such as the start and end times of the window, data quality indicators, and anomaly marks, and obtains a series of data windows with quality marks.

[0040] For each data window passing the quality inspection (SNR > 10 dB, stable baseline, normal spectrum), it is processed sequentially to generate a real-time sequence. The first step of preprocessing is baseline correction. Using the median filtering method, a sliding window of 60 sampling points is used to calculate the local baseline, and then the original signal is subtracted by the baseline to obtain the corrected signal. The second step is band-pass filtering. A fourth-order Butterworth filter is used, and the cut-off frequencies are set to 0.5 Hz - 45 Hz to remove power frequency interference and other high-frequency noises. The phase of the filtered signal needs to be corrected to compensate for the phase delay introduced by the filter. For the preprocessed window data, the system performs normalization using the z-score normalization method, and the calculation formula is Z = (X - μ) / σ, where X is the original value, μ is the mean of the data within the window, and σ is the standard deviation. To ensure the continuity of the real-time sequence, the system uses weighted averaging for data fusion in the overlapping regions of adjacent windows. The weight coefficient uses the cosine function, which can achieve smooth transition. During the process of generating the real-time sequence, the system establishes a multi-level data quality control mechanism: First is the artifact detection based on autocorrelation analysis. When obvious periodic artifacts are detected, wavelet transform is used for local denoising. Second is to ensure data continuity by calculating the correlation coefficient between adjacent windows (the default threshold is 0.6). Finally, the statistical characteristics of the sequence are monitored, including mean drift detection and variance stability analysis. The system also calculates the correlation between multi-channel data to evaluate the spatial consistency of the signal, and finally obtains the real-time sequence after preprocessing and quality control.

[0041] Exemplarily, the feature marking of the real-time sequence and the extraction of status indicators include: extracting the signal amplitude change characteristics of multi-channel EEG data in time-domain analysis, decomposing the signal energy distribution characteristics of multi-channel EEG data in frequency-domain analysis, obtaining the signal dynamic evolution characteristics of multi-channel EEG data in joint time-frequency domain analysis, and extracting nonlinear dynamics characteristics; constructing the status indicators according to the signal amplitude change characteristics, signal energy distribution characteristics, signal dynamic evolution characteristics, and nonlinear dynamics characteristics.

[0042] Feature mark the three types of key data (preprocessed time series, phase-corrected signal data, and spatial correlation metrics) obtained in the above steps. Conduct time-domain feature analysis on the preprocessed sequence, and calculate the statistical features within each fixed time window, including mean, variance, skewness, and kurtosis. For example, in the scenario of detecting high school students' classroom attentiveness, when students are highly concentrated, their EEG signals usually exhibit lower variance (reflecting signal stability) and higher kurtosis (reflecting signal regularity). In a distracted or fatigued state, the variance of the signal will increase significantly, the kurtosis value will decrease, and at the same time, the absolute value of skewness will also increase. The variation patterns of these statistical features provide a reliable basis for real-time cognitive state assessment. To improve the accuracy of statistical features, the system divides the fixed window into several sub-segments according to the local stationarity characteristics of the signal in the above steps. Calculate the zero-crossing rate feature using a sliding window on each sub-segment to characterize the fast-changing characteristics of the signal. This fast change is particularly obvious when the student's state changes, for example, during the transition from concentration to distraction, the zero-crossing rate will show an obvious change pattern. Then, perform Hilbert transform on the phase-corrected signal to extract instantaneous amplitude and phase information. Combine the spatial correlation metrics in the above steps, calculate the waveform consistency between multiple channels, and extract various characteristic wave parameters, including occurrence time, duration, peak value, and amplitude, etc. Conduct morphological analysis on the identified characteristic waves, and calculate steepness, sharpness, and asymmetry metrics. These morphological features can effectively distinguish the EEG characteristic waves in different cognitive states and provide an important basis for subsequent state recognition. Finally, perform multi-scale decomposition on the signal through continuous wavelet transform to obtain the time-frequency energy distribution of different frequency bands. Through these processes, four types of state metrics are obtained: statistical features reflecting the overall distribution of the signal, temporal features describing signal changes, characteristic wave parameters characterizing waveform features, and time-frequency features depicting frequency characteristics.

[0043] Construct a feature set using the four types of output status indicators. For the statistical feature group (including mean, variance, skewness, kurtosis), use a robust normalization method to map it to a unified interval. For the time series feature group (including zero-crossing rate, instantaneous amplitude change rate), combine the time window information for standardization. The characteristic wave parameter group (duration and amplitude characteristics of various brain waves) and the time-frequency feature group (energy distribution of each frequency band) are normalized according to their respective physical meanings. Calculate the correlation coefficient matrix for all the normalized indicators and identify strongly correlated indicator pairs. In the correlation analysis, the system adopts a segmented strategy to ensure the stability of the correlation over different time periods. For the marked strongly correlated indicator pairs, retain the features with strong discrimination ability through signal-to-noise ratio and discrimination ratio analysis. For the time-frequency features, based on the energy distribution features obtained in the first stage, use an improved spectral estimation method for feature extraction. The feature extraction process takes into account the time-varying characteristics of the data and improves the time resolution of the features through an adaptive window technique. Evaluate the significance and stability of each feature through variance analysis and resampling techniques, and adopt a stratified sampling strategy to ensure the representativeness of the evaluation samples. In the resampling process, the system comprehensively considers the balance of time continuity and state distribution to avoid bias in feature selection. Finally, select the features with the best performance to form an optimized feature set.

[0044] Generate a multi-dimensional feature representation based on the optimized feature set selected by segments. This feature set includes statistically significant verified statistics (such as standardized variance, kurtosis), stable time series features (such as zero-crossing rate in a specific frequency band), key waveform parameters (such as the duration ratio of the α band), and important frequency domain features (such as the θ / α energy ratio). First, perform centering and standardization preprocessing on these features to eliminate scale differences. During the standardization process, the interquartile range-based method is used to improve the robustness to outliers. Then, perform principal component analysis for dimensionality reduction. By calculating the covariance matrix of the standardized features, singular value decomposition is used to solve for eigenvalues and eigenvectors. The system adaptively determines the number of principal components based on the inflection point information of the eigenvalue decay curve, achieving effective dimensionality reduction while maintaining information integrity. In the construction of the principal component space, a local structure preservation constraint is introduced to ensure that the dimensionality reduction process does not destroy the essential associations of the features. Cross-validation is used to determine the optimal number of principal components, and various configurations are evaluated through feature reconstruction error. The system designs a hierarchical verification strategy to ensure the reliability of the verification results. After projecting the original features into the principal component space, construct high-order feature combinations. The feature combination process adopts a progressive strategy, gradually constructing from low-order to high-order. At each step, the effectiveness of the combined features is evaluated through the information gain criterion. The kernel function method is used for non-linear mapping, and the non-linear relationships between features are captured through parameter optimization. During the kernel mapping process, the idea of multi-kernel fusion is adopted to improve the flexibility of feature expression. Finally, construct a three-dimensional tensor structure according to the time dimension, space dimension, and feature type dimension, realizing the multi-dimensional organization of features and obtaining the final multi-dimensional feature representation.

[0045] Perform time-domain transformation processing on the multi-dimensional feature representation obtained by combining the above steps. The system makes full use of the time-dimensional feature sequence in the three-dimensional tensor structure (including the changing trends of various statistics over time), the spatial-dimensional features (the spatial distribution features of multi-channel signals), and the feature-type dimension (the combination patterns of different types of features). For the time-dimensional features, wavelet packet transform is used for multi-resolution analysis, and the most suitable wavelet basis function is selected for different feature types. For example, for rapidly changing attention indicators, the Daubechies wavelet basis with good time localization characteristics is selected; for slowly changing fatigue indicators, the Symlet wavelet basis with better frequency localization characteristics is selected. In the spatial dimension, based on the spatial distribution features in the tensor, spatial filtering technology is used to enhance the feature representation of specific brain regions. According to the functional characteristics of different brain regions, corresponding spatial filter banks are designed: for the prefrontal region, the β-band features reflecting executive control are mainly extracted; for the parietal region, the α-band features related to attention are emphasized. Through matrix decomposition technology, the feature-type dimension in the tensor is projected onto different feature subspaces to achieve efficient expression of features. After these time-domain transformation processes, two types of signal components are obtained: one is the multi-resolution components reflecting cognitive features at different time scales, including rapidly changing components (reflecting instantaneous state changes, such as attention fluctuations), moderately changing components (reflecting short-term state trends, such as changes in cognitive load), and slowly changing components (reflecting long-term state evolution, such as fatigue accumulation); the other is the enhanced spatial components reflecting the activities of key brain regions, which contain the enhanced signals of each brain region of interest and their interaction features.

[0046] Perform frequency domain decomposition on the obtained multi-resolution components and enhanced spatial components. The system uses short-time Fourier transforms with different parameter configurations for signal components of each time scale: for rapidly changing components, a shorter analysis window (256 sampling points) is used to ensure time resolution; for moderately changing components, a medium window length (512 sampling points) is adopted to balance time-frequency resolution; for slowly changing components, a longer analysis window (1024 sampling points) is used to obtain a more accurate spectrum estimate. The window length is adaptively adjusted according to the local stationarity of the signal, and the optimal window size is dynamically determined by calculating the local stability index of the signal (LSI = local variance / global variance). For the enhanced spatial components, the system conducts frequency domain analysis from three levels: first, calculate the power spectral density of a single brain region to obtain the energy distribution characteristics of each frequency band; then, analyze the phase synchrony between brain regions to construct a synchronization matrix based on the phase locking value (PLV); finally, establish a functional connection network between brain regions based on coherence analysis. During the network construction process, an adaptive threshold technique is used to remove weak connections, and the threshold is set to the 75th percentile of the connection strength distribution. Through these frequency domain decomposition processes, a complete set of frequency domain features is obtained, including: the segmented power spectral diagrams of signals of each time scale, the phase synchronization matrix between brain regions (reflecting the collaborative activity patterns of different regions), and the characteristics of the dynamic functional connection network (describing the information interaction intensity between brain regions).

[0047] Use the above frequency-domain features to mark the characteristic intervals. Based on the power spectrum distribution obtained by decomposition, the system combines the phase synchronization matrix and functional connectivity features and adopts a multi-dimensional interval partitioning strategy. In the frequency dimension, an improved clustering algorithm is used to automatically determine the personalized frequency band boundaries. First, the main frequency band centers are identified based on the peak distribution of the power spectrum, and then the frequency band boundaries are determined using the gradient change of the spectral density. For each identified frequency band, the system calculates three types of features: energy features based on the power spectrum (including absolute energy, relative energy, and energy ratio), synchronization features based on the phase synchronization matrix (including average synchronization intensity, synchronization stability, and clustering coefficient of the synchronization network), and network features based on functional connectivity (including connection density, node centrality, and modularity). The system uses a multi-feature fusion method to evaluate the cognitive relevance of each frequency band, and calculates the feature significance score S = w1×energy significance + w2×synchronization significance + w3×connection significance, where the weight coefficients are obtained through training with historical data. The frequency bands with significance scores exceeding the threshold are finely divided. For example, the classical α band (8 - 13 Hz) is subdivided into low α (8 - 10 Hz, mainly related to alertness) and high α (10 - 13 Hz, mainly related to cognitive processing) according to functional characteristics. The system also establishes a dynamic evaluation mechanism for the characteristic intervals, and evaluates their stability by calculating the time-variability of the interval features. After these processes, a set of multi-dimensionally marked characteristic intervals are obtained, and each interval includes: frequency range definition, energy distribution characteristics, phase synchronization characteristics, functional connectivity characteristics, state correlation degree, and stability score.

[0048] In some embodiments, the obtaining the state representation corresponding to the signal component includes: performing frequency-domain decomposition on the signal component to mark the characteristic intervals through the frequency-domain decomposition; generating a state representation according to the characteristic intervals.

[0049] Generate a state representation based on the above-mentioned feature intervals with multi-dimensional tags. The system designs a hierarchical state mapping framework to systematically transform the multi-dimensional information of the feature intervals into a state representation. In the time dimension, based on the energy change laws of different feature intervals, a multi-scale description of state evolution is constructed. For feature intervals with rapid changes (such as the β band), the system tracks the instantaneous changes in their energy to identify critical moments of state transitions; for feature intervals with slower changes (such as the θ band), the system focuses on the cumulative effect of their energy to evaluate the long-term trends of the state. In the spatial dimension, the system comprehensively utilizes phase synchronization features and functional connectivity features to construct multi-level spatial state representations: first, analyze the immediate cooperation patterns of different brain regions based on the phase synchronization matrix, then use the functional connectivity network features to evaluate the information interaction efficiency between brain regions, and finally integrate these spatial features into a regional activation intensity map. In the frequency dimension, the system performs weighted combination of the features of each tagged interval, and the weight coefficients are dynamically adjusted according to the state correlation degree of the interval. For example, in attention assessment, the weight of the high α wave interval will adaptively change according to the current task type. The system uses tensor decomposition technology to compress these multi-dimensional features into a compact state representation while retaining the key state information. The finally generated state representation is a multi-dimensional feature vector, which contains multi-scale state evolution features in the time dimension, topological organization features of brain region activities in the spatial dimension, and interaction features between frequency bands in the frequency dimension.

[0050] Perform hierarchical processing on the state representation obtained from the above steps. The system directly constructs a hierarchical structure using the multi-dimensional state representation (including temporal continuity features, spatial distribution features, and frequency features) output from the above steps. In the temporal dimension, the temporal continuity features in the state representation are divided into three levels according to the feature change rate: a fast-changing layer (sampling interval < 1 second, reflecting instantaneous state), a medium-changing layer (sampling interval 1 - 10 seconds, reflecting short-term trends), and a slow-changing layer (sampling interval > 10 seconds, reflecting long-term trends). For each temporal level, the system calculates the statistical moment correlation of the features and constructs an intra-layer feature association network. In the spatial dimension, a three-level spatial hierarchy is established based on the spatial distribution features in the state representation: a local layer (activity features of a single brain region), a regional layer (interaction features of a group of brain regions), and a global layer (overall spatial pattern features). Each spatial level determines the spatial aggregation degree of the features by calculating the spatial autocorrelation index. In the frequency dimension, a frequency band hierarchy is constructed using the frequency features in the state representation: a δ-wave layer (1 - 4 Hz), a θ-wave layer (4 - 8 Hz), an α-wave layer (8 - 13 Hz), and a β-wave layer (13 - 30 Hz). At each frequency band level, the system analyzes the energy distribution and phase coupling features of the frequency components. Calculate the inter-layer correlation degree matrix R for all layer features, where Rij represents the correlation strength between the i-th layer feature and the j-th layer feature. Through this systematic hierarchical analysis, a hierarchical feature structure including three dimensions of time, space, and frequency and with complete inter-dimensional associations is obtained.

[0051] Mark cognitive attributes based on the above hierarchical feature structure. The system first extracts significant cognitive state feature patterns at different levels of each dimension. At each level in the temporal dimension: the fast-changing layer analyzes instantaneous attention fluctuation features and identifies significant fluctuation points by calculating the first-order difference of the feature sequence; the medium-changing layer extracts cognitive load change features and detects load level transitions using a sliding variance window; the slow-changing layer analyzes fatigue accumulation features and determines the fatigue development stage based on trend analysis of long time series. At each level in the spatial dimension: the local layer analyzes the activation pattern of a single brain region, such as the β-wave activity intensity in the prefrontal region; the regional layer analyzes the cooperative pattern of functionally related brain regions, such as the synchronous activity of the fronto-parietal network; the global layer evaluates the organizational features of the overall brain network, such as the clustering coefficient and path length of the network. At each level in the frequency dimension: analyze the energy distribution features of different frequency bands, such as α-wave suppression indicating attention concentration and θ-wave enhancement reflecting cognitive load. The system combines the features at each level to construct a multi-dimensional discriminant model of cognitive state. Based on the inter-layer correlation degree matrix in the hierarchical feature structure, calculate the contribution weights of each feature combination to different cognitive attributes. Through multiple rounds of iterative optimization, finally obtain a feature mapping table including the feature-attribute mapping relationship and hierarchical association strength.

[0052] Construct a state mapping using the above feature mapping table with cognitive property tags. The system designs a multi-layer probabilistic graph model to describe the dynamic evolution law of cognitive states. First, based on the feature-attribute correspondence in the feature mapping table, construct a state node network, where each node represents a specific combination of cognitive states. The connection weights between nodes are determined by two parts: one is the hierarchical association strength in the feature mapping table, and the other is the state transition frequency observed in historical data. For each state node, the system calculates three types of probability distributions: the state maintenance probability (PM), which represents the possibility of the state remaining stable; the state transition probability (PT), which describes the possible paths to other states; and the state recovery probability (PR), which reflects the trend of recovering from an abnormal state. During the network construction process, the system pays special attention to the interaction between cognitive properties. For example, there is an obvious negative correlation between the attention level and cognitive load, and the system captures this correlation through conditional probability modeling. For each possible state transition path, calculate the temporal correlation RT and the attribute correlation RA, and comprehensively obtain the path significance score S = w1×RT + w2×RA. Based on the path significance, the system establishes a priority mechanism for state transitions, giving priority to the transition paths with high significance. Through this systematic modeling, a complete state mapping network including state nodes, transition probabilities, path weights, and significance scores is finally obtained.

[0053] According to the above state mapping network, the system generates a detection sequence. The system makes full use of four key elements in the network: calculates the activation probability of each cognitive state using state nodes, predicts the possibility of state transitions using the transition probabilities between nodes, determines the priority of state transitions by combining path weights, and judges the credibility of transitions through significance scoring. Based on these elements, the system implements a multi-scale state sequence generation mechanism. On a short time scale (2 - 3 seconds), the system mainly uses state nodes and transition probabilities to detect state changes by calculating the activation probability of nodes in real time. When the activation probability of a certain state node exceeds a preset threshold (usually 0.75), the system generates a state transition marker. On a medium time scale (30 seconds), the system combines path weights and significance scoring to analyze the evolution pattern of the state path and predict possible state transition sequences. For example, if it is detected that the attention level follows a gradual path of 'high → medium → low' and this path has a high significance score, the system will generate an early warning of attention decay. On a long time scale (5 - 10 minutes), the system evaluates the overall stability of the state network, quantifies the uncertainty of the state distribution by calculating the network entropy value H = -∑Pi×log(Pi), where Pi is the probability of the i-th state. For each detected state, the system outputs four key indicators simultaneously: state label (the type of the current cognitive state), confidence level (calculated based on the activation probability), change trend (predicted based on path significance), and warning level (evaluated based on the network entropy value). The system also establishes a quality control mechanism for the state sequence to evaluate the reliability of the detection results by calculating the temporal consistency CI and spatial consistency CS of the sequence. The finally generated detection sequence contains complete state evolution information, which can not only accurately reflect the current state but also predict future state change trends.

[0054] Step S102, perform a temporal analysis on the detection sequence, extract change features, mark state nodes based on the change features, construct a cognitive sequence using the state nodes, and generate a state description according to the cognitive sequence; perform a trend analysis on the state description, extract features from the analysis results corresponding to the trend analysis, so as to construct a transition rule according to the extracted features, and generate a prediction sequence according to the transition rule.

[0055] Specifically, perform a timing analysis on the detection sequence obtained in the above steps. The system fully utilizes four types of key information in the detection sequence: status tags, confidence levels, change trends, and warning levels. Establish a time transition matrix for the status tag sequence, and perform weighting in combination with the confidence level to obtain a reliable status transition pattern. Calculate the stability index (the product of the status tag and the confidence level) for each time point to identify the status stable interval. In the change trend analysis, use the difference method to remove the linear trend to obtain the trend characteristics, and perform fluctuation analysis on the remaining sequence to obtain the fluctuation characteristics. Conduct a cumulative effect analysis on the warning sequence and introduce a time decay factor to calculate the warning cumulative index. The system establishes a multi-scale mutation detection mechanism: detect the status tag jump points at the short time scale (5 seconds), analyze the trend inflection points at the medium time scale (30 seconds), and identify the significant changes in the warning level at the long time scale (5 minutes). For example, in the classroom teaching scenario, the system can identify the sudden change in students' attention (short scale), the continuous downward trend (medium scale), and the cumulative fatigue effect (long scale). Through these systematic timing analyses, four groups of feature vectors are finally obtained: trend characteristics (including status transition probability and trend slope), fluctuation characteristics (including fluctuation amplitude and frequency), mutation characteristics (including jump point position and intensity), and periodic characteristics (including period length and phase information).

[0056] In some embodiments, the status nodes include slope change points, abnormal fluctuation points, and fluctuation nodes; the marking of the status nodes based on the change characteristics includes: performing piecewise linear fitting on the trend characteristics corresponding to the change characteristics and identifying the slope change points; detecting the abnormal fluctuation points of the fluctuation characteristics corresponding to the change characteristics by using an adaptive threshold method; calculating the local fluctuation intensity corresponding to the change characteristics and marking the fluctuation nodes exceeding the standard deviation of the historical mean value.

[0057] Based on the above four groups of feature vectors, state nodes are marked. For trend features, significant slope change points are identified through piecewise linear fitting, and when the change rate exceeds the threshold, they are marked as trend nodes. For fluctuation features, the adaptive threshold method is used to detect abnormal fluctuations, and the local fluctuation intensity is calculated. When it exceeds two standard deviations of the historical mean, it is marked as a fluctuation node. For mutation features, the cumulative sum statistic is used to identify mutation points, and when the statistic exceeds the decision threshold, it is marked as a mutation node. For periodic features, the main period is determined through autocorrelation analysis, and the start point and peak-valley points of the period are marked as periodic nodes. For each marked node, three key indicators are calculated: node significance (weighted sum based on slope change, fluctuation intensity, and mutation amplitude), influence duration (determined by analyzing state changes before and after the node), and propagation range (based on connectivity analysis of the state space). In practical applications, such as in the remote teaching scenario, when the system detects that multiple students simultaneously show significant fluctuations in attention, a high-significance fluctuation node will be generated, and its influence range will be evaluated. Through this multi-dimensional node marking, a complete set of state nodes is obtained, and each node contains a timestamp, node type, state features, significance score, duration, and influence range.

[0058] Use the marked set of state nodes to construct a cognitive sequence. The system first establishes a temporal skeleton based on the timestamps of the nodes to determine the basic framework of the sequence. Analyze the dominant change pattern in each time window according to the node type (trend, fluctuation, mutation, period), and calculate the distribution ratio of different types of nodes. Use the state feature data to construct a state transition matrix to describe the jump rules between states. Normalize the significance score of each node as the weight coefficient of state importance, and high-significance nodes obtain greater weights in sequence reconstruction. Evaluate the stability of the state based on the duration information of the node, and the longer the duration of the state, the higher the reliability score in the sequence. Determine the spatial distribution characteristics of the state by analyzing the influence range of the node, and the system performs special processing on nodes with overlapping influence ranges to maintain spatial consistency. Three key factors are considered in sequence construction: temporal continuity (ensuring the rationality of state transitions between adjacent nodes), spatial consistency (evaluating state distribution based on the influence range of nodes), and state stability (analyzing through the duration of nodes). For example, in classroom teaching, the system can construct a sequence reflecting the changes in students' cognitive states based on consecutive attention fluctuation nodes: first, determine the sequence framework according to the timestamps, identify the change pattern by combining the node type, analyze the change rule through state features, highlight important transition points using the significance score, evaluate the state stability based on the duration, and finally consider the influence range to ensure reasonable spatial distribution. Through iterative optimization, a state sequence with the highest reliability score is obtained, including complete temporal relationships, state transition rules, and importance weights.

[0059] In some embodiments, generating a state description according to the cognitive sequence includes: obtaining the time intervals between adjacent states corresponding to the cognitive sequence to generate a basic description; extracting the state transition path rules in the basic description to generate a behavior pattern description; screening out high-weight state transitions in the pattern description by combining importance weights to generate a priority description; dynamically adjusting the detail level of the pattern description according to the application scenario requirements to generate a state description including time series statistics and state evolution processes.

[0060] Generate a state description based on the constructed cognitive state sequence. Adopt a multi-level description generation framework, combining quantitative information and qualitative analysis. At the micro level, the system generates a basic description based on the temporal relationship in the sequence: analyze the time intervals between adjacent states, calculate the frequency distribution of the occurrence of states, and identify key temporal patterns, such as '3 state transitions occurred continuously within 10 minutes'. Generate a behavior pattern description based on the state transition rules: analyze the main path of state transition, such as 'the attention level experienced a gradual change from high to medium to low'; extract the typical sequence of state transition, such as '3-4 attention fluctuations usually occur before fatigue appears'; summarize the periodic rules of state change, such as 'a significant decrease in attention occurs every 20-25 minutes'. Use importance weights to highlight key state changes: high-weight state transitions are preferentially described, such as 'the most significant state change occurred at the 45th minute of the course, and the attention level dropped suddenly'; the weight information is also used to screen and sort the importance of the description content. The system dynamically adjusts the detail level of the description according to different application scenario requirements: for real-time teacher monitoring, generate short prompts for high-weight state changes; for teaching evaluation, generate a detailed report including complete time series analysis and transition rules. The final complete state representation output includes two parts: a quantitative analysis report (including time series statistics, transition probabilities, and weight distributions) and a qualitative description report (including a textual description of the state evolution process, key transition points, and periodic rules).

[0061] Perform trend analysis on the state representation obtained from the above steps. The system processes the information in the quantitative analysis report and the qualitative description report separately. Decompose the statistical index sequences (including attention level, cognitive load, fatigue degree, etc.) in the quantitative analysis report. Use the STL method to decompose each index sequence into three components: the trend component reflects the long-term change trend and is extracted through locally weighted regression; the periodic component represents the cyclic change characteristics and is determined based on the periodic patterns identified in the previous analysis (such as the 45-minute attention fluctuation cycle); the random component contains short-term fluctuation information. At the same time, analyze the state descriptions in the qualitative description report and extract key state transition information. For example, the gradual change trend of "high → medium → low" is converted into a state sequence. For each sequence, calculate the trend intensity (variance of the trend component / total variance) to reflect the significance of the trend, the periodic intensity (variance of the periodic component / total variance) to represent the degree of periodic change, and the noise ratio (variance of the random component / total variance) to measure the intensity of fluctuations. The system also analyzes the interaction between components, calculates the correlation degree between the trend component and the periodic component, and between the periodic component and the random component, and evaluates the independence of each component. In practical applications, such as in the classroom attention monitoring scenario, this decomposition can clearly identify the long-term change trend of students' attention (such as the decline caused by fatigue accumulation), the inherent cycle (such as the natural fluctuation of 45 minutes), and the random perturbation (such as the fluctuation caused by external interference). Through this systematic trend analysis, a complete set of trend characteristics is finally obtained, including the trend functions, periodic patterns, fluctuation characteristics of each index, and the correlation indexes between them.

[0062] Feature extraction is performed on the above-mentioned trend features. The system first processes the trend function, identifies the inflection point positions through piecewise linear fitting, calculates the slope of each segment to represent the change rate, and when the slope change between adjacent segments exceeds the preset threshold, marks this point as a trend turning point. At the same time, the acceleration feature of the trend is calculated to evaluate the change rate. In the periodic pattern analysis, the system extracts the main periodic components through Fourier analysis, and calculates the energy contribution and phase characteristics for each significant period. The system pays special attention to the periods related to cognitive laws, such as the natural fluctuation period of attention (about 45 minutes) and the fatigue accumulation period (about 2 hours). For the fluctuation features, the system calculates the local fluctuation intensity, and marks it as a high-fluctuation interval when the fluctuation intensity exceeds two standard deviations of the global mean; identifies the clustering effect and pattern features of the fluctuations through autocorrelation analysis. The system also uses correlation indicators for in-depth analysis to study the temporal correlation and influence relationship between different features. The system establishes a multi-scale risk assessment mechanism: assesses risks based on the slope change rate at the trend scale, judges risks through the abnormal degree of amplitude at the periodic scale, and determines the risk level based on the intensity accumulation effect at the fluctuation scale. For example, in the remote teaching scenario, when it is detected that the attention trend drops rapidly (high slope), the periodic fluctuation amplitude increases abnormally, and the local fluctuations are frequent, the system will mark it as a high-risk state. Through these feature extractions, a complete feature expression set is obtained, including the key time point sequence, periodic feature vector, fluctuation pattern description, and multi-scale risk ratings.

[0063] Based on the above feature expression set, conversion rules are constructed. The system maps the key time point sequence to the state space, determines the state transition type and specific time corresponding to each time point; at the same time, constructs periodic transition constraints using the periodic feature vector to accurately define the periodic law of state transitions; establishes state fluctuation rules according to the fluctuation pattern description to depict the specific manifestations of state fluctuations; determines the priority level and trigger threshold of the rules based on the multi-scale risk ratings. The system uses the decision tree algorithm to construct the initial rule framework: uses the key time points as splitting nodes, and uses the periodic features and fluctuation features as classification attributes. The form of each rule is condition-transition correspondence, such as "when the high fluctuation lasts for more than 10 minutes and the trend slope is less than the threshold, predict a significant decrease in attention". Calculate the reliability indicators for each rule: trigger probability (rule usage frequency), transition accuracy (rule success rate), time correlation (stability of the rule at different times), and spatial correlation (applicability of the rule in different scenarios). Based on these indicators, confidence scores are assigned to the rules. For complex state transitions, rule sequences are constructed through rule combination operators (AND, OR, THEN). In the teaching scenario, the system can generate combination rules such as "continuous three attention fluctuations and the trend is downward, predict that cognitive fatigue is about to occur". Finally, a complete conversion rule library is obtained, including trigger conditions, transition probabilities, time windows, and reliability scores.

[0064] Generate a prediction sequence based on the conversion rule library. The system monitors the feature stream in real time, matches the current state with the triggering conditions, and calculates the matching degree; once the rule is activated, the system generates possible state sequences according to their conversion probabilities and performs multi-step predictions within the time window; at the same time, the reliability score is used as a weight coefficient to generate multiple weighted evolution paths through the Monte Carlo method. In the sequence integration stage, the system combines multiple prediction paths, and the weight coefficient is determined by the reliability of the rule. The system evaluates the uncertainty of the prediction: quantifies the degree of uncertainty by calculating the entropy value of the prediction distribution, and estimates the confidence interval based on historical errors. In applications, such as classroom concentration prediction, the system can give an early warning of possible attention decline 3-5 minutes in advance, and give the probability estimate and confidence interval of the early warning. The finally generated prediction sequence includes state prediction values, probability distributions, timestamps, and confidence intervals, providing predictive guidance for teaching adjustment.

[0065] In step S103, perform threshold division on the prediction sequence, mark the state interval, perform feature aggregation according to the state interval, construct a criterion sequence using the aggregation result corresponding to the feature aggregation, and generate an evaluation result according to the criterion sequence; perform signal detection on the evaluation result to obtain detection information.

[0066] Specifically, threshold partitioning is performed on the prediction sequence obtained in the above steps. The system processes the state prediction values, probability distributions, timestamps, and confidence intervals in the prediction sequence respectively. First, multi-level thresholds are set based on the state prediction values, and an adaptive strategy is adopted to determine the key segmentation points. The system calculates the local statistical characteristics of the prediction sequence, including mean, variance, and skewness, to establish an initial threshold candidate set; for each candidate threshold, its classification effect is calculated, and the optimal threshold combination is selected. By analyzing the multi-modal characteristics of the probability distribution, the system sets separation thresholds between the peaks, and at the same time introduces an adaptive smoothing strategy to avoid overly frequent fluctuations of the thresholds. Sequence segmentation is performed by combining timestamp information, a temporal correlation matrix is established, and the temporal law of state transitions is analyzed to ensure that the threshold boundaries are aligned with the temporal characteristics. At the same time, the confidence interval is used to adjust the thresholds. When the prediction confidence is between 0.7 and 0.9, the threshold range is appropriately relaxed; when the confidence is below 0.7, the threshold tolerance is significantly expanded to reduce the risk of misjudgment. In the distance education scenario, the system divides the attention level into four levels: highly concentrated (prediction value > 0.8 and confidence interval fluctuation < 0.1), normally concentrated (0.6 - 0.8), attention dispersed (0.4 - 0.6), and attention missing (< 0.4). During the threshold adjustment process, a two-way verification mechanism is introduced: forward verification ensures the applicability of the thresholds to historical data by sliding a time window (typical window size is 30 minutes), and backward verification divides the most recent data into multiple subsets to evaluate the generalization ability of the thresholds to new data. In addition, the system fine-tunes the thresholds every 10 minutes, and the adjustment range does not exceed 5% of the current threshold to maintain the stability of the system. Through this multi-dimensional threshold partitioning strategy, a set of state intervals with grade labels is finally obtained.

[0067] Feature aggregation is performed based on the above-mentioned state intervals with level markings. The system designs a multi-level feature aggregation strategy to integrate and analyze the features of state intervals from two dimensions: time and space. In the time dimension, continuous intervals with the same state level are merged, and an adaptive sliding window is used to calculate the duration and state stability of the merged intervals. The system dynamically adjusts the window size by analyzing the frequency characteristics of state changes, with a typical range between 30 seconds and 5 minutes. A smaller window is used when the state changes rapidly, and a larger window is used when the state is stable. The state duration index is calculated for each merged interval, which not only considers the interval length and prediction confidence but also introduces a state intensity factor to reflect the significance of the state. When consecutive "attention dispersion" intervals are found, the system calculates their cumulative impact degree, assigns a 1.5-fold weight to intervals with a confidence level higher than 0.8, and also considers the temporal position of the intervals. The weight of the later intervals will be appropriately increased to reflect the fatigue accumulation effect. In the space dimension, the system analyzes the distribution density and transition frequency of intervals with different state levels, establishes a state transition matrix, and identifies the critical moments and typical patterns of state transitions. The system also introduces a state evolution graph to describe the transition relationship between states through a graph structure. The weight of the edge represents the transition probability, and the size of the node reflects the stability of the state. For example, in classroom teaching, the system can identify the typical transition pattern of the student group from concentration to fatigue and quantify the significance of this transition by calculating the probability of the state path. Through this multi-dimensional feature aggregation, a set of complete comprehensive features is obtained, including state persistence indicators, state transition indicators, stability indicators, and trend prediction indicators.

[0068] Construct a criterion sequence using the above comprehensive features. The system processes various types of indicators separately: construct a criterion in the time dimension based on the state persistence indicator, divide the state duration into three levels: short-term (within 5 minutes), medium-term (5 - 15 minutes), and long-term (more than 15 minutes), and set different thresholds according to different state types. Establish a dynamic criterion using the state transition indicator, not only calculate the frequency characteristics of state transitions, but also analyze the directionality and intensity of the transitions, and assign higher risk weights to rapid back-and-forth state transitions. Set a fluctuation monitoring criterion according to the stability indicator, establish a multi-level fluctuation warning threshold by calculating the state variance and coefficient of variation within a local time window. Construct a trend evaluation criterion in combination with the trend prediction indicator, and the system uses weighted moving average and trend extrapolation methods to predict the evolution trend of the state. The system also designs a combined criterion based on fuzzy rules, constructs a composite evaluation condition by connecting different characteristic indicators, and introduces temporal constraints to ensure the temporal coherence of the criterion. For example, when the trend prediction indicator shows a continuous downward trend for more than 10 minutes, and the state transition indicator exceeds the preset threshold three times within 5 minutes, the criterion for a significant decrease in attention is triggered. Each criterion contains an evaluation condition, a trigger threshold, a weight coefficient, and a reliability score. The reliability of the criterion is verified and calculated through historical data, and a forgetting factor mechanism is introduced to make the most recent data have higher reference value. Criteria with a score lower than 0.6 will be automatically adjusted or eliminated. In teaching applications, the system can generate specific criteria such as "when the attention level fluctuates more than 3 times within 10 minutes and the trend is downward". By integrating these different types of criteria, a complete criterion sequence is formed.

[0069] Generate the evaluation results according to the above criterion sequence. The system first processes the judgment conditions, trigger thresholds, weight coefficients, and reliability scores in the criterion sequence to construct a multi-level evaluation framework. For state recognition through judgment conditions, the system uses the analytic hierarchy process to calculate the importance of different conditions and dynamically adjusts the weights according to real-time feedback. To determine the warning level using the trigger threshold, the system sets up a three-level warning mechanism, and each level of warning is equipped with corresponding intervention suggestions. For weighted calculation based on the weight coefficient, the weight values are obtained through training with historical data, and a dynamic update mechanism is set. To ensure the accuracy of the evaluation by combining the reliability score, the system introduces confidence weighting, and high-reliability criteria have a greater impact on the final evaluation. For example, when the prediction accuracy of a certain criterion continuously exceeds 90% for more than 30 minutes, its weight will increase by 20%. The system designs an adaptive warning mechanism that not only considers the threshold overstep of a single indicator but also analyzes the coordinated changes of multiple indicators. When the score is lower than the warning threshold, graded warning signals are automatically generated according to the degree and duration of overstep. In the evaluation report, the system not only provides quantitative score results but also gives improvement suggestions in combination with specific trigger criteria, such as "It is detected that there are periodic fluctuations in attention for 20 minutes. It is recommended to take a short break for 3 - 5 minutes every 20 minutes and appropriately adjust the teaching pace." By real-time monitoring the change trend of the evaluation score and combining with the historical data pattern, the system can predict potential state problems 3 - 5 minutes in advance and provide preventive suggestions. The finally generated evaluation results include the overall score, sub-item scores for each dimension, a status diagnosis report, and specific improvement suggestions, providing comprehensive decision-making support for teaching adjustment.

[0070] Perform signal detection on the evaluation results obtained from the above steps. The system processes the overall score, sub-item scores, status diagnosis reports, and improvement suggestions in the evaluation results respectively. First, establish a baseline analysis for the overall score sequence, calculate the local mean and standard deviation through an adaptive sliding window, and the window size is dynamically adjusted within the range of 1 - 5 minutes according to the data fluctuation degree. The system sets three levels of detection thresholds: mild deviation (baseline ±1.5σ), moderate deviation (baseline ±2σ), and severe deviation (baseline ±3σ), and introduces a time-weighting mechanism to make the most recent data have a higher reference value. Adopt a differential detection strategy for sub-item scores. For example, set a more sensitive fluctuation threshold (±1.2σ) for the attention dimension and focus on mutation features, while for the fatigue dimension, use a wider threshold (±2.5σ) but pay more attention to the detection of cumulative effects. The system monitors mutation anomalies at a micro scale of 10 seconds and analyzes trend anomalies at a macro scale of 5 minutes. The system combines the status diagnosis report, establishes a semantic analysis module to extract key status description information, identifies description fragments related to anomalies through natural language processing technology, and establishes a problem-symptom mapping relationship. Conduct a systematic effect tracking for improvement suggestions, set a before-and-after comparison window (5 minutes before the suggestion vs 5 minutes after the suggestion), and record and quantify the degree of status improvement after the implementation of the suggestion. In a teaching scenario, the system can monitor individual anomalies and group anomalies simultaneously. When it detects that more than 30% of the students have the same type of anomaly, it triggers a group anomaly alarm. Through this multi-dimensional and multi-scale signal analysis, detection results including the degree of anomaly, duration, and scope of influence are obtained.

[0071] Classify and label the above detection results containing the degree of abnormality, duration, and scope of influence. The system adopts a multi-level labeling strategy. First, a five-level classification system is established based on the degree of abnormality: normal fluctuation (within ±σ), slight abnormality (±1.5σ), moderate abnormality (±2σ), significant abnormality (±2.5σ), and severe abnormality (±3σ and above). Different visual markings and processing priorities are set for each level of abnormality. Severe abnormality is marked with a red flash and triggers a real-time alarm. In the duration analysis, the system designs a time-series classification algorithm that divides the abnormal patterns into three categories: instantaneous abnormality (lasting < 30 seconds) is identified by calculating the change rate ΔS between adjacent time points. When the change rate exceeds 40% and lasts for no more than 3 sampling points, it is marked as a mutation point; fluctuating abnormality (lasting 30 seconds - 3 minutes) is identified by spectral analysis, calculating the fluctuation frequency and amplitude within a local time window; cumulative abnormality (lasting > 3 minutes) is identified by trend analysis methods, calculating the cumulative deviation index and change slope within the time window. The evaluation of the scope of influence uses a spatial clustering method to calculate the distribution density and diffusion speed of the abnormality in the population, and divides the scope of influence into local influence (< 10% of the population), regional influence (10% - 30% of the population), and global influence (> 30% of the population). The system also establishes a composite marking mechanism that comprehensively considers multiple dimensional characteristics of the abnormality. For example, in the remote teaching scenario, when it is detected that the attention indicators of more than 25% of the students simultaneously show moderate or above abnormalities and last for more than 2 minutes, the system will mark this abnormality as "significant decline in group attention" and automatically upgrade the alarm level. A time-series consistency constraint is introduced during the marking process to prevent misjudgment caused by frequent fluctuations in marking. Through this multi-dimensional and multi-level classification and marking, a marking result set containing the type of abnormality, degree level, and spatial distribution is obtained.

[0072] Anomaly feature recognition is carried out based on the above-mentioned marked result set including anomaly types, severity levels, and spatial distributions. First, the system constructs a hierarchical feature pattern library according to the anomaly types, storing feature templates of various typical anomalies, such as main anomaly patterns like attention fluctuation type, cognitive load type, and fatigue accumulation type. For each anomaly type, the system designs a specific feature extraction algorithm: for attention-related anomalies, it focuses on extracting the energy change of the α band and eye movement pattern features; for cognitive load-related anomalies, it emphasizes analyzing the θ band activity and task switching frequency; for fatigue accumulation-related anomalies, it pays attention to features such as the decrease in β band energy and the prolongation of reaction time. The system uses the severity level information to establish a feature significance evaluation mechanism, assigning a feature weight of 1.5 times to high-level anomalies to ensure that the features of severe anomalies can be preferentially analyzed. Based on the spatial distribution data, the system uses spatial statistical methods to analyze the propagation pattern and aggregation effect of anomalies, and calculates the Moran's spatial correlation coefficient to evaluate the spatial autocorrelation of anomalies. During the feature extraction process, the system adopts a segmented statistical method, setting three time windows: before the anomaly occurs (the first 2 minutes), during the occurrence (during the anomaly duration), and after the occurrence (the next 2 minutes), to analyze the complete process of state changes. For example, for anomalies of sudden attention drop, the system analyzes its triggering conditions (such as the presentation of consecutive high-difficulty knowledge points), development process (such as the step-by-step decline in attention level, usually experiencing 4 - 5 stages), and impact consequences (such as the continuous downturn in learning efficiency, usually lasting 5 - 8 minutes). The system evaluates the significance by calculating the discrimination degree of features in normal samples and abnormal samples (measured by F-score), and features with a significance score exceeding 0.8 are retained as key features. The system also establishes a combined analysis mechanism for features, discovers synergistic effects and redundant information by calculating the mutual information and conditional entropy between features, and constructs a minimum feature set to ensure the interpretation efficiency. In classroom teaching applications, the system can identify typical anomaly patterns such as "group attention fluctuations induced by content difficulty" and extract their key feature combinations. Through this comprehensive feature analysis, a set of feature vectors describing the anomaly patterns is obtained, including temporal features, spatial features, and correlation features.

[0073] In some embodiments, the obtaining of detection information by performing signal detection on the evaluation result includes: classifying and marking the detection result corresponding to the signal detection, identifying anomaly features based on the classification and marking; and generating the detection information according to the anomaly features.

[0074] Generate detection information based on the above abnormal feature vectors that include temporal features, spatial features, and correlation features. The system adopts a structured information generation framework and designs three layers of information templates: the core layer contains the basic description of the abnormality; the analysis layer contains quantitative indicators and spatio-temporal features; the suggestion layer contains intervention measures and expected effects. The system first analyzes the development process of the abnormality using temporal features, identifies key turning points and evolution stages through time-frequency analysis of the feature sequence, and accurately describes the occurrence time, duration, and evolution trend of the abnormality. For example, for abnormal attention fluctuations, the system can accurately describe that "the fluctuations started at the 32nd minute of the course, experienced 3 significant drops, each drop lasting about 45 seconds, and generally showing a step-by-step decreasing trend". The system determines the influence range and distribution pattern of the abnormality through spatial features, uses heat map visualization technology to display the distribution of the abnormality in the group, and calculates the spatial autocorrelation index to evaluate the aggregation effect of the abnormality. For example, the system can detect that "the decrease in attention first appeared in the back row of the classroom, then spread forward at a speed of 0.8 people per minute, and finally affected 78% of the student group". The system identifies the causal relationship and chain effect between abnormalities based on correlation features, constructs an abnormal association network, and determines the root cause and influence path of the abnormality through Bayesian network analysis. During the detection information generation process, the system adjusts the detail level of the description according to the type and severity of the abnormality, and provides more detailed analysis information for high-risk abnormalities. For different application scenarios, the system customizes the information output format: for real-time teacher monitoring, generate a concise abnormality summary and key indicators; for teaching evaluation, provide a detailed report containing a complete analysis and historical comparison. The system also establishes an information update mechanism. When the abnormal state changes, it automatically updates the detection information and marks the change trend. The finally generated detection information includes an abnormality description ("a significant decrease in group attention occurred at the 32nd minute of the course"), quantitative indicators ("influence range 78%, duration 8.5 minutes, severity 4.2 / 5"), development trend ("showing a step-by-step decrease, without a natural recovery trend"), and correlation analysis ("showing a significant correlation with the previous difficult knowledge points, confidence level 0.92"), providing precise guidance for teaching intervention.

[0075] Step S104: Perform data conversion on the detection information, construct a display sequence, organize the states based on the display sequence, and mark the state type according to the organization result of the state organization; generate cognitive state information according to the state type.

[0076] Specifically, data conversion is performed on the detection information obtained in the above steps. Four types of data in the detection information are processed systematically: anomaly description, quantitative metrics, development trends, and correlation analysis results. First, semantic analysis is performed on the anomaly description to extract key event features and time markers, and an event-time mapping relationship is established. Then, dimensionality reduction mapping is performed on the quantitative metrics, and multidimensional scaling technology is used to map high-dimensional features to a visualization space. During the mapping process, the degree of anomaly is converted into display intensity, the scope of influence is mapped to spatial distribution, and temporal correlation is converted into connection strength. The system designs an adaptive normalization scheme to ensure the comparability of different types of metrics. For example, the fluctuation amplitude of the attention level and the cumulative value of fatigue degree are uniformly converted to the display range of 0-100 for intuitive comparison. For the development trend information, the system establishes a temporal mapping rule, converts the trend slope into a direction vector, and maps the trend significance to vector intensity. The system also introduces a trend prediction component to generate short-term predictions through regression methods. For the correlation analysis results, a network structure representation is constructed, where the correlation strength determines the connection weight, and different correlation types are represented by different line types. For example, in a classroom scenario, the system can convert the attention distribution pattern of the student group into an area map with different shades of color, and the propagation path of attention fluctuations is represented as a directional connection. Through this systematic data conversion, a set of standardized display sequences are obtained.

[0077] State organization is performed based on the above standardized display sequences. The system adopts a hierarchical organizational structure to systematically integrate different elements in the display sequences. The system uses the display intensity data in the display sequences to evaluate the significance of the state, sets three levels of attention thresholds; divides the individual, local, and overall levels through spatial distribution information; analyzes the change trend of the state based on the direction vector and vector intensity; and constructs the correlation relationship between states using the network structure representation. Taking the classroom scenario as an example, the spatial dimension is divided into three levels: the individual level (single student), the local level (regions such as the front row, middle row, and back row), and the overall level (the whole class range). For the individual level, the system establishes a state timeline based on the display intensity in the display sequences and records key state nodes, such as attention peaks, trough points, and stable intervals. At the local level, using the spatial distribution information of the display sequences, a density clustering method is used to identify student groups with similar states and calculate the representative features of the groups. For example, when it is found that there is a general phenomenon of inattention among the students in the back row, the system will organize this local aggregation effect separately and calculate its influence intensity. At the overall level, a global view of state evolution is constructed by combining the direction vector and vector intensity of the display sequences, and at the same time, the correlation pattern between states is analyzed using the network structure representation to identify the path and speed of state propagation. Through multi-level state organization, a hierarchical state structure containing individual characteristics, group characteristics, and overall characteristics is obtained.

[0078] Use the above hierarchical state structure including individual characteristics, group characteristics, and overall characteristics to perform state type marking. Based on cognitive science theory, the system establishes a multi-dimensional state classification system in combination with the specific manifestations in the hierarchical structure. At the individual characteristic level, the system establishes a multi-dimensional feature space, including key indicators such as attention level, cognitive load, fatigue degree, and engagement. Based on the combination patterns of these indicators, the cognitive states of students are classified into basic types such as active engagement, normal learning, attention distraction, and cognitive fatigue. The system uses a fuzzy classification method to calculate the membership degree of the state type according to the similarity between the feature vector and the standard template. At the group characteristic level, the system identifies common collective phenomena and constructs a group state dictionary, including patterns such as group attention fluctuations caused by difficulties in understanding knowledge points and overall fatigue caused by continuous high-intensity learning. The system analyzes the group consistency index to evaluate the synchronization degree of the state. Based on the spatial distribution characteristics in the overall characteristics, the system analyzes the propagation effects of different types of states, such as the positive impact of the focused state of students in the front row on the surrounding students. The system calculates the spillover effect and aggregation degree of the state through spatial statistical methods to identify key influencing points. For each state type, the system establishes a detailed feature description template, including quantitative discrimination criteria and qualitative judgment bases. The system also implements a state transition rule library to describe common state transition sequences and triggering conditions. Through this systematic state classification, a complete set of state type markings is obtained.

[0079] Generate cognitive state information according to the above-mentioned state type tag set including individual characteristics, group characteristics, and overall characteristics. The system designs a hierarchical information generation architecture to convert state characteristics in different dimensions into specific cognitive state descriptions. The system converts individual characteristics into state descriptions and suggestions at the personal level, and uses natural language generation technology to select appropriate expression templates according to the state type. At the individual level, a personal state report including the current state type, risk level, and warning information is generated, and the update frequency of the state report is dynamically adjusted according to the state change speed. For example, the system will push a personalized state reminder like "Your attention level has shown a fluctuating downward trend in the past 15 minutes, which is 35% lower than your personal baseline. It is recommended to take a proper rest or adjust your learning strategy" to the user in real time. The system generates a collective state report at the team or class level based on group characteristics, including state distribution statistics and typical group characteristics. In group application scenarios, such as a multi-person collaborative office environment, the system can analyze the state coordination of team members. For example, when it detects that multiple members of a project team are simultaneously inattentive, the system will generate state information like "In the project team (5 people), there has been a general decline in attention level in the past 30 minutes, and the average attention index has dropped to 65%. It is recommended to arrange a short break or switch the task type in time". The system uses overall characteristics to construct a macro trend analysis to identify long-term state patterns and periodic change rules. For the deep work scenario of an individual, the system will also combine the macro trend in the overall characteristics to generate a more targeted state report, such as "Your attention pattern shows an obvious fluctuation every 70 - 80 minutes. It is recommended to adjust your work rhythm". Through this scenario-based and personalized information generation method, a complete set of cognitive state information is finally obtained.

[0080] The method has the following beneficial effects:

[0081] 1. Innovatively adopt a dynamic data caching and multiple window processing mechanism, combined with time-domain transformation and frequency-domain decomposition techniques, to achieve data continuity and signal integrity in the feature extraction process, ensuring the stable performance of the portable device during data acquisition and processing.

[0082] 2. By establishing a hierarchical state representation and a time-series analysis system, combined with a trend prediction and evaluation mechanism, the system can adaptively adjust detection parameters to achieve precise monitoring of different individuals' cognitive states.

[0083] 3. Construct a complete state anomaly detection and information generation process, integrating multi-dimensional feature analysis and state association recognition functions, enabling the system to comprehensively describe and display the changes in the user's cognitive state.

[0084] To execute the real-time cognitive state detection method of the attention glasses corresponding to the above method embodiment to achieve the corresponding functions and technical effects. See Figure 2, Figure 2 The structure block diagram of a cognitive state real-time detection device 200 provided by an embodiment of the present application is shown. For ease of explanation, only the parts related to this embodiment are shown. The cognitive state real-time detection device 200 provided by the embodiment of the present application includes:

[0085] A data acquisition unit 201, configured to acquire multi-channel electroencephalogram data of an attention glasses, acquire multi-dimensional features corresponding to the multi-channel electroencephalogram data; perform time-domain transformation according to the multi-dimensional features to obtain signal components, acquire state representations corresponding to the signal components; perform hierarchical processing on the state representations to determine feature levels; mark cognitive attributes based on the feature levels; construct a state mapping according to the cognitive attributes, and generate a detection sequence according to the state mapping;

[0086] A timing analysis unit 202, configured to perform timing analysis on the detection sequence, extract change features, mark state nodes based on the change features, construct a cognitive sequence using the state nodes, and generate a state description according to the cognitive sequence; perform trend analysis on the state description, extract features from the analysis results corresponding to the trend analysis, so as to construct a conversion rule according to the extracted features, and generate a prediction sequence according to the conversion rule;

[0087] A threshold division unit 203, configured to perform threshold division on the prediction sequence, mark state intervals, perform feature aggregation according to the state intervals, construct a criterion sequence using the aggregation results corresponding to the feature aggregation, and generate an evaluation result according to the criterion sequence; perform signal detection on the evaluation result to obtain detection information;

[0088] A data conversion unit 204, configured to perform data conversion on the detection information, construct a display sequence, perform state organization based on the display sequence, mark state types according to the organization results of the state organization; generate cognitive state information according to the state types.

[0089] The above-mentioned cognitive state real-time detection device 200 can implement the cognitive state real-time detection method of the attention glasses in the above method embodiment. The optional items in the above method embodiment are also applicable to this embodiment, which will not be elaborated here. The remaining content of the embodiment of the present application can refer to the content of the above method embodiment, and will not be repeated in this embodiment.

[0090] Figure 3 The structure schematic diagram of a computer device provided by an embodiment of the present application. As Figure 3 shown, the computer device 3 in this embodiment includes: at least one processor 30 ( Figure 3Only one is shown in the figure), a memory 31, and a computer program 32 stored in the memory 31 and executable on the at least one processor 30. When the processor 30 executes the computer program 32, the steps in any of the above method embodiments are implemented.

[0091] The computer device 3 may be a computing device such as a smart phone, a tablet computer, a desktop computer, and a cloud server. The computer device may include, but is not limited to, a processor 30 and a memory 31. Those skilled in the art can understand that Figure 3 This is only an example of the computer device 3 and does not limit the computer device 3. It may include more or fewer components than shown in the figure, or combine certain components, or different components. For example, it may also include input / output devices, network access devices, etc.

[0092] The so-called processor 30 may be a central processing unit (CPU). The processor 30 may also be other general-purpose processors, digital signal processors (DSPs), application specific integrated circuits (ASICs), off-the-shelf programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc.

[0093] In some embodiments, the memory 31 may be an internal storage unit of the computer device 3, such as the hard disk or memory of the computer device 3. In other embodiments, the memory 31 may also be an external storage device of the computer device 3, such as a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card, etc. equipped on the computer device 3. Further, the memory 31 may also include both the internal storage unit and the external storage device of the computer device 3. The memory 31 is used to store an operating system, application programs, a boot loader, data, and other programs, such as the program code of the computer program. The memory 31 may also be used to temporarily store data that has been output or will be output.

[0094] In addition, an embodiment of the present application further provides a computer-readable storage medium storing a computer program, and when the computer program is executed by a processor, the steps in any of the above method embodiments are implemented.

[0095] An embodiment of the present application provides a computer program product. When the computer program product runs on a computer device, the computer device is caused to implement the steps in each of the above method embodiments when executed.

[0096] In several embodiments provided by the present application, it can be understood that each block in the flowchart or block diagram may represent a module, a program segment, or a part of code, and the module, the program segment, or the part of code includes one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions marked in the blocks may occur in a different order from that marked in the drawings. For example, two consecutive blocks may actually be executed substantially in parallel, and they may sometimes be executed in the reverse order, depending on the functions involved.

[0097] If the above functions are implemented in the form of software function modules and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on such an understanding, the technical solution of the present application, in essence, or the part that contributes to the prior art, or a part of this technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for causing a computer device to execute all or part of the steps of the methods described in the various embodiments of the present application. The foregoing storage medium includes: various media such as USB flash drives, mobile hard disks, read-only memories (ROM, Read-Only Memory), random access memories (RAM, Random Access Memory), magnetic disks, or optical discs that can store program codes.

[0098] The above specific embodiments have further elaborated in detail the purpose, technical solution, and beneficial effects of the present application. It should be understood that the above are only specific embodiments of the present application and are not used to limit the protection scope of the present application. It is particularly pointed out that for those skilled in the art, any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.

Claims

1. A real-time detection method for the cognitive state of an attention glasses, characterized in that, Including: Obtaining multi-channel electroencephalogram data of an attention-getting glasses, and obtaining multi-dimensional features corresponding to the multi-channel electroencephalogram data; Performing time-domain transformation according to the multi-dimensional features to obtain signal components, and obtaining state representations corresponding to the signal components; performing hierarchical processing on the state representations to determine feature levels; Marking cognitive attributes based on the feature levels; constructing a state mapping according to the cognitive attributes, and generating a detection sequence according to the state mapping; Performing time-series analysis on the detection sequence, extracting change features, marking state nodes based on the change features, constructing a cognitive sequence using the state nodes, and generating a state description according to the cognitive sequence; Performing trend analysis on the state description, extracting features from the analysis results corresponding to the trend analysis, constructing a conversion rule according to the extracted features, and generating a prediction sequence according to the conversion rule; Performing threshold division on the prediction sequence, marking state intervals, performing feature aggregation according to the state intervals, constructing a criterion sequence using the aggregation results corresponding to the feature aggregation, and generating an evaluation result according to the criterion sequence; performing signal detection on the evaluation result to obtain detection information; Performing data conversion on the detection information, constructing a display sequence, performing state organization based on the display sequence, and marking state types according to the organization results of the state organization; Generating cognitive state information according to the state types.

2. The method according to claim 1, wherein The obtaining of the multi-dimensional features corresponding to the multi-channel electroencephalogram data includes: Establishing a data cache for the multi-channel electroencephalogram data; Performing window division based on the data cache, and generating a real-time sequence according to the division results of the window division; performing feature marking on the real-time sequence and extracting state indicators; Constructing a feature set based on the state indicators, and generating multi-dimensional features according to the feature set.

3. The method according to claim 2, characterized in that The establishing of the data cache for the multi-channel electroencephalogram data includes: Adopting a double-buffer mechanism for cache management to calculate the kurtosis and skewness of the divided multi-channel electroencephalogram data to evaluate data quality.

4. The method according to claim 2, wherein The performing of feature marking on the real-time sequence and extracting state indicators includes: Extracting signal amplitude change features of multi-channel electroencephalogram data in time-domain analysis, decomposing signal energy distribution features of multi-channel electroencephalogram data in frequency-domain analysis, obtaining signal dynamic evolution features of multi-channel electroencephalogram data in joint time-frequency domain analysis, and extracting non-linear dynamics features; Constructing the state indicators according to the signal amplitude change features, signal energy distribution features, signal dynamic evolution features and non-linear dynamics features.

5. The method according to claim 1, characterized in that, The state nodes include slope change points, abnormal fluctuation points and fluctuation nodes; the marking of state nodes based on the change features includes: Performing piecewise linear fitting on the trend features corresponding to the change features and identifying slope change points; Detecting abnormal fluctuation points of the fluctuation features corresponding to the change features using an adaptive threshold method; Calculating the local fluctuation intensity corresponding to the change features and marking fluctuation nodes exceeding the standard deviation of the historical mean.

6. The method according to claim 1, characterized in that, The generating of the state description according to the cognitive sequence includes: Obtaining the time intervals between adjacent states corresponding to the cognitive sequence to generate a basic description; Extract the state transition path law in the basic description to generate a behavior pattern description; Combine the importance weights to screen the high-weight state transitions in the pattern description to generate a priority description; Dynamically adjust the detail level of the pattern description according to the application scenario requirements to generate a state description including time series statistics and state evolution process.

7. The method according to claim 1, wherein The obtaining of the state representation corresponding to the signal component includes: Perform frequency domain decomposition on the signal component to mark the characteristic interval through the frequency domain decomposition; Generate a state representation according to the characteristic interval.

8. The method according to claim 1, wherein The performing signal detection on the evaluation result to obtain detection information includes: Classify and mark the detection result corresponding to the signal detection, and identify abnormal features based on the classification mark; Generate the detection information according to the abnormal features.

9. A real-time detection device for cognitive state, characterized in that, Includes: A data acquisition unit, configured to acquire multi-channel EEG data of an attention glasses, and acquire multi-dimensional features corresponding to the multi-channel EEG data; Perform time domain transformation according to the multi-dimensional features, acquire a signal component, and acquire a state representation corresponding to the signal component; perform hierarchical processing on the state representation to determine a feature level; Mark cognitive attributes based on the feature level; construct a state mapping according to the cognitive attributes, and generate a detection sequence according to the state mapping; A time series analysis unit, configured to perform time series analysis on the detection sequence, extract change features, mark state nodes based on the change features, construct a cognitive sequence using the state nodes, and generate a state description according to the cognitive sequence; Perform trend analysis on the state description, extract features from the analysis result corresponding to the trend analysis, construct a conversion rule according to the extracted features, and generate a prediction sequence according to the conversion rule; A threshold division unit, configured to perform threshold division on the prediction sequence, mark a state interval, perform feature aggregation according to the state interval, construct a criterion sequence using the aggregation result corresponding to the feature aggregation, and generate an evaluation result according to the criterion sequence; perform signal detection on the evaluation result to obtain detection information; A data conversion unit, configured to perform data conversion on the detection information, construct a display sequence, perform state organization based on the display sequence, and mark a state type according to the organization result of the state organization; Generate cognitive state information according to the state type.

10. A computer device, characterized in that, Includes a memory and a processor; the memory is used to store a computer program; the processor is used to execute the computer program and implement the method according to any one of claims 1 to 8 when executing the computer program.

Images