An organization member information collection management system

Abstract

The present application relates to a kind of organization member information collection management systems, including adaptive multi-modal acquisition unit, holographic correlation modeling unit, forecast formula dynamic updating unit and heterogeneous encryption storage unit;Adaptive multi-modal acquisition unit starts voice collection after identity is verified by voiceprint recognition, switches to image interaction mode in silence, synchronously interfaces wearable equipment and collects physiological parameters and pauses collection when abnormal;Holographic correlation modeling unit constructs three-dimensional holographic information model, and strong correlation information combination is filtered using quantum particle swarm optimization algorithm;Forecast formula dynamic updating unit establishes information attenuation prediction model, and generates pre-populated template by pushing update reminder in advance;Heterogeneous encryption storage unit stores information using space-time slicing encryption method, the purpose of the present application solves the problems that traditional organization member information collection management system is single in information collection mode, low in multi-source data correlation degree, information update lags behind and data storage security is insufficient.

Description

Technical Field

[0001] This invention belongs to the field of information management and settlement, and relates to an organizational member information collection and management system. Background Technology

[0002] In actual production and use, traditional organizational member information collection and management systems have revealed numerous problems. First, the information collection methods are relatively simplistic, mostly limited to manually filled forms or simple electronic questionnaires. This approach not only consumes a significant amount of members' time and energy, easily triggering resistance, but also struggles to capture information about members in dynamic scenarios, such as real-time work status and sudden emotional changes, resulting in a significant reduction in the completeness and timeliness of information collection. Second, the correlation between multiple data sources is low. Organizations typically obtain member information from various channels, such as work performance systems, attendance records, and training feedback, but this data is often isolated, lacking effective integration and correlation analysis. This prevents organizations from gaining a comprehensive understanding of members' overall situation from a holistic perspective, making it difficult to uncover the potential value behind the data and provide strong support for decision-making. Third, information updates are severely lagging. Members' work status, skill development, and personal circumstances are constantly changing, but traditional systems often rely on manual, periodic updates, failing to keep pace with these changes. This leads to outdated information used for decision-making, potentially causing errors. Finally, data storage security is insufficient. With the frequent occurrence of data breaches, the security of organizational member information faces severe challenges. Traditional systems have vulnerabilities in data encryption, access control, and storage architecture design, making them susceptible to external attacks and internal violations, which can lead to the leakage of member information and cause serious losses to the organization and its members.

[0003] After reviewing relevant materials, the main solutions to the above problems are as follows: Some patents propose improving information collection using a single information technology, such as Chinese patent CN202411253608.5, "An Information Collection and Release System for Enterprise Management Personnel." This system guides employees to fill in information through a specific information collection page, which standardizes the information collection process to some extent, but it still essentially remains limited to manual input and cannot solve the problem of a single collection method. Regarding data association, some systems attempt to use simple data integration techniques to centrally store data from different sources in the same database, but lack in-depth mining and analysis algorithms for the inherent relationships between data, making it difficult to achieve efficient association of multi-source data. For information updates, some systems use periodic reminder mechanisms, such as sending notifications of updated information to members at fixed time points. However, this method lacks real-time perception and intelligent prediction of information changes, and cannot respond in advance to information change needs. Regarding data storage security, some systems use conventional encryption algorithms, such as symmetric encryption algorithms, to encrypt the overall data. However, this encryption method is somewhat inadequate in the face of complex and ever-changing network attack methods, and once the encryption key is leaked, the entire data storage will face serious risks. While existing solutions address some of the problems of traditional systems, they fundamentally fail to comprehensively and systematically resolve the deep-seated issues in organizational member information collection and management systems, including information collection methods, multi-source data correlation, information updates, and data storage security. Therefore, a novel and comprehensive technological solution is urgently needed to address these challenges and meet the high standards of modern organizations for member information management. Summary of the Invention

[0004] This invention provides an organizational member information collection and management system, which solves the problems of traditional organizational member information collection and management systems in terms of single information collection methods, low correlation between multi-source data, delayed information updates, and insufficient data storage security.

[0005] To solve the above problems, the technical solution adopted by the invention is as follows:

[0006] An organization member information collection and management system includes:

[0007] S1 Adaptive Multimodal Acquisition Unit: is configured to perform the following steps to acquire organization member information: S11: after verifying the identity of organization members through voiceprint recognition, start the voice information acquisition mode, convert the voice information into text information in real time, and capture the emotional feature parameters in the voice at the same time;

[0008] S12: When it is detected that an organization member has been in a silent state for more than a preset time, it automatically switches to the image interaction mode, displays the dynamic information input interface, collects the member's body movement information through the camera, and converts the body movement into the corresponding information input command by combining the preset action command library.

[0009] S13: Synchronously connect to the wearable devices of members to collect their physiological status parameters. When the physiological status parameters exceed the normal range, pause the current collection process and push a rest reminder.

[0010] S2 holographic correlation modeling unit: connected to the adaptive multimodal acquisition unit, performs the following method steps to process information:

[0011] S21: Based on the collected text information, emotional feature parameters, body movement conversion instructions and physiological state parameters, a three-dimensional holographic information model is constructed, in which text information constitutes the basic data layer of the model, emotional feature parameters and body movement conversion instructions constitute the behavioral feature layer, and physiological state parameters constitute the state association layer.

[0012] S22: The quantum particle swarm optimization algorithm is used to iteratively optimize the three-dimensional holographic information model, calculate the coupling coefficient of information in different dimensions, and screen out strongly correlated information combinations with coupling coefficients higher than the threshold.

[0013] S3 Predictive Dynamic Update Unit: Connected to the holographic correlation modeling unit, it performs the following steps to update information:

[0014] S31: Based on the combination of strongly correlated information, establish an information decay prediction model and calculate the timeliness decay curve of each piece of information;

[0015] S32: Before the information actually changes, push the information update reminder in advance according to the timeliness decay curve, and predict the possible update content based on historical update data to generate a pre-filled update template;

[0016] S4 Heterogeneous Encrypted Storage Unit: Connected to the predictive dynamic update unit, it uses a spatiotemporal sharding encryption method to store information. This method includes: sharding the information according to timestamps and spatial features, encrypting each shard using a chaotic mapping-based encryption algorithm, and dynamically generating encryption keys for different shards based on the decryption results of the previous shard.

[0017] The principle and advantages of this solution are as follows:

[0018] Voiceprint recognition verifies identity, initiating voice acquisition mode to convert speech to text and capture emotional features. When silent, it automatically switches to image interaction mode, obtaining input commands through body language. Simultaneously, it connects to wearable devices to collect physiological parameters, pausing acquisition and prompting a rest period in case of abnormalities. This comprehensive approach gathers basic information, behavioral characteristics, and physiological states of members. Next, the holographic correlation modeling unit constructs a three-dimensional holographic information model based on the collected multi-type information, dividing it into a basic data layer, a behavioral feature layer, and a state correlation layer. The model is then optimized using a quantum particle swarm optimization algorithm, selecting combinations of strongly correlated information. The predictive dynamic update unit establishes a decay prediction model based on strongly correlated information, proactively pushing update reminders and generating pre-filled templates. Finally, the heterogeneous encrypted storage unit uses spatiotemporal sharding encryption, combined with chaotic mapping encryption algorithms and dynamically generated keys to store information. All units work collaboratively to form a complete information acquisition and management system.

[0019] Compared to existing technologies, in terms of information collection, current technologies mostly employ single text or voice input modes, resulting in a rigid collection process that can easily lead to member fatigue. This solution's adaptive multimodal collection automatically switches modes based on member status and incorporates physiological parameters to ensure a better collection experience. For example, when a member is emotionally agitated during voice collection, the system can capture their emotional characteristics; when silent, it automatically switches to image interaction; and when physiological parameters are abnormal, it pauses collection, making it more humane and efficient. Regarding information processing, traditional systems have low correlation with multi-source data, making it difficult to uncover deep connections. This solution, through constructing a 3D holographic information model and optimizing algorithms, can filter strongly correlated information combinations. For example, it can correlate members' emotions, body movements, and physiological parameter changes when under work pressure, providing more valuable references for management. In terms of information updates, existing technologies often lag behind actual changes. This solution's predictive updates can push reminders in advance and generate pre-filled templates. For instance, members receive reminders before their address changes and can pre-fill possible new address ranges, reducing information obsolescence. In terms of storage security, traditional encryption methods with fixed keys are easily cracked. This solution's spatiotemporal sharding encryption and dynamic key generation make each shard encryption independent and the keys associated, which greatly improves information security and brings about an efficient, accurate and secure organizational member information management effect that cannot be achieved by existing technologies.

[0020] Furthermore, in the adaptive multimodal acquisition unit, the voiceprint recognition in step S11 adopts a method combining deep learning and wavelet transform. First, the high-frequency features of the speech signal are extracted through wavelet transform, and then input into a convolutional neural network for voiceprint feature matching.

[0021] Furthermore, in the adaptive multimodal acquisition unit, the dynamic information input interface in step S12 is displayed using AR augmented reality technology, which overlays the information input box onto the real environment image. The acquisition frame rate of limb movement information is adjusted in real time according to the movement amplitude, and the larger the movement amplitude, the higher the acquisition frame rate.

[0022] Furthermore, in the holographic association modeling unit, the three-dimensional holographic information model in step S21 is stored using a holographic projection data structure, with the basic data layer, behavioral feature layer, and state association layer corresponding to different holographic grating parameters.

[0023] Furthermore, in the holographic correlation modeling unit, the quantum particle swarm optimization algorithm in step S22 introduces a quantum entanglement factor, which makes the position updates of different particles correlated with each other. The termination condition of the iterative optimization is that the rate of change of the coupling coefficient in N consecutive iterations is less than a preset minimum value.

[0024] Furthermore, in the heterogeneous encrypted storage unit, the timestamp of the spatiotemporal sharding encryption method is generated using atomic clock synchronization technology, the spatial features are divided based on the member's job level and department affiliation, and the initial parameters of the chaotic mapping are generated by hashing the organization's unique identifier code and the current timestamp.

[0025] Furthermore, it also includes an anomaly tracing unit, which is configured to: when an information anomaly is detected, trace back the construction process of the three-dimensional holographic information model, locate the source dimension of the abnormal information through the iterative record of the quantum particle swarm optimization algorithm, and generate an anomaly tracing report by combining the spatiotemporal positioning data of the wearable device.

[0026] Furthermore, the anomaly tracing unit also includes a virtual reproduction module, which reproduces the anomaly tracing process in the form of virtual animation, and marks the characteristic changes of the anomaly information in each processing stage.

[0027] Furthermore, based on the analysis results of the behavioral feature layer and state association layer of the three-dimensional holographic information model, the access permissions of different users to organizational member information are adjusted in real time. When the deviation between a user's access behavior and the preset permission model exceeds a threshold, secondary authentication is automatically triggered. Attached Figure Description

[0028] Figure 1 This is a system flowchart of the present invention. Detailed Implementation

[0029] Example 1, as Figure 1 As shown, an organizational member information collection and management system includes:

[0030] S1 Adaptive Multimodal Acquisition Unit: is configured to perform the following steps to acquire organization member information: S11: after verifying the identity of organization members through voiceprint recognition, start the voice information acquisition mode, convert the voice information into text information in real time, and capture the emotional feature parameters in the voice at the same time;

[0031] S12: When it is detected that an organization member is in a silent state for more than a preset time, the preset time is 5 seconds, or can be customized by the organization according to actual needs, the system will automatically switch to image interaction mode, display dynamic information input interface, collect the member's body movement information through camera, and convert the body movement into corresponding information input instructions in combination with the preset action instruction library.

[0032] S13: Synchronously connect to the wearable devices of members to collect their physiological status parameters. When physiological status parameters, including heart rate and blood pressure, exceed the normal range (referring to the "Guidelines for the Prevention and Treatment of Hypertension in China (2024 Revised Edition)", WS / T610-2018, "Multidisciplinary Expert Consensus on Heart Rate Management in Hypertensive Patients in China (2021 Edition)", and WS / T101-1988), the current data collection process is paused and a rest reminder is sent.

[0033] S2 holographic correlation modeling unit: connected to the adaptive multimodal acquisition unit, performs the following method steps to process information:

[0034] S21: Based on the collected text information, emotional feature parameters, body movement conversion instructions and physiological state parameters, a three-dimensional holographic information model is constructed, in which text information constitutes the basic data layer of the model, emotional feature parameters and body movement conversion instructions constitute the behavioral feature layer, and physiological state parameters constitute the state association layer.

[0035] S22: The quantum particle swarm optimization algorithm is used to iteratively optimize the three-dimensional holographic information model, calculate the coupling coefficient of information in different dimensions, and screen out strongly correlated information combinations with coupling coefficients higher than the threshold. The threshold is 0.8. The termination condition of the iterative optimization is that the rate of change of the coupling coefficient is less than 0.001 for 10 consecutive iterations.

[0036] S3 Predictive Dynamic Update Unit: Connected to the holographic correlation modeling unit, it performs the following steps to update information:

[0037] S31: Based on the combination of strongly correlated information, establish an information decay prediction model and calculate the timeliness decay curve of each piece of information;

[0038] S32: Before the information actually changes, push the information update reminder in advance according to the timeliness decay curve, and predict the possible update content based on historical update data to generate a pre-filled update template;

[0039] S4 Heterogeneous Encrypted Storage Unit: Connected to the predictive dynamic update unit, it uses a spatiotemporal sharding encryption method to store information. This method includes: sharding the information according to timestamps and spatial features, encrypting each shard using a chaotic mapping-based encryption algorithm, and dynamically generating encryption keys for different shards based on the decryption results of the previous shard.

[0040] The system verifies identity using voiceprint recognition, activates voice acquisition mode, converts speech to text, and captures emotional features. When silent, it automatically switches to image interaction mode, allowing users to input commands through body language. Simultaneously, it connects to wearable devices to collect physiological parameters, pausing acquisition and prompting a rest reminder in case of abnormalities, comprehensively acquiring basic information, behavioral characteristics, and physiological states of members. Next, the holographic correlation modeling unit constructs a three-dimensional holographic information model based on the collected multi-type information, dividing it into a basic data layer, a behavioral feature layer, and a state correlation layer. The model is then optimized using a quantum particle swarm optimization algorithm, selecting combinations of strongly correlated information. The predictive dynamic update unit establishes a decay prediction model based on strongly correlated information, proactively pushing update reminders and generating pre-filled templates. Finally, the heterogeneous encrypted storage unit uses a spatiotemporal sharding encryption method, combined with a chaotic mapping encryption algorithm and dynamically generated keys to store information. All units work collaboratively to form a complete information acquisition and management system.

[0041] Compared to existing technologies, in terms of information collection, current technologies mostly employ single text or voice input modes, resulting in a rigid collection process that can easily lead to member fatigue. This solution's adaptive multimodal collection automatically switches modes based on member status and incorporates physiological parameters to ensure a better collection experience. For example, when a member is emotionally agitated during voice collection, the system can capture their emotional characteristics; when silent, it automatically switches to image interaction; and when physiological parameters are abnormal, it pauses collection, making it more humane and efficient. Regarding information processing, traditional systems have low correlation with multi-source data, making it difficult to uncover deep connections. This solution, through constructing a 3D holographic information model and optimizing algorithms, can filter strongly correlated information combinations. For example, it can correlate members' emotions, body movements, and physiological parameter changes when under work pressure, providing more valuable references for management. In terms of information updates, existing technologies often lag behind actual changes. This solution's predictive updates can push reminders in advance and generate pre-filled templates. For instance, members receive reminders before their address changes and can pre-fill possible new address ranges, reducing information obsolescence. In terms of storage security, traditional encryption methods with fixed keys are easily cracked. This solution's spatiotemporal sharding encryption and dynamic key generation make each shard encryption independent and the keys associated, which greatly improves information security and brings about an efficient, accurate and secure organizational member information management effect that cannot be achieved by existing technologies.

[0042] In the adaptive multimodal acquisition unit, the voiceprint recognition in step S11 employs a combination of deep learning and wavelet transform. First, wavelet transform extracts high-frequency features of the speech signal, which are then input into a convolutional neural network for voiceprint feature matching. Wavelet transform excels at multi-scale analysis of non-stationary signals, accurately extracting high-frequency features from the speech signal. These high-frequency features often contain unique details of the voiceprint and are key information for distinguishing different individuals. Compared to traditional methods that rely solely on low-frequency features, this significantly improves feature recognition accuracy. Convolutional neural networks, as a typical deep learning model, possess powerful feature learning and pattern matching capabilities. Inputting the high-frequency features extracted by wavelet transform into the network allows for automatic learning of complex relationships between features through the nonlinear mapping of multiple neural networks, achieving more accurate voiceprint feature matching. This combination leverages the targeted advantages of wavelet transform in high-frequency feature extraction and the superiority of convolutional neural networks in complex feature matching. It effectively reduces the interference of background noise, speech intonation variations, and other factors on the recognition results. For example, in noisy environments, traditional voiceprint recognition may lead to misjudgments due to noise obscuring key features. However, this method can filter out some noise and extract clear high-frequency features through wavelet transform, followed by accurate matching through convolutional neural networks. This significantly improves the accuracy and robustness of identity verification, laying a solid foundation for the security and reliability of subsequent information collection.

[0043] In the adaptive multimodal acquisition unit, the dynamic information input interface in step S12 uses AR augmented reality technology to display the information input box overlaid on the real environment image. The acquisition frame rate of body movement information is adjusted in real time according to the movement amplitude; the larger the movement amplitude, the higher the acquisition frame rate. The use of AR augmented reality technology and real-time adjustment of the body movement acquisition frame rate in the dynamic information input interface in step S12 can significantly improve the efficiency and accuracy of information acquisition. AR technology overlays the information input box on the real environment image, breaking the traditional sense of separation between the interface and physical space. Members do not need to stare at the screen to perceive the position of the input box through peripheral vision. The correspondence between body movements and information input is more intuitive, reducing the learning cost of operation. For example, in an office setting, members can naturally raise their hands to correspond to the virtual input box to complete the selection, avoiding the redundancy of frequently looking down at the screen. Meanwhile, the frame rate for capturing body movements is dynamically adjusted according to the amplitude, ensuring the accuracy of capturing key information while avoiding resource waste: when the movement amplitude is large, such as a large wave to switch input items, the high frame rate can accurately capture the details of the movement trajectory, reducing misjudgments of commands caused by excessively fast movements; when the movement amplitude is small, such as a subtle gesture to confirm input, the low frame rate can reduce the device's computational load while meeting recognition requirements. Compared with fixed frame rate acquisition, this dynamic adaptation mode can ensure the accuracy of command conversion during fast operations and save system resources during slow operations, making the entire image interaction mode smoother and more efficient. Especially in scenarios where members are not familiar with the operation or where there is a lot of environmental interference, it can effectively reduce operational errors and improve the continuity of information acquisition.

[0044] In the holographic association modeling unit, the three-dimensional holographic information model in step S21 is stored using a holographic projection data structure. The basic data layer, behavioral feature layer, and state association layer each correspond to different holographic grating parameters. This significantly improves information storage efficiency and association retrieval capabilities. The holographic projection data structure itself has three-dimensional storage characteristics, breaking through the capacity limitations of traditional planar data storage. It allows multiple layers of information to be superimposed and stored in the same spatial dimension, saving storage space while laying the foundation for rapid interaction between different layers. Different layers corresponding to different holographic grating parameters ensure relative independence of information at each layer, avoiding data confusion. For example, the text information in the basic data layer corresponds to highly stable grating parameters, ensuring long-term complete storage of basic data; the emotional and physical movement information in the behavioral feature layer corresponds to highly dynamic responsive grating parameters, accurately recording instantaneous changes in behavior; and the physiological parameters in the state association layer correspond to highly sensitive grating parameters, promptly capturing subtle fluctuations in physiological states. This design enables the rapid location and extraction of target-level information by adjusting raster parameters when retrieving information. Simultaneously, leveraging the inherent interconnectivity of the holographic structure, it facilitates easy cross-layer information retrieval. For instance, when analyzing the physiological reasons behind a member's behavior, the behavioral feature layer and state association layer information under the corresponding raster parameters can be simultaneously retrieved to quickly obtain their associated data. Compared to traditional techniques where information is stored independently at each layer and requires complex algorithms to establish connections, this approach not only improves information processing efficiency but also more accurately uncovers potential connections between multiple layers, providing a more reliable foundation for subsequent model optimization and information applications.

[0045] In the predictive dynamic update unit, the information decay prediction model in step S31 combines information type, historical change frequency, and external environmental influencing factors. These external environmental influencing factors include the frequency of industry policy changes and the organizational business adjustment cycle. This combination significantly improves the accuracy of information timeliness prediction. Different information types exhibit fundamentally different decay characteristics. For example, basic personal information such as name and gender is relatively stable, while information such as job position and project participation tends to change over time. Combining information type provides the foundation for prediction. Historical change frequency reflects the dynamic patterns of information; for instance, if a member's job adjustment frequency is high, the decay rate of their related information will naturally be faster, providing data support for prediction. Introducing external environmental factors further overcomes the limitations of relying solely on internal data: the frequency of industry policy changes directly impacts the compliance information of organizational members, such as qualification certifications and training records; the more frequent the policy changes, the faster the timeliness of this information decays. The organizational business adjustment cycle is related to members' job responsibilities and project affiliations; the more frequent the business adjustments, the more urgent the need to update this information. For example, when industry policies are issued intensively, the model can predict, based on the frequency of policy changes, that members' qualification information needs to be updated in advance, avoiding compliance risks due to outdated information. When the organization enters a business adjustment period, it can remind members to update their project participation information in advance based on the adjustment cycle. Compared to traditional simple decay predictions based solely on time spans, this multi-dimensional model can more comprehensively capture the driving factors of information decay, making the prediction results more aligned with actual scenarios. This provides accurate guidance for subsequent update reminders and template pre-filling, effectively reducing management problems caused by information lag.

[0046] In the heterogeneous encrypted storage unit, the timestamps of the spatiotemporal sharding encryption method are generated using atomic clock synchronization technology. Spatial features are divided based on members' job levels and department affiliations. The initial parameters of the chaotic mapping are generated by hashing the organization's unique identifier code and the current timestamp. This multi-dimensional approach enhances the security and management efficiency of information storage. The timestamps generated by atomic clock synchronization technology have extremely high precision and uniqueness, ensuring accurate and unique time dimension identification for each information shard, avoiding shard confusion or missed detection due to timestamp errors, and providing a reliable time reference for subsequent information traceability. Dividing spatial features based on job levels and department affiliations allows information sharding to match the organizational management structure, facilitating fine-grained access control based on permissions, allowing only specific levels of personnel to access information shards in their corresponding departments, and enabling rapid location of target areas during information retrieval, thus improving management efficiency. The initial parameters of the chaotic mapping are generated by hashing the organization's unique identifier and the current timestamp, making the encryption starting point of each segment unpredictable. The organization's unique identifier ensures a strong binding between the initial parameters and the organization, while the current timestamp allows the parameters to change dynamically over time. Combined with the one-way nature of hashing, this greatly reduces the possibility of the initial parameters being cracked, thus enhancing the anti-attack capability of chaotic mapping encryption. For example, even if the encryption mode of a certain segment is partially cracked, because the key of the next segment is dynamically generated from the decryption result of the previous segment, and the initial parameters are always dynamically changing, it is difficult for attackers to crack all segment information in batches. Compared with the design of fixed timestamps, simple spatial division, and static initial parameters in traditional encryption methods, this multi-dimensional dynamic encryption mechanism can effectively resist risks such as time-series attacks, privilege escalation, and brute-force attacks, providing more comprehensive and reliable security for organizational member information.

[0047] It also includes an anomaly tracing unit, configured to: when an information anomaly is detected, trace back the construction process of the 3D holographic information model, locate the source dimension of the anomaly information through the iterative record of the quantum particle swarm optimization algorithm, and generate an anomaly tracing report by combining the spatiotemporal positioning data of the wearable device. This anomaly tracing unit significantly improves the accuracy and efficiency of information anomaly handling by tracing back the construction process of the 3D holographic information model, relying on the iterative record of the quantum particle swarm optimization algorithm to locate the source dimension of the anomaly information, and generating an anomaly tracing report by combining the spatiotemporal positioning data of the wearable device. When information anomalies occur, traditional methods often struggle to quickly locate the root cause of the problem. However, this unit, by leveraging the construction trajectory of the 3D holographic information model, can systematically investigate the information generation process of the basic data layer, behavioral feature layer, and state association layer. The iterative record of the quantum particle swarm optimization algorithm acts like an "operation log," clearly reflecting the changes in information of each dimension during coupling optimization, thereby accurately pinpointing the specific dimension to which the anomaly information belongs, such as a text information input error or a deviation in the recognition of emotional feature parameters. Simultaneously, by combining spatiotemporal positioning data from wearable devices, abnormal information can be correlated with the physical location and time of members, providing contextualized evidence for source tracing. For example, if a member's job information is abnormal, combining their location data during business adjustments can quickly determine whether it was caused by a failure to update organizational business information in a timely manner. Compared to the traditional method of relying solely on manual verification, this multi-dimensional source tracing mechanism not only shortens the time for anomaly investigation but also avoids missed detections and misjudgments, making the analysis of anomaly causes more targeted. This provides a reliable basis for subsequent information correction and system optimization, effectively reducing the management risks caused by information anomalies.

[0048] The anomaly tracing unit also includes a virtual reproduction module, which reproduces the anomaly tracing process in the form of virtual animation, marking the characteristic changes of the anomaly information at each processing stage. The virtual animation can concretize the abstract tracing process, dynamically demonstrating the interactions at each level in the construction of the 3D holographic information model, the parameter changes in the quantum particle swarm optimization algorithm iteration, and the complete trajectory of the anomaly information from its appearance to its identification. It also highlights the characteristics of the anomaly information in each stage, such as incorrect input points in text information and abnormal coupling coefficients in the behavioral feature layer, allowing relevant personnel to clearly and intuitively grasp the ins and outs of the anomaly. For example, when a member's job information is abnormal, the animation can reproduce the text conversion deviation during voice acquisition, the abnormal coupling with departmental affiliation information during modeling, and the process that was not correctly corrected during optimization iteration. Even managers unfamiliar with the system's technical details can quickly understand the root cause of the anomaly. Compared to pure text reports, this visual presentation not only lowers the threshold for information interpretation but also reduces communication errors, facilitating rapid consensus within the team and the development of targeted corrective measures, further improving the efficiency and accuracy of anomaly handling.

[0049] It also includes a dynamic permission allocation unit. Based on the analysis results of the behavioral feature layer and state association layer of the 3D holographic information model, it adjusts the access permissions of different users to organizational member information in real time. When the deviation of a user's access behavior from the preset permission model exceeds a threshold, it automatically triggers secondary authentication. The dynamic permission allocation unit adjusts access permissions in real time based on the analysis results of the behavioral feature layer and state association layer of the 3D holographic information model, and automatically triggers secondary authentication when the deviation of a user's access behavior from the preset permission model exceeds a threshold, which can greatly improve the security and flexibility of information access. The emotional feature parameters of the behavioral feature layer, the body movement commands, and the physiological state parameters of the state association layer can comprehensively reflect the user's true operational intentions and state. For example, when a user's operational behavior is abnormal due to emotional fluctuations or abnormal physiological state, the system can detect it in time through these features. Based on this dynamic adjustment of permissions, it can ensure the convenience of information access for legitimate users in reasonable scenarios, such as business backbones temporarily obtaining more information viewing permissions due to project needs, while avoiding the security vulnerabilities caused by rigid permissions. When access behavior deviates beyond a threshold, two-factor authentication is triggered, effectively blocking abnormal access. For example, if an unauthorized user attempts to access an authorized user's account by mimicking their user habits, the system will prevent their behavior through two-factor authentication, such as additional password or biometric verification. Compared to the traditional static permission allocation model, this mechanism, which combines dynamic adjustment with two-factor authentication, can adapt to the flexible information usage needs within an organization while accurately identifying potential security risks.

[0050] In actual use

[0051] (a) Adaptive Multimodal Acquisition Unit

[0052] This unit is used to collect information about organization members, specifically by performing the following steps: The voice information collection mode is initiated and executed. First, voiceprint recognition is performed to verify the identity of organization members, using a combination of deep learning and wavelet transform. Wavelet transform is used to perform multi-scale analysis on the collected voice signal, extracting high-frequency features. These high-frequency features contain unique details of the voiceprint and are key to distinguishing different individuals. Subsequently, the extracted high-frequency features are input into a convolutional neural network, which utilizes its powerful feature learning and pattern matching capabilities to perform voiceprint feature matching, completing the identity verification.

[0053] After identity verification, the voice information collection mode is activated, converting the voice information into text information in real time. At the same time, professional voice emotion analysis tools are used to capture emotional feature parameters in the voice, such as changes in speech rate and pitch.

[0054] Image interaction mode switching and execution

[0055] The system monitors the status of organization members in real time. When it detects that an organization member has been in a silent state for more than a preset time (in this embodiment, the preset time is set to 5 seconds), it automatically switches to the image interaction mode.

[0056] The dynamic information input interface displayed in this mode uses AR (Augmented Reality) technology, overlaying the information input box onto a real-world image captured by a camera. Simultaneously, the camera captures the user's body movements, with the frame rate adjusting in real-time according to the amplitude of the movement; the larger the movement, the higher the frame rate, ensuring accuracy and efficiency in information acquisition. For example, when a user waves their hand dramatically to switch input items, the frame rate is 30 frames per second; when a user makes a subtle gesture to confirm input, a lower frame rate, such as 10 frames per second, is used to reduce the device's computational load. The system has a pre-set motion command library, which is used to convert body movements into corresponding information input commands.

[0057] Wearable device docking and physiological parameter collection

[0058] It synchronously connects to members' wearable devices, such as smart bracelets and smartwatches, and collects their physiological parameters, including heart rate and blood pressure, through the built-in sensors of the devices.

[0059] The system has preset normal ranges for physiological state parameters. When the collected physiological state parameters exceed these ranges, the current collection process is immediately paused, and a rest reminder is sent to the member through a display interface or voice prompt.

[0060] (II) Holographic Relational Modeling Unit

[0061] This unit is connected to the adaptive multimodal acquisition unit to process the acquired information. The specific steps are as follows:

[0062] Construction of 3D Holographic Information Model

[0063] Based on the collected text information, emotional feature parameters, body movement conversion instructions, and physiological state parameters, a three-dimensional holographic information model is constructed. The text information constitutes the basic data layer of the model, the emotional feature parameters and body movement conversion instructions constitute the behavioral feature layer, and the physiological state parameters constitute the state association layer.

[0064] This 3D holographic information model uses a holographic projection data structure for storage. The basic data layer, behavioral feature layer, and state association layer each correspond to different holographic grating parameters. The text information in the basic data layer corresponds to grating parameters with high stability, ensuring the long-term complete storage of the basic data; the information in the behavioral feature layer corresponds to grating parameters with high dynamic responsiveness, accurately recording instantaneous changes in behavior; and the physiological parameters in the state association layer correspond to grating parameters with high sensitivity, promptly capturing subtle fluctuations in physiological state.

[0065] Optimization of 3D Holographic Information Model

[0066] The quantum particle swarm optimization algorithm is used to iteratively optimize the three-dimensional holographic information model. The quantum entanglement factor is introduced to make the position updates of different particles correlated, thereby enhancing the global search capability and avoiding getting trapped in local optima.

[0067] Calculate the coupling coefficients of information from different dimensions. The termination condition for iterative optimization is N consecutive iterations (N is set to 10), and the rate of change of the coupling coefficient during iteration is less than a preset minimum value (0.001). Select strongly correlated information combinations with coupling coefficients higher than a threshold value (0.8).

[0068] (III) Predictive Dynamic Update Unit

[0069] This unit is connected to the holographic correlation modeling unit for information updating. The specific steps are as follows:

[0070] Information decay prediction model establishment

[0071] An information decay prediction model is established based on strongly correlated information combinations. This model combines information type, historical change frequency, and external environmental influencing factors, including the frequency of industry policy changes and the organization's business adjustment cycle. For example, for qualification certification information closely related to industry policies, the decay rate is predicted by combining the frequency of industry policy changes; for project participation information related to the organization's business, the decay is predicted by combining the organization's business adjustment cycle. The model calculates the timeliness decay curves for each piece of information.

[0072] Information update reminders and pre-filled template generation

[0073] Before the information actually changes, push the information update reminder in advance according to the timeliness decay curve. The reminder can be sent via system message, email, etc.

[0074] Based on historical update data, data mining algorithms are used to predict potential update content and generate pre-populated update templates, enabling members to quickly update their information. For example, for members with multiple job adjustment records, the pre-populated template will include common job options when a potential job change is predicted.

[0075] (iv) Heterogeneous encrypted storage unit

[0076] This unit is connected to the predictive dynamic update unit and uses a spatiotemporal sharding encryption method to store information, as follows:

[0077] Information fragmentation: Information is fragmented based on both timestamps and spatial characteristics. Timestamps are generated using atomic clock synchronization technology to ensure high precision and uniqueness; spatial characteristics are based on members' job levels and department affiliations, ensuring that information fragmentation matches the organizational management structure.

[0078] Segmented encryption: Each segment is encrypted using a chaotic mapping-based encryption algorithm. The initial parameters of the chaotic mapping are generated by hashing the organization's unique identifier and the current timestamp. Furthermore, the encryption keys for different segments are dynamically generated based on the decryption results of the previous segment, greatly improving the security of information storage.

[0079] (v) Anomaly Tracing Unit

[0080] This unit is configured to perform anomaly tracing when an anomaly is detected, as detailed below:

[0081] Source dimension location of abnormal information: By tracing back the construction process of the three-dimensional holographic information model and analyzing the changes of information in each dimension during coupling optimization through the iterative record of the quantum particle swarm optimization algorithm, the source dimension of abnormal information is located, such as text information input errors and deviations in the recognition of emotional feature parameters.

[0082] Anomaly tracing report generation: By combining the spatiotemporal positioning data of wearable devices, the abnormal information is associated with the physical location and time of the member to generate an anomaly tracing report, providing a basis for anomaly handling.

[0083] Virtual Reproduction: The anomaly tracing unit also includes a virtual reproduction module, which reproduces the anomaly tracing process in the form of virtual animation, clearly marking the characteristic changes of the anomaly information in each processing stage, such as the error input points of text information, the anomaly coupling coefficient of the behavioral feature layer, etc., so that relevant personnel can intuitively understand the anomaly situation.

[0084] (vi) Dynamic permission allocation unit

[0085] Based on the analysis results of the behavioral feature layer and state association layer of the 3D holographic information model, this unit adjusts the access permissions of different users to organizational member information in real time. For example, when key business personnel need more information support due to project requirements, the system will appropriately expand their access permissions. At the same time, the system presets a permission model. When it detects that a user's access behavior deviates from the model by more than a threshold (set at 30%), it automatically triggers secondary authentication, such as requiring the user to enter an additional verification code or perform fingerprint recognition, to ensure the security of information access.

Claims

1. An organization member information collection management system characterized by comprising: Comprise: S1 adaptive multimodal acquisition unit: configured to perform the following method steps to collect organization member information: S11: after verifying the identity of the organization member through voiceprint recognition, start the voice information acquisition mode, convert the voice information into text information in real time, and capture the emotional feature parameters in the voice; S12: when it is detected that the organization member is in a silent state for more than a preset length of time, automatically switch to an image interaction mode, display a dynamic information entry interface, collect the member's body action information through a camera, and convert the body action into corresponding information input instructions in combination with a preset action instruction library; S13: synchronously interface with the member's wearable device, collect the member's physiological state parameters, and when the physiological state parameters exceed the normal range, pause the current acquisition process and push a rest prompt; S2 holographic correlation modeling unit: connected with the adaptive multimodal acquisition unit, and performing the following method steps to process information: S21: based on the collected text information, emotional feature parameters, body action conversion instructions, and physiological state parameters, construct a three-dimensional holographic information model, wherein the text information constitutes a basic data layer of the model, the emotional feature parameters and the body action conversion instructions constitute a behavior feature layer, and the physiological state parameters constitute a state correlation layer; S22: use a quantum particle swarm optimization algorithm to iteratively optimize the three-dimensional holographic information model, calculate the coupling coefficients of different dimension information, and select strong correlation information combinations with coupling coefficients higher than a threshold value; S3 predictive dynamic updating unit: connected with the holographic correlation modeling unit, and performing the following method steps to update information: S31: establish an information decay prediction model according to the strong correlation information combinations, and calculate the timeliness decay curves of each information; S32: before the information actually changes, push an information update reminder in advance according to the timeliness decay curves, and generate a pre-filled update template based on the predicted possible update content and historical update data; S4 heterogeneous encryption storage unit: connected with the predictive dynamic updating unit, and storing information using a spatiotemporal slicing encryption method, which comprises: double slicing the information according to time stamps and spatial characteristics, encrypting each slice using a chaotic mapping-based encryption algorithm, and dynamically generating encryption keys for different slices through the decryption results of the previous slice.

2. The system according to claim 1, wherein In S1, the voiceprint recognition of step S11 uses a method combining deep learning and wavelet transform, which first extracts the high-frequency features of the voice signal through wavelet transform, and then inputs the convolutional neural network for voiceprint feature matching.

3. The system of claim 1, wherein: In S1, the dynamic information entry interface of step S12 uses AR augmented reality technology to display, and the information entry box is superimposed on the real environment image. The frame rate of the body action information collection is adjusted in real time according to the action amplitude, and the greater the action amplitude, the higher the collection frame rate.

4. The system according to claim 1, wherein In S2, the three-dimensional holographic information model of step S21 is stored using a holographic projection data structure, and the basic data layer, the behavior feature layer, and the state correlation layer correspond to different holographic grating parameters, respectively.

5. The system according to claim 1, wherein In the S2, the quantum particle swarm optimization algorithm of the step S22 introduces a quantum entanglement factor, so that the position updates of different particles are correlated with each other, and the termination condition of the iterative optimization is that the change rate of the coupling coefficient in the last N iterations is less than a preset minimum value.

6. The system of claim 1, wherein, In the S4, the timestamp of the space-time slicing encryption method is generated by using an atomic clock synchronization technology, the spatial features are divided based on the position level and department affiliation of the members, and the initial parameters of the chaotic mapping are generated by hash operation on the unique identification code of the organization and the current timestamp.

7. The system of claim 1, wherein: The abnormality tracing unit is further configured to: when detecting information abnormality, backtrack the construction process of the three-dimensional holographic information model, locate the source dimension of the abnormal information through the iterative record of the quantum particle swarm optimization algorithm, and generate an abnormality tracing report in combination with the space-time positioning data of the wearable device.

8. The system of claim 7, wherein, The abnormality tracing unit further includes a virtual reproduction module configured to reproduce the abnormality tracing process in the form of a virtual animation and mark the feature changes of the abnormal information at each processing link.

9. The system of claim 1, wherein, The permission dynamic allocation unit is further configured to: based on the analysis results of the behavior feature layer and the state association layer of the three-dimensional holographic information model, adjust the access permission of different users to the information of the members of the organization in real time, and when detecting that the deviation degree of the access behavior of the user from the preset permission model exceeds a threshold value, automatically trigger secondary identity verification.

Images