Intelligent auxiliary system and method for emotion regulation based on cloud computing

Abstract

The invention discloses an emotion regulation intelligent auxiliary system and method based on cloud computing, and belongs to the technical field of mental health, and the system comprises a data collection module, an edge computing module, a cloud analysis module, an emotion analysis module, an intervention strategy module, a user interaction module and a feedback optimization module. The data acquisition module is in bidirectional signal connection with the edge calculation module, the edge calculation module is in bidirectional signal connection with the cloud analysis module, the cloud analysis module is in bidirectional signal connection with the emotion analysis module, and the emotion analysis module is in bidirectional signal connection with the intervention strategy module. The intervention strategy module is in bidirectional signal connection with the user interaction module, and the user interaction module is in bidirectional signal connection with the feedback optimization module. According to the emotion regulation intelligent auxiliary system and method based on cloud computing, the emotion state of the user can be comprehensively and accurately evaluated by fully utilizing cloud analysis, and the advantages of high capture accuracy, good intervention persistence effect and efficient emotion regulation are achieved.

Description

Technical Field

[0001] This invention relates to the field of mental health technology, specifically to a cloud-based intelligent auxiliary system and method for emotion regulation. Background Technology

[0002] Traditional emotion recognition relies on single-modal data, which has significant drawbacks: text analysis can only capture 60% of emotional cues, speech recognition is affected by environmental noise with an error rate of up to 25%, and facial expression recognition accuracy drops by 40% in cross-cultural scenarios; emotion regulation places stringent requirements on system response time: the golden window for intervention after triggering negative emotions is only 3-5 seconds, while traditional cloud architectures have a latency of 2-3 seconds; there are significant differences in user acceptance of emotion regulation solutions: the 25-35 age group prefers VR immersive training, while users over 45 years old prefer breathing guidance methods; emotion data is highly sensitive information, and current technical protection systems are weak, making information leakage easy.

[0003] Therefore, there is an urgent need for a cloud-based intelligent assistance system and method for emotion regulation to solve the problems mentioned above. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this invention provides a cloud-based intelligent auxiliary system and method for emotion regulation, which has advantages such as high accuracy in capturing emotions, good sustained intervention effects, and high efficiency in emotion regulation, thus solving the problems mentioned in the background technology.

[0005] To achieve the above objectives, the present invention provides the following technical solution:

[0006] The cloud-based intelligent assistance system for emotion regulation includes a data acquisition module, an edge computing module, a cloud analysis module, an emotion analysis module, an intervention strategy module, a user interaction module, and a feedback optimization module. The data acquisition module is bidirectionally connected to the edge computing module, the edge computing module is bidirectionally connected to the cloud analysis module, the cloud analysis module is bidirectionally connected to the emotion analysis module, the emotion analysis module is bidirectionally connected to the intervention strategy module, the intervention strategy module is bidirectionally connected to the user interaction module, and the user interaction module is bidirectionally connected to the feedback optimization module.

[0007] The data acquisition module, as the system's input, is responsible for collecting multimodal emotion data, including facial expression images, voice signals, text input, and physiological indicators. It captures users' emotional characteristics in real time through cameras, microphones, and smart bracelets, providing raw data support for subsequent analysis.

[0008] The edge computing module generates preliminary emotion recognition results by quickly processing facial expression data on a local device, reducing latency.

[0009] The cloud-based analysis module uses speech emotion recognition technology to analyze speech signals and generate online emotion recognition results.

[0010] The emotion analysis module identifies the user's current emotional state by analyzing facial expressions, voice tone, text semantics, and physiological signals, and assesses the intensity and trend of emotion changes, providing a scientific basis for intervention strategies.

[0011] The intervention strategy module matches personalized intervention plans from the strategy library based on the emotion analysis results, and pushes them to users in the form of natural language interaction through an intelligent agent to help them regulate their emotions.

[0012] The user interaction module provides a visual interface and multiple interaction methods, allowing users to view emotion reports, select intervention strategies, and set preference parameters. At the same time, it enhances the intervention experience through immersive scenarios, improving user participation and system usability.

[0013] The feedback optimization module is used to continuously collect user feedback data on intervention strategies, combine it with long-term emotional change trajectories, and use machine learning algorithms to optimize the emotion analysis model and intervention strategy library. Through iterative updates, the system's adaptability is improved, ensuring that the intervention effect continues to improve.

[0014] Furthermore, the data acquisition module includes a face acquisition module, a voice acquisition module, a physiological acquisition module, and a text acquisition module, all of which are bidirectionally connected to the data acquisition module.

[0015] Furthermore, the facial capture module captures changes in the user's facial expressions in real time through a camera, extracts micro-expression and muscle movement features, identifies emotion-related facial action units, and combines computer vision algorithms to convert facial features into emotion labels, providing the system with intuitive evidence of emotion expression. The input signal of the facial capture module is connected to a camera.

[0016] Furthermore, the voice acquisition module records the user's voice signal through a microphone, analyzes the acoustic features of tone, speech rate, volume and pauses, extracts semantic content by combining natural language processing technology, and uses a voice emotion recognition model to determine the emotional state contained in the voice. The input signal of the voice acquisition module is connected to a microphone.

[0017] Furthermore, the physiological acquisition module collects physiological signals, including heart rate variability, skin conductance, and electroencephalogram (EEG), through wearable devices or non-contact sensors. The input terminals of the physiological acquisition module are respectively connected to a smart bracelet and a camera.

[0018] Furthermore, the text acquisition module obtains user text content through keyboard input, speech-to-text conversion, or social media data capture, and uses NLP technology to analyze the text's emotional polarity, emotion category, and contextual information, supplementing emotional cues from other modalities. The input terminals of the text acquisition module are respectively connected to a microphone and a keyboard.

[0019] Furthermore, the intervention strategy module includes intervention programs such as cognitive behavioral therapy guidance, relaxation training, and music scene adjustment.

[0020] Another technical problem this invention aims to solve is to provide a cloud-based intelligent assistance method for emotion regulation, comprising the following steps:

[0021] S1. Multimodal data acquisition and processing: comprehensively capture user emotional cues, eliminate noise interference, and provide high-quality data for subsequent analysis;

[0022] S2, Cloud-based Multimodal Emotion Recognition: Integrates multi-dimensional data to generate accurate emotion labels and intensity assessments;

[0023] S3. Personalized intervention strategy generation: Recommend the most effective adjustment plan based on the user's emotional state, historical preferences and situational factors;

[0024] S4. Interactive intervention implementation feedback: Guide users to perform regulatory behaviors through natural interaction and collect feedback data;

[0025] S5. System Optimization and Model Iteration: Based on user feedback and long-term data, improve the accuracy of emotion recognition and the effectiveness of intervention;

[0026] S6. Privacy Protection and Security Mechanisms: Ensure the security of user data throughout its entire lifecycle and comply with ethical standards.

[0027] Furthermore, the multimodal data acquisition and processing in S1 includes the following steps:

[0028] S1-1 Facial Data Acquisition: Use a camera to capture facial expressions in real time and extract the movement trajectory of key points (such as the corners of the eyes and mouth);

[0029] S1-2, Voice Data Acquisition: Record the voice signal through a microphone, segment it into short time frames and extract acoustic features;

[0030] S1-3, Physiological Data Acquisition: Heart rate variability and skin electrical activity signals are collected via a smart bracelet;

[0031] S1-4. Text Data Acquisition: Acquire user text content through keyboard input or speech-to-text conversion, segment words, and remove stop words.

[0032] Furthermore, the interactive intervention implementation feedback in S4 includes the following steps:

[0033] S4-1, Multi-channel interaction:

[0034] Voice interaction: The intelligent agent guides the user with empathetic language;

[0035] Visual interaction: AR / VR scene simulation of natural environment, combined with breathing guidance animation;

[0036] Tactile feedback: The smart bracelet vibrates to indicate breathing rhythm;

[0037] S4-2. Real-time effect monitoring: During the intervention, facial, voice and physiological data are continuously collected to assess emotional changes. If the effect is not good, the backup strategy is automatically switched.

[0038] S4-3. User feedback collection: After the intervention, subjective evaluations are collected through scales or open-ended questions.

[0039] Compared with existing technologies, this invention provides a cloud-based intelligent auxiliary system and method for emotion regulation, which has the following beneficial effects:

[0040] 1. This cloud-based intelligent auxiliary system and method for emotion regulation, through a bidirectional signal connection between the data acquisition module and the edge computing module, enables the collected emotion-related data to be preliminarily processed and analyzed at the edge, close to the data source. The edge computing module can quickly filter, denoise, and extract features from the data, reducing unnecessary data transmission and avoiding network congestion and delays caused by directly uploading large amounts of raw data to the cloud. This effectively improves data processing efficiency and real-time performance. Through the bidirectional signal connection between the cloud analysis module and the emotion analysis module, the cloud analysis can be fully utilized to comprehensively and accurately assess the user's emotional state, achieving the advantage of high accuracy in capturing emotions.

[0041] 2. This cloud-based intelligent auxiliary system and method for emotion regulation, through a two-way signal connection between the emotion analysis module and the intervention strategy module, can adjust the intervention plan in real time according to the dynamic changes in emotion analysis, ensuring that the intervention measures always match the user's emotional state, thereby improving the effectiveness and pertinence of emotion regulation. Through a two-way signal connection between the intervention strategy module and the user interaction module, it can understand user feedback in real time and adjust the interaction content and method according to user choices and inputs, making the entire interaction process smoother and more convenient, and improving the user experience. It achieves the advantages of good intervention sustainability and efficient emotion regulation.

[0042] 3. This cloud-based intelligent assistance system and method for emotion regulation, through the deep integration of multimodal perception, edge-cloud collaboration, personalized intervention and privacy protection, constructs a complete closed loop from emotion recognition to behavior regulation. It provides a scalable and low-cost digital solution for solving the global mental health crisis. Through closed-loop design (data collection → analysis → intervention → feedback → optimization), it forms a continuously improving intelligent system, providing scientific and efficient assistance for emotion regulation. Attached Figure Description

[0043] Figure 1 This is a system block diagram of the cloud-based intelligent auxiliary system for emotion regulation proposed in this invention.

[0044] Figure 2 This is a system block diagram of the data acquisition module of the cloud-based intelligent auxiliary system for emotion regulation proposed in this invention;

[0045] Figure 3 This is a flowchart illustrating the cloud-based intelligent assistance method for emotion regulation proposed in this invention. Detailed Implementation

[0046] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0047] Example 1:

[0048] Please see Figures 1 to 2 The cloud-based intelligent assistance system for emotion regulation in this embodiment includes a data acquisition module, an edge computing module, a cloud analysis module, an emotion analysis module, an intervention strategy module, a user interaction module, and a feedback optimization module. The data acquisition module and the edge computing module are bidirectionally connected, the edge computing module and the cloud analysis module are bidirectionally connected, the cloud analysis module and the emotion analysis module are bidirectionally connected, the emotion analysis module and the intervention strategy module are bidirectionally connected, the intervention strategy module and the user interaction module are bidirectionally connected, and the user interaction module and the feedback optimization module are bidirectionally connected.

[0049] Specifically, the data acquisition module, as the system's input, is responsible for collecting multimodal emotion data, including facial expression images, voice signals, text input, and physiological indicators (such as brain waves and heart rate). It captures users' emotional characteristics in real time through cameras, microphones, and smart bracelets, providing raw data support for subsequent analysis.

[0050] Specifically, the edge computing module generates preliminary emotion recognition results (such as Elocal) by quickly processing facial expression data on local devices, thereby reducing latency.

[0051] Specifically, the cloud-based analytics module uses speech emotion recognition technology to analyze speech signals and generate online emotion recognition results (such as Eonline); the edge computing module and the cloud-based analytics module use a decision-level fusion algorithm (such as the weighted value method) to comprehensively determine the emotional state, taking into account both real-time performance and accuracy.

[0052] Specifically, the emotion analysis module uses deep learning algorithms (such as CNN and LSTM) to extract features and classify emotions from multimodal data. By analyzing facial expressions, voice tone, text semantics and physiological signals, it identifies the user's current emotional state (such as happiness, anxiety, depression) and assesses the intensity and trend of emotion, providing a scientific basis for intervention strategies.

[0053] Specifically, the intervention strategy module matches personalized intervention plans from the strategy library based on the emotion analysis results, and pushes them to users in the form of natural language interaction through intelligent agents (such as chatbots and virtual psychological counselors) to help them regulate their emotions.

[0054] Specifically, the user interaction module provides a visual interface and multiple interaction methods (such as voice, text, and touch) to support users in viewing emotion reports, selecting intervention strategies, and setting preference parameters. At the same time, it enhances the intervention experience through immersive scenarios (such as virtual natural environments), thereby improving user participation and system usability.

[0055] Specifically, the feedback optimization module is used to continuously collect user feedback data on intervention strategies (such as the degree of mood improvement and strategy acceptance). Combining this data with long-term mood change trajectories, the module uses machine learning algorithms to optimize the mood analysis model and intervention strategy library. Through iterative updates, the system's adaptability is improved, ensuring that the intervention effect continues to improve.

[0056] In this embodiment, the data acquisition module includes a face acquisition module, a voice acquisition module, a physiological acquisition module, and a text acquisition module. The face acquisition module, the voice acquisition module, the physiological acquisition module, and the text acquisition module are all bidirectionally connected to the data acquisition module.

[0057] Specifically, the facial capture module uses a camera to capture real-time changes in the user's facial expressions, extracting micro-expressions and muscle movement features to identify emotion-related facial action units (such as frowning and raising the corners of the mouth). Combined with computer vision algorithms (such as OpenCV, DLIB, or deep learning models like FERNet), facial features are converted into emotion labels (such as happiness, anger, and sadness), providing the system with intuitive evidence of emotional expression. The input signal of the facial capture module is connected to a camera.

[0058] The value of facial recognition modules in cloud computing systems:

[0059] Real-time performance: Edge computing can quickly process facial data and generate preliminary emotion results (such as Elocal), reducing cloud transmission latency.

[0060] Multi-scene adaptation: Supports facial expression recognition in dynamic scenes (such as video conferencing and daily interactions), and adapts to different lighting and angle conditions.

[0061] Non-invasive: Data can be collected naturally without the need for active user cooperation, thus improving the user experience.

[0062] Specifically, the voice acquisition module records the user's voice signal through a microphone, analyzes the acoustic features of tone, speech rate, volume and pauses, extracts semantic content (such as keywords and sentiment) by combining natural language processing (NLP) technology, and uses a voice emotion recognition model (such as an end-to-end network based on LSTM or Transformer) to determine the emotional state (such as excitement, anxiety, calm) contained in the voice. The input signal of the voice acquisition module is connected to a microphone.

[0063] The value of voice acquisition modules in cloud computing systems:

[0064] Rich in emotional dimensions: Voice can convey emotional details that are difficult to capture with facial expressions (such as a trembling voice when nervous).

[0065] Cloud-based collaborative processing: Raw voice data is uploaded to the cloud and deep analysis is performed using high-performance computing resources to generate online emotion results (such as Eonline), which are then fused with facial data to improve accuracy.

[0066] Cross-language support: Adapts to the emotion recognition needs of users of different languages ​​through multilingual emotion dictionaries and transfer learning techniques.

[0067] Specifically, the physiological data acquisition module collects physiological signals, including heart rate variability (HRV), electrical skin activity (EDA), and electroencephalogram (EEG), through wearable devices (such as smart bracelets and EEG caps) or non-contact sensors (such as cameras measuring heart rate). These signals reflect the activity of the autonomic nervous system and are closely related to emotional states (e.g., heart rate increases when anxious, HRV increases when relaxed). The input terminals of the physiological data acquisition module are connected to a smart bracelet and a camera, respectively.

[0068] The value of physiological data acquisition modules in cloud computing systems:

[0069] Objectivity: Physiological signals are not easily affected by subjective expression, providing evidence of the physiological basis of emotions.

[0070] Long-term monitoring: By leveraging the storage capabilities of cloud computing, it is possible to track long-term changes in users' physiological indicators and assist in analyzing patterns of emotional fluctuations.

[0071] Health Links: Identify potential health risks (such as a weakened immune system due to chronic stress) through correlation analysis of emotional and physiological data.

[0072] Specifically, the text acquisition module obtains user text content (such as diaries, chat logs, and search history) through keyboard input, speech-to-text conversion, or social media data capture. It uses NLP technology to analyze the text's emotional polarity (positive / negative), emotion category (such as anger or gratitude), and contextual information, supplementing emotional cues from other modalities. The input terminals of the text acquisition module are connected to a microphone and a keyboard.

[0073] The value of text acquisition modules in cloud computing systems:

[0074] Deep semantic understanding: Text can reveal users' implicit emotions (such as sarcasm or irony), which need to be accurately interpreted by combining context and domain knowledge.

[0075] Personalized intervention: By analyzing users' historical text, a database of emotion-triggered events is built to provide a basis for personalized intervention strategies (such as recommending time management suggestions to users with high work pressure).

[0076] Multimodal fusion: It integrates text emotion results with facial, voice and physiological data to overcome the limitations of single modality (such as contradictory scenarios where the face is smiling but the text expresses sadness).

[0077] Advantages of collaboration between facial recognition, voice recognition, physiological data acquisition, and text recognition modules, and cloud computing:

[0078] Data fusion: Data from each module is integrated through decision-level fusion algorithms (such as weighted voting and DS evidence theory) to determine the emotional state and improve recognition robustness.

[0079] Elastic computing resources: The cloud can dynamically allocate computing resources to handle high-concurrency data streams (such as facial / voice analysis when multiple people use the system at the same time).

[0080] Privacy protection: Data is encrypted during transmission and storage, and the federated learning framework is used to isolate model training data from user data.

[0081] Continuous optimization: Through cloud-based feedback loops, the emotion recognition model is iteratively updated using historical user data to adapt to individual differences and the diversity of emotion expression.

[0082] Specifically, the intervention strategies module includes intervention programs such as cognitive behavioral therapy guidance, relaxation training (such as deep breathing and meditation), and music scene adjustment.

[0083] Example 2:

[0084] Please see Figure 3 The cloud-based intelligent assistance method for emotion regulation in this embodiment includes the following steps:

[0085] S1. Multimodal data acquisition and processing: comprehensively capture user emotional cues, eliminate noise interference, and provide high-quality data for subsequent analysis;

[0086] S2, Cloud-based Multimodal Emotion Recognition: Integrates multi-dimensional data to generate accurate emotion labels and intensity assessments;

[0087] S3. Personalized intervention strategy generation: Recommend the most effective adjustment plan based on the user's emotional state, historical preferences and situational factors;

[0088] S4. Interactive intervention implementation feedback: Guide users to perform regulatory behaviors through natural interaction and collect feedback data;

[0089] S5. System Optimization and Model Iteration: Based on user feedback and long-term data, improve the accuracy of emotion recognition and the effectiveness of intervention;

[0090] S6. Privacy Protection and Security Mechanisms: Ensure the security of user data throughout its entire lifecycle and comply with ethical standards.

[0091] In this embodiment, the multimodal data acquisition and processing in S1 includes the following steps:

[0092] S1-1 Facial Data Acquisition: Use a camera to capture facial expressions in real time and extract the movement trajectory of key points (such as the corners of the eyes and mouth).

[0093] Preprocessing: Eliminate the effects of angle and lighting through face alignment and illumination normalization (such as histogram equalization);

[0094] S1-2, Voice Data Acquisition: Record the voice signal through the microphone, segment it into short time frames (e.g., 25ms) and extract acoustic features (e.g., Mel-frequency cepstral coefficients MFCC, fundamental frequency F0).

[0095] Preprocessing: noise reduction (e.g., spectral subtraction), endpoint detection (distinguishing between speech and non-speech segments);

[0096] S1-3, Physiological Data Acquisition: Heart rate variability (HRV) and electrical skin activity (EDA) signals are acquired through a smart bracelet.

[0097] Preprocessing: filtering and noise reduction (such as Butterworth filter), sliding window smoothing;

[0098] S1-4. Text Data Acquisition: Acquire user text content through keyboard input or speech-to-text conversion, segment words, and remove stop words.

[0099] Preprocessing: sentiment lexicon matching (e.g., Chinese sentiment lexicon ontology), word vector representation (e.g., Word2Vec).

[0100] In this embodiment, cloud-based multimodal emotion recognition in S2 includes the following steps:

[0101] S2-1, Unimodal Emotion Classification: Facial Emotion Recognition: Input preprocessed facial features into a CNN model (such as ResNet), and output the emotion category (such as happiness, anger) and confidence score.

[0102] Voice emotion recognition: LSTM network is used to analyze acoustic feature sequences, and attention mechanism is combined to focus on key emotional segments.

[0103] Physiological emotion assessment: Based on HRV and EDA signals, emotional activation (e.g., high / low arousal) is determined using random forest or SVM models.

[0104] Text sentiment analysis: Extract semantic sentiment polarity (positive / negative / neutral) using a BERT pre-trained model.

[0105] S2-2, Multimodal Fusion Decision: Feature Layer Fusion: Concatenate the features of each modality into a joint vector and input it into the fusion model (such as Multimodal Transformer).

[0106] Decision-making level integration: A weighted voting method is adopted, and weights are allocated according to modal reliability (e.g., physiological signals have higher weights under stress).

[0107] Output: Generates a comprehensive emotion label (e.g., "moderate anxiety") and a trend chart of emotion changes;

[0108] In this embodiment, the generation of personalized intervention strategies in S3 includes the following steps:

[0109] S3-1. User profile construction: Integrate basic user information (age, gender), emotional history records (such as the frequency of anxiety attacks), and intervention feedback data.

[0110] Use clustering algorithms (such as K-means) to segment users into categories (such as "high-pressure workplace people" and "socially anxious people").

[0111] S3-2, Strategy library matching: Based on the rule engine: trigger preset strategies based on emotion tags (such as "anger → deep breathing training").

[0112] Based on reinforcement learning: optimize strategy selection (e.g., adjust music type or guiding style) through user feedback (e.g., post-intervention emotion rating).

[0113] S3-3, Context-Aware Adjustment: Dynamically adjust the strength of the strategy based on time (e.g., late at night), location (e.g., office), and social relationship (e.g., alone / group chat).

[0114] Example: Recommend mindfulness meditation to users who feel depressed when alone, and push progressive muscle relaxation to users who feel tense during meetings.

[0115] In this embodiment, the interactive intervention implementation feedback in S4 includes the following steps:

[0116] S4-1, Multi-channel interaction:

[0117] Voice interaction: Intelligent agents (such as chatbots) guide users with empathetic language (e.g., "I notice you seem a little nervous, would you like to do some breathing exercises together?");

[0118] Visual interaction: AR / VR scene simulation of natural environments (such as forests and beaches), combined with breathing-guided animations;

[0119] Tactile feedback: The smart bracelet vibrates to indicate breathing rhythm (e.g., vibrates when inhaling and stops when exhaling);

[0120] S4-2. Real-time effect monitoring: During the intervention, facial, voice and physiological data are continuously collected to assess emotional changes (such as decreased heart rate and increased smiling frequency). If the effect is not good, the backup strategy is automatically switched (such as switching from music therapy to cognitive reconstruction dialogue).

[0121] S4-3. User feedback collection: After the intervention, subjective evaluations are collected using scales (e.g., 1-5 points) or open-ended questions (e.g., "Did this method alleviate your anxiety?").

[0122] In this embodiment, system optimization and model iteration in S5 include the following steps:

[0123] S5-1, Data Labeling and Augmentation: Manually label fuzzy emotion samples (such as "smiling with tears in his eyes"), and combine them with automatic labeling tools to generate a large-scale training set.

[0124] Generative Adversarial Networks (GANs) can be used to augment data for a few emotion categories (e.g., synthesizing new data when there are insufficient samples for "surprise").

[0125] S5-2, Model Retraining: Regularly fine-tune the emotion recognition model with new data (e.g., update the facial CNN weights every two weeks).

[0126] Transfer learning can be used to adapt a general model to a specific user group (such as optimizing speech emotion recognition parameters for the elderly).

[0127] S5-3. Strategy effectiveness evaluation: Compare the long-term effects of different strategies through a control group experiment (e.g., changes in anxiety scale scores after 4 weeks of continuous meditation intervention).

[0128] Eliminate inefficient strategies (e.g., reduce the recommendation priority of a certain type of music if it is ineffective for 30% of users).

[0129] In this embodiment, the privacy protection and security mechanism in S6 includes the following steps:

[0130] S6-1, Data Encryption:

[0131] Transport layer: Data streams are encrypted using the TLS 1.3 protocol;

[0132] Storage layer: Sensitive data (such as physiological signals) is stored using AES-256 encryption.

[0133] S6-2, Access Control: Role-Based Access Control (RBAC) restricts developers to accessing only anonymized data.

[0134] Dynamic desensitization: Automatically hides user ID information during the analysis phase.

[0135] S6-3 Compliance Audit: Regularly generate data usage reports that comply with GDPR and HIPAA regulations.

[0136] Users can export or delete their personal data at any time, and the system supports the "right to be forgotten".

[0137] The cloud-based intelligent assistance method for emotion regulation in this embodiment has the following effects:

[0138] Accuracy: Multimodal fusion overcomes the limitations of single modality (such as contradictory emotions like a smiling face but a trembling voice).

[0139] Real-time performance: Edge-cloud collaboration enables low-latency interaction (total latency <1 second).

[0140] Personalization: By dynamically optimizing strategies through user profiling and reinforcement learning, it adapts to individual differences.

[0141] Scalability: Cloud computing resources can be scaled elastically to support large-scale concurrent use by users.

[0142] In summary:

[0143] This cloud-based intelligent auxiliary system and method for emotion regulation uses a bidirectional signal connection between a data acquisition module and an edge computing module. This allows the collected emotion-related data to be preliminarily processed and analyzed at the edge, close to the data source. The edge computing module can quickly filter, reduce noise, and extract features from the data, reducing unnecessary data transmission and avoiding network congestion and latency caused by directly uploading large amounts of raw data to the cloud. This effectively improves data processing efficiency and real-time performance. Through the bidirectional signal connection between the cloud analysis module and the emotion analysis module, cloud analysis can be fully utilized to comprehensively and accurately assess the user's emotional state, achieving the advantage of high accuracy in capturing emotions.

[0144] This cloud-based intelligent auxiliary system and method for emotion regulation, through a two-way signal connection between the emotion analysis module and the intervention strategy module, can adjust the intervention plan in real time according to the dynamic changes in emotion analysis, ensuring that the intervention measures always match the user's emotional state and improving the effectiveness and pertinence of emotion regulation. Through a two-way signal connection between the intervention strategy module and the user interaction module, it can understand user feedback in real time and adjust the interaction content and method according to user choices and inputs, making the entire interaction process smoother and more convenient, and improving the user experience. It achieves the advantages of good sustained intervention effect and efficient emotion regulation.

[0145] This cloud-based intelligent assistance system and method for emotion regulation, through the deep integration of multimodal perception, edge-cloud collaboration, personalized intervention and privacy protection, constructs a complete closed loop from emotion recognition to behavior regulation. It provides a scalable and low-cost digital solution for addressing the global mental health crisis. Through closed-loop design (data collection → analysis → intervention → feedback → optimization), it forms a continuously improving intelligent system, providing a scientific and efficient auxiliary means for emotion regulation.

[0146] It should be noted that, in this document, relational terms such as "first" and "second" are used merely to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0147] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A cloud-based intelligent assistance system for emotion regulation, characterized in that: It includes a data acquisition module, an edge computing module, a cloud analysis module, a sentiment analysis module, an intervention strategy module, a user interaction module, and a feedback optimization module. The data acquisition module is bidirectionally connected to the edge computing module, the edge computing module is bidirectionally connected to the cloud analysis module, the cloud analysis module is bidirectionally connected to the sentiment analysis module, the sentiment analysis module is bidirectionally connected to the intervention strategy module, the intervention strategy module is bidirectionally connected to the user interaction module, and the user interaction module is bidirectionally connected to the feedback optimization module. The data acquisition module, as the system's input, is responsible for collecting multimodal emotion data, including facial expression images, voice signals, text input, and physiological indicators. It captures users' emotional characteristics in real time through cameras, microphones, and smart bracelets, providing raw data support for subsequent analysis. The edge computing module generates preliminary emotion recognition results by quickly processing facial expression data on a local device, reducing latency. The cloud-based analysis module uses speech emotion recognition technology to analyze speech signals and generate online emotion recognition results. The emotion analysis module identifies the user's current emotional state by analyzing facial expressions, voice tone, text semantics, and physiological signals, and assesses the intensity and trend of emotion changes, providing a scientific basis for intervention strategies. The intervention strategy module matches personalized intervention plans from the strategy library based on the emotion analysis results, and pushes them to users in the form of natural language interaction through an intelligent agent to help them regulate their emotions. The user interaction module provides a visual interface and multiple interaction methods, allowing users to view emotion reports, select intervention strategies, and set preference parameters. At the same time, it enhances the intervention experience through immersive scenarios, improving user participation and system usability. The feedback optimization module is used to continuously collect user feedback data on intervention strategies, combine it with long-term emotional change trajectories, and use machine learning algorithms to optimize the emotion analysis model and intervention strategy library. Through iterative updates, the system's adaptability is improved, ensuring that the intervention effect continues to improve.

2. The cloud-based intelligent assistance system for emotion regulation according to claim 1, characterized in that: The data acquisition module includes a face acquisition module, a voice acquisition module, a physiological acquisition module, and a text acquisition module. The face acquisition module, voice acquisition module, physiological acquisition module, and text acquisition module are all bidirectionally connected to the data acquisition module.

3. The cloud-based intelligent assistance system for emotion regulation according to claim 2, characterized in that: The facial capture module captures the user's facial expression changes in real time through a camera, extracts micro-expression and muscle movement features, identifies emotion-related facial action units, and combines computer vision algorithms to convert facial features into emotion labels, providing the system with intuitive evidence of emotion expression. The input signal of the facial capture module is connected to a camera.

4. The cloud-based intelligent assistance system for emotion regulation according to claim 2, characterized in that: The voice acquisition module records the user's voice signal through a microphone, analyzes the acoustic features of tone, speech rate, volume and pauses, extracts semantic content by combining natural language processing technology, and uses a voice emotion recognition model to determine the emotional state contained in the voice. The input signal of the voice acquisition module is connected to a microphone.

5. The cloud-based intelligent assistance system for emotion regulation according to claim 2, characterized in that: The physiological acquisition module collects physiological signals, including heart rate variability, skin conductance, and electroencephalogram (EEG), through wearable devices or non-contact sensors. The input terminals of the physiological acquisition module are connected to a smart bracelet and a camera, respectively.

6. The cloud-based intelligent assistance system for emotion regulation according to claim 2, characterized in that: The text acquisition module obtains user text content through keyboard input, speech-to-text conversion, or social media data capture. It uses NLP technology to analyze the text's emotional polarity, emotion category, and context, and supplements emotional cues from other modalities. The input terminals of the text acquisition module are respectively connected to a microphone and a keyboard.

7. The cloud-based intelligent auxiliary system for emotion regulation according to claim 1, characterized in that: The intervention strategy module includes intervention programs such as cognitive behavioral therapy guidance, relaxation training, and music scene adjustment.

8. A cloud-based intelligent assistance method for emotion regulation, employing the cloud-based intelligent assistance system for emotion regulation as described in any one of claims 1-7, characterized in that, Includes the following steps: S1. Multimodal data acquisition and processing: comprehensively capture user emotional cues, eliminate noise interference, and provide high-quality data for subsequent analysis; S2, Cloud-based Multimodal Emotion Recognition: Integrates multi-dimensional data to generate accurate emotion labels and intensity assessments; S3. Personalized intervention strategy generation: Recommend the most effective adjustment plan based on the user's emotional state, historical preferences and situational factors; S4. Interactive intervention implementation feedback: Guide users to perform regulatory behaviors through natural interaction and collect feedback data; S5. System Optimization and Model Iteration: Based on user feedback and long-term data, improve the accuracy of emotion recognition and the effectiveness of intervention; S6. Privacy Protection and Security Mechanisms: Ensure the security of user data throughout its entire lifecycle and comply with ethical standards.

9. The cloud-based intelligent assistance method for emotion regulation according to claim 8, characterized in that, The multimodal data acquisition and processing in S1 includes the following steps: S1-1 Facial Data Acquisition: Use a camera to capture facial expressions in real time and extract the movement trajectory of key points (such as the corners of the eyes and mouth); S1-2, Voice Data Acquisition: Record the voice signal through a microphone, segment it into short time frames and extract acoustic features; S1-3, Physiological Data Acquisition: Heart rate variability and skin electrical activity signals are collected via a smart bracelet; S1-4. Text Data Acquisition: Acquire user text content through keyboard input or speech-to-text conversion, segment words, and remove stop words.

10. The cloud-based intelligent assistance method for emotion regulation according to claim 8, characterized in that: The interactive intervention implementation feedback in S4 includes the following steps: S4-1, Multi-channel interaction: Voice interaction: The intelligent agent guides the user with empathetic language; Visual interaction: AR / VR scene simulation of natural environment, combined with breathing guidance animation; Tactile feedback: The smart bracelet vibrates to indicate breathing rhythm; S4-2. Real-time effect monitoring: During the intervention, facial, voice and physiological data are continuously collected to assess emotional changes. If the effect is not good, the backup strategy is automatically switched. S4-3. User feedback collection: After the intervention, subjective evaluations are collected through scales or open-ended questions.

Images