A robot micro-expression-triggered instant empathetic reply generation method

Abstract

This invention relates to the field of artificial intelligence technology and discloses a method for generating instant empathetic responses for robots based on micro-expression triggers. The method includes collecting facial micro-expression data of the interacting object, generating standardized micro-expression data through preprocessing, generating feature vector data based on micro-expression feature extraction, generating classification data by combining emotional state analysis, generating trigger signals for empathetic responses, generating empathetic response text data using natural language generation technology, and outputting instant empathetic responses through speech synthesis. Simultaneously, the micro-expression analysis model is optimized based on interaction feedback data. This invention ensures the reliability of the data foundation for generating empathetic response trigger signals, guarantees the real-time performance and reliability of interactive responses, and improves the friendliness and overall effect of robot interaction.

Description

Technical Field

[0001] This invention relates to the field of artificial intelligence technology, specifically to a method for generating instant empathetic responses for robots based on micro-expression triggers. Background Technology

[0002] Artificial intelligence is a new technological science that studies, develops, and applies theories, methods, technologies, and application systems to simulate, extend, and expand human intelligence. As a branch of computer science, artificial intelligence attempts to understand the essence of intelligence and produce a new kind of intelligent machine that can react in a way similar to human intelligence. Research in this field includes robotics, speech recognition, image recognition, natural language processing, and expert systems.

[0003] Currently, due to various technical limitations in the process of generating instant empathetic responses based on micro-expressions, the sensors equipped with them are affected by ambient lighting, facial occlusion, and individual differences when collecting micro-expression data. They cannot detect in real time whether there is data distortion and noise interference in key areas of micro-expressions. When the quality of the collected micro-expression data is low, it will cause large errors in the extraction of emotional features and cannot guarantee the accuracy of emotional state analysis.

[0004] Therefore, a method for generating instant empathetic responses for robots based on micro-expression triggers is proposed to solve the above problems. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this invention provides a robot-based instant empathic response generation method triggered by micro-expressions. This method solves the problems mentioned in the background section, such as the inability to detect data distortion and noise interference in key areas of micro-expressions in real time, and the inability to guarantee the accuracy of emotional state analysis.

[0006] To achieve the above objectives, the present invention provides the following technical solution: a method for generating instant empathetic responses for robots based on micro-expression triggers, the method comprising the following steps: S1. Collect facial micro-expression data of the robot's interaction objects; S2. Preprocess the facial micro-expression data to generate standardized micro-expression data; S3. Based on the standardized micro-expression data, perform micro-expression feature extraction processing to generate micro-expression feature vector data; S4. Perform emotional state analysis and processing based on the micro-expression feature vector data to generate emotional state classification data; S5. Based on the emotional state classification data, perform empathy response trigger judgment processing to generate an empathy response trigger signal; S6. Based on the empathy response trigger signal and the emotional state classification data, perform real-time empathy response content generation processing to generate empathy response text data; S7. Perform speech synthesis and output processing on the empathetic response text data to generate the robot's instant empathetic response; S8. Based on the interactive feedback data of the instant empathy response, perform model optimization processing to generate an updated micro-expression analysis model.

[0007] Preferably, the step S1 of collecting facial micro-expression data of the robot's interaction object includes the following steps: S11. Collect facial video stream data of the interactive object during the dialogue process using the high-definition camera mounted on the robot; S12. Perform frame extraction processing on the facial video stream data to generate facial image sequence data; S13. Perform micro-expression region localization processing based on the facial image sequence data to generate facial micro-expression data containing key micro-expression regions, wherein the key micro-expression regions include the eyebrows, corners of the mouth, and facial muscle regions. The localization accuracy evaluation formula is as follows: ; in, The positioning accuracy score is given. These are the actual coordinates of the feature points. To predict coordinates, For tolerance parameters, Given the total number of feature points, generate facial micro-expression data containing key micro-expression regions, including the eyebrows, corners of the mouth, and facial muscle areas.

[0008] Preferably, the preprocessing of the facial micro-expression data in step S2 to generate standardized micro-expression data includes the following steps: S21. Obtain the facial micro-expression data; S22. Perform image denoising and illumination normalization on the facial micro-expression data to generate preprocessed facial image data; S23. Based on the preprocessed facial image data, perform size standardization and alignment processing to generate standardized micro-expression data, wherein the standardization processing includes adjusting the image resolution to a uniform pixel size and aligning based on facial key points.

[0009] Preferably, step S3, which involves extracting micro-expression features based on the standardized micro-expression data to generate micro-expression feature vector data, includes the following steps: S31. Obtain the standardized micro-expression data; S32. Use a convolutional neural network model to extract features from the standardized micro-expression data to generate preliminary micro-expression feature map data; S33. The preliminary micro-expression feature map data is subjected to pooling and dimensionality reduction processing to generate micro-expression feature vector data, wherein the micro-expression feature vector data includes multi-dimensional feature values ​​representing the intensity, duration and type of micro-expression, and the feature fusion formula is as follows: ; in, As a feature of fusion, The weight of the k-th feature. The coefficient of the interaction term. To generate micro-expression feature vector data through element-wise multiplication.

[0010] Preferably, step S4, which involves performing emotional state analysis based on the micro-expression feature vector data to generate emotional state classification data, includes the following steps: S41. Obtain the micro-expression feature vector data; S42. Input the micro-expression feature vector data into a pre-trained emotion classification model to perform emotion state analysis and generate emotion probability distribution data; S43. Based on the emotional probability distribution data, perform threshold judgment processing to generate emotional state classification data, wherein the emotional state classification includes positive, negative and neutral states, and the corresponding specific emotional categories include joy, sadness and anger.

[0011] Preferably, step S5, which involves performing empathy response trigger judgment processing based on the emotional state classification data to generate an empathy response trigger signal, includes the following steps: S51. Obtain the emotional state classification data; S52. Based on the preset emotional intensity threshold and micro-expression change rate, determine the triggering conditions and generate an initial triggering signal; S53. Combine historical interaction data to perform context verification processing on the initial trigger signal to generate an empathy response trigger signal, wherein the empathy response trigger signal indicates whether to generate an empathy response immediately.

[0012] Preferably, in step S6, the instant empathy response content generation process based on the empathy response trigger signal and the emotional state classification data, to generate empathy response text data, includes the following steps: S61. Obtain the empathy response trigger signal and the emotional state classification data; S62. Based on the emotional state classification data, retrieve matching similar response template data from the response template library; S63. Use a natural language generation model to personalize the empathy response template data and generate empathy response text data, wherein the personalization includes adding the name of the interaction object and emotional resonance phrases.

[0013] Preferably, step S7, which involves speech synthesis and output processing of the empathetic response text data to generate the robot's instant empathetic response, includes the following steps: S71. Obtain the similar reply text data; S72. Perform speech synthesis processing on the similar response text data to generate speech audio data; S73. Output the voice audio data through the robot's speaker and simultaneously display the text reply to generate the robot's instant empathetic reply.

[0014] Preferably, the step S8, which involves optimizing the model based on the interactive feedback data of the immediate empathetic response to generate an updated micro-expression analysis model, includes the following steps: S81. Collect the interactive feedback data of the instant empathetic response, including the facial reaction and voice response data of the interactive object; S82. Based on the interactive feedback data, perform effect evaluation processing to generate model optimization index data; S83. Using a reinforcement learning algorithm, update the parameters of the micro-expression analysis model based on the model optimization index data to generate the updated micro-expression analysis model.

[0015] Preferably, the method further includes the following steps: S91. Monitor the performance data of the updated micro-expression analysis model in real time; S92. Based on performance data, adaptive adjustment processing is performed to optimize the efficiency of generating similar responses and ensure real-time responsiveness.

[0016] Compared with existing technologies, this invention provides a method for generating instant empathetic responses for robots based on micro-expression triggers, which has the following beneficial effects: 1. In this invention, when generating instant empathic responses for robots based on micro-expression triggers, the standardized preprocessing of facial micro-expression data and the extraction of micro-expression feature vector data are used to detect and correct data distortion caused by ambient lighting and individual differences in real time, ensuring the accuracy of feature extraction of key micro-expression regions, thereby reducing the analysis error of emotional state classification data, improving the accuracy of emotional state classification, and further ensuring the reliability of the data basis for generating empathic response trigger signals.

[0017] 2. In this invention, when generating instant empathic responses for robots based on micro-expression triggers, the real-time verification of emotional state classification data and the dynamic judgment mechanism of empathic response trigger signals are used to avoid misjudgment caused by subtle changes in micro-expressions. This makes the generation of empathic response trigger signals more timely and accurate. Furthermore, when the triggering conditions are abnormal, the micro-expression analysis model can be adaptively adjusted to quickly correct the situation, ensuring the real-time performance and reliability of the interactive response.

[0018] 3. In this invention, when generating instant empathetic responses for the robot based on micro-expression triggers, the generation process of empathetic response text data is dynamically adjusted by combining multi-dimensional emotion adaptation and personalized response optimization with the dialogue context. This achieves a high degree of matching between instant empathetic responses and the user's actual emotional needs, reduces the risk of deviation in response content, and thus enhances the friendliness and overall effect of robot interaction. Attached Figure Description

[0019] Figure 1 This is a flowchart illustrating the steps of a robot-based instant empathy response generation method triggered by micro-expressions, as described in this invention. Detailed Implementation

[0020] 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.

[0021] Please see Figure 1 The specific implementation of a robot-based instant empathy response generation method based on micro-expression triggers is as follows, and the method includes the following steps: S1. Collect facial micro-expression data of the robot's interaction objects; S2. Preprocess the facial micro-expression data to generate standardized micro-expression data; S3. Based on standardized micro-expression data, perform micro-expression feature extraction and processing to generate micro-expression feature vector data; S4. Perform emotional state analysis and processing based on micro-expression feature vector data to generate emotional state classification data; S5. Based on the emotional state classification data, perform empathy response trigger judgment processing to generate empathy response trigger signals; S6. Based on the empathy response trigger signal and emotional state classification data, perform real-time empathy response content generation processing to generate empathy response text data; S7. Perform speech synthesis and output processing on the empathetic response text data to generate the robot's instant empathetic response; S8. Based on the interactive feedback data of immediate empathetic responses, optimize the model to generate an updated micro-expression analysis model.

[0022] Step S1 involves collecting facial micro-expression data of the robot's interactive objects, including the following steps: S11. Collect facial video stream data of the interactive object during the dialogue process using the high-definition camera mounted on the robot; S12. Perform frame extraction processing on the facial video stream data to generate facial image sequence data; S13. Perform micro-expression region localization processing based on facial image sequence data to generate facial micro-expression data containing key micro-expression regions, including the eyebrows, corners of the mouth, and facial muscle regions. The localization accuracy evaluation formula is as follows: ; in, The positioning accuracy score is given. These are the actual coordinates of the feature points. To predict coordinates, For tolerance parameters, Given the total number of feature points, generate facial micro-expression data containing key micro-expression regions, including the eyebrows, corners of the mouth, and facial muscle areas.

[0023] Step S2 involves preprocessing facial micro-expression data to generate standardized micro-expression data, including the following steps: S21. Obtain facial micro-expression data; S22. Perform image denoising and illumination normalization on the facial micro-expression data to generate preprocessed facial image data. The illumination compensation formula is as follows: ; in, The image after normalization. For the original image, and The mean and standard deviation of a local region. and Target mean and standard deviation; generate preprocessed facial image data; S23. Based on the preprocessed facial image data, perform size standardization and alignment processing to generate standardized micro-expression data, wherein the standardization processing includes adjusting the image resolution to a uniform pixel size and aligning based on facial key points.

[0024] Step S3 involves extracting micro-expression features from standardized micro-expression data to generate micro-expression feature vector data, including the following steps: S31. Obtain standardized micro-expression data; S32. A convolutional neural network model is used to extract features from standardized micro-expression data. The convolutional neural network model contains three convolutional layers with kernel sizes of 5x5, 3x3, and 3x3, respectively. Each convolutional layer is followed by a ReLU activation function and a 2x2 max pooling layer. The model is pre-trained using the FER2013 dataset. During training, the cross-entropy loss function and the Adam optimizer are used, and the learning rate is set to 0.001 to generate preliminary micro-expression feature map data. S33. Perform pooling and dimensionality reduction processing on the preliminary micro-expression feature map data to generate micro-expression feature vector data. The micro-expression feature vector data includes multi-dimensional feature values ​​representing the intensity, duration, and type of micro-expressions. The feature fusion formula is as follows: ; in, As a feature of fusion, The total number of features, For feature index, The weight of the k-th feature. The coefficient of the interaction term. To generate micro-expression feature vector data through element-wise multiplication, This is the interaction term between the k-th feature and the j-th feature.

[0025] Step S4 involves performing emotional state analysis based on the micro-expression feature vector data to generate emotional state classification data, including the following steps: S41. Obtain micro-expression feature vector data; S42. Input the micro-expression feature vector data into the pre-trained sentiment classification model for sentiment state analysis. The sentiment classification model adopts the support vector machine algorithm, the kernel function is the radial basis function, and the training data comes from the AffectNet dataset, which contains eight sentiment labels. During training, the hyperparameters are optimized through cross-validation to generate sentiment probability distribution data. The formula for calculating its emotional confidence is: ; in, For emotional confidence, For the i-th sentiment score, For the j-th type of sentiment score, For the number of emotion categories, Generate sentiment probability distribution data based on the mean sentiment score; S43. Based on the emotional probability distribution data, threshold judgment processing is performed to generate emotional state classification data, where the emotional state classification includes positive, negative and neutral states, and the corresponding specific emotional categories include joy, sadness and anger.

[0026] Step S5, which involves processing the empathy response trigger based on the emotion state classification data to generate an empathy response trigger signal, includes the following steps: S51. Obtain emotional state classification data; S52. Based on the preset emotional intensity threshold and micro-expression change rate, the triggering condition is judged, and an initial triggering signal is generated. The triggering decision function is: ; in, To trigger the decision value, For emotional intensity, For emotional threshold, For the rate of change, , For weight parameters, It is a sigmoid function. The hyperbolic tangent function is used to generate an initial trigger signal; S53. Combine historical interaction data to perform contextual verification processing on the initial trigger signal and generate an empathy response trigger signal, wherein the empathy response trigger signal indicates whether to generate an empathy response immediately.

[0027] Step S6 involves generating real-time empathic response content based on empathic response trigger signals and emotional state classification data. This generation of empathic response text data includes the following steps: S61. Obtain empathy response trigger signals and emotional state classification data; S62. Based on the emotional state classification data, retrieve matching similar response template data from the response template library. The response template library is manually constructed and contains text templates for different emotional states. The template structure includes fixed sentence patterns and variable placeholders. S63. A natural language generation model is used to personalize the empathy response template data. The natural language generation model is based on a rule engine and adds the name of the interaction object and emotional resonance phrases by replacing placeholders. The emotional resonance phrases are matched from the NRC sentiment dictionary to generate empathy response text data. The personalization includes adding the name of the interaction object and emotional resonance phrases. The text quality evaluation formula is as follows: ; in, A text quality score is given. , , These are respectively fluency, empathy, and relevance. , , These are the fluency function, the empathy function, and the relevance function, respectively. , , Using these as weighting coefficients, similar response text data is generated.

[0028] Step S7 involves speech synthesis and output processing of the empathetic response text data to generate the robot's instant empathetic response, including the following steps: S71. Obtain the similar reply text data; S72. Perform speech synthesis processing on the similar response text data to generate speech audio data; S73. Output voice audio data through the robot's speaker and simultaneously display text responses to generate the robot's instant empathetic response.

[0029] Step S8 involves optimizing the model based on the interactive feedback data from immediate empathetic responses to generate an updated micro-expression analysis model, including the following steps: S81. Collect interactive feedback data for real-time empathetic responses, including facial reactions and voice responses of the interactive object; S82. Based on the interactive feedback data, perform effect evaluation processing to generate model optimization index data. The model update rule is as follows: ; in, , For the parameters of the old and new models, For learning rate, This represents the partial derivative of the loss function with respect to the model parameters. For loss function, To provide feedback on reward values, model optimization metrics data are generated; S83. The micro-expression analysis model parameters are updated based on the model optimization index data using a reinforcement learning algorithm. The reinforcement learning algorithm adopts the Q-learning algorithm, the state space is the model parameter set, the action space is the parameter adjustment amount, and the reward function is calculated based on user satisfaction feedback to generate the updated micro-expression analysis model.

[0030] The method also includes the following steps: S91. Real-time monitoring of the performance data of the updated micro-expression analysis model; S92. Based on performance data, adaptive adjustment processing is performed to optimize the efficiency of generating similar responses and ensure real-time responsiveness.

[0031] The operation steps of a robot-based instant empathy response generation method triggered by micro-expressions are as follows: Step 1: Collect facial micro-expression data The method first uses a high-definition camera mounted on the robot to collect facial video stream data of the interactive object during the dialogue process, and performs frame extraction processing on the video stream data to generate facial image sequence data. Then, based on the facial image sequence data, micro-expression region localization processing is performed to generate facial micro-expression data containing key micro-expression regions, including eyebrows, corners of the mouth, and facial muscle regions. This step ensures the acquisition of raw data and provides a reliable foundation for subsequent processing.

[0032] Step 2: Preprocessing to generate standardized micro-expression data After acquiring facial micro-expression data, image denoising and illumination normalization are performed to generate preprocessed facial image data. Then, size standardization and alignment are performed on the preprocessed facial image data to generate standardized micro-expression data. The standardization process includes adjusting the image resolution to a uniform pixel size and aligning based on facial key points, thereby reducing environmental interference and improving data consistency.

[0033] Step 3: Micro-expression feature extraction and generation of micro-expression feature vector data Based on standardized micro-expression data, a convolutional neural network model is used for feature extraction to generate preliminary micro-expression feature map data. Then, pooling and dimensionality reduction processing are performed on the preliminary micro-expression feature map data to generate micro-expression feature vector data. This vector data contains multi-dimensional feature values ​​such as micro-expression intensity, duration, and type, providing structured input for sentiment analysis.

[0034] Step 4: Sentiment state analysis generates sentiment state classification data. Based on micro-expression feature vector data, the data is input into a pre-trained emotion classification model to analyze the emotion state and generate emotion probability distribution data. Based on the emotion probability distribution data, threshold judgment is performed to generate emotion state classification data, which includes positive, negative and neutral states, and corresponding to specific emotion categories such as joy, sadness and anger. This step realizes the recognition of emotion state.

[0035] Step 5: Generating an empathy response trigger signal based on the empathy response trigger judgment. Based on emotional state classification data, the system determines the triggering conditions by using a preset emotional intensity threshold and micro-expression change rate, generating an initial trigger signal. This initial trigger signal is then combined with historical interaction data for contextual verification, generating an empathetic response trigger signal. This signal indicates whether to generate an empathetic response immediately, ensuring the timeliness and accuracy of the triggering.

[0036] Step Six: Generate Immediate Empathic Response Content Generate empathic response text data. Based on the empathy response trigger signals and emotional state classification data, matching empathy response template data is retrieved from the response template library. A natural language generation model is used to personalize the template data and generate empathy response text data. Personalization includes adding the name of the interaction object and emotional resonance phrases to make the response content more approachable.

[0037] Step 7: Speech synthesis and output generation of instant empathetic responses The empathetic response text data is processed by speech synthesis to generate audio data. The audio data is then output through the robot's speaker, and the text response is displayed simultaneously to generate the robot's instant empathetic response. This step enables multimodal output and enhances the interactive experience.

[0038] Step 8: Model Optimization and Generation of Updated Micro-expression Analysis Model Based on interactive feedback data from real-time empathetic responses, facial and voice response data of the interactive object are collected, and effect evaluation is performed to generate model optimization index data. Reinforcement learning algorithms are used to update the micro-expression analysis model parameters based on the index data, generating an updated micro-expression analysis model. This step ensures continuous system optimization and improves its adaptability.

[0039] It should be noted that, in this document, relational terms such as "first" and "second" are used only 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.

[0040] 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 method for generating instant empathetic responses for robots based on micro-expression triggers, characterized in that, The method includes the following steps: S1. Collect facial micro-expression data of the robot's interaction objects; S2. Preprocess the facial micro-expression data to generate standardized micro-expression data; S3. Based on the standardized micro-expression data, perform micro-expression feature extraction processing to generate micro-expression feature vector data; S4. Perform emotional state analysis and processing based on the micro-expression feature vector data to generate emotional state classification data; S5. Based on the emotional state classification data, perform empathy response trigger judgment processing to generate an empathy response trigger signal; S6. Based on the empathy response trigger signal and the emotional state classification data, perform real-time empathy response content generation processing to generate empathy response text data; S7. Perform speech synthesis and output processing on the empathetic response text data to generate the robot's instant empathetic response; S8. Based on the interactive feedback data of the instant empathy response, perform model optimization processing to generate an updated micro-expression analysis model.

2. The method for generating instant empathetic responses for robots based on micro-expression triggers according to claim 1, characterized in that, The step S1 of collecting facial micro-expression data of the robot's interaction object includes the following steps: S11. Collect facial video stream data of the interactive object during the dialogue process using the high-definition camera mounted on the robot; S12. Perform frame extraction processing on the facial video stream data to generate facial image sequence data; S13. Perform micro-expression region localization processing based on the facial image sequence data to generate facial micro-expression data containing key micro-expression regions, wherein the key micro-expression regions include the eyebrows, corners of the mouth, and facial muscle regions. The localization accuracy evaluation formula is as follows: ； in, The positioning accuracy score is given. These are the actual coordinates of the feature points. To predict coordinates, For tolerance parameters, Given the total number of feature points, generate facial micro-expression data containing key micro-expression regions, including the eyebrows, corners of the mouth, and facial muscle areas.

3. The method for generating instant empathetic responses for robots based on micro-expression triggers according to claim 1, characterized in that, The preprocessing of the facial micro-expression data in step S2 to generate standardized micro-expression data includes the following steps: S21. Obtain the facial micro-expression data; S22. Perform image denoising and illumination normalization on the facial micro-expression data to generate preprocessed facial image data; S23. Based on the preprocessed facial image data, perform size standardization and alignment processing to generate standardized micro-expression data, wherein the standardization processing includes adjusting the image resolution to a uniform pixel size and aligning based on facial key points.

4. The method for generating instant empathetic responses for robots based on micro-expression triggers according to claim 1, characterized in that, Step S3, which involves extracting micro-expression features from the standardized micro-expression data to generate micro-expression feature vector data, includes the following steps: S31. Obtain the standardized micro-expression data; S32. Use a convolutional neural network model to extract features from the standardized micro-expression data to generate preliminary micro-expression feature map data; S33. The preliminary micro-expression feature map data is subjected to pooling and dimensionality reduction processing to generate micro-expression feature vector data, wherein the micro-expression feature vector data includes multi-dimensional feature values ​​representing the intensity, duration and type of micro-expression, and the feature fusion formula is as follows: ； in, As a feature of fusion, The weight of the k-th feature. The coefficient of the interaction term. To generate micro-expression feature vector data through element-wise multiplication.

5. A method for generating instant empathetic responses for robots based on micro-expression triggers according to claim 1, characterized in that, Step S4, which involves performing emotional state analysis based on the micro-expression feature vector data to generate emotional state classification data, includes the following steps: S41. Obtain the micro-expression feature vector data; S42. Input the micro-expression feature vector data into a pre-trained emotion classification model to perform emotion state analysis and generate emotion probability distribution data; S43. Based on the emotional probability distribution data, perform threshold judgment processing to generate emotional state classification data, wherein the emotional state classification includes positive, negative and neutral states, and the corresponding specific emotional categories include joy, sadness and anger.

6. The method for generating instant empathetic responses for robots based on micro-expression triggers according to claim 1, characterized in that, The step S5, which involves performing empathy response trigger judgment processing based on the emotional state classification data to generate an empathy response trigger signal, includes the following steps: S51. Obtain the emotional state classification data; S52. Based on the preset emotional intensity threshold and micro-expression change rate, determine the triggering conditions and generate an initial triggering signal; S53. Combine historical interaction data to perform context verification processing on the initial trigger signal to generate an empathy response trigger signal, wherein the empathy response trigger signal indicates whether to generate an empathy response immediately.

7. A method for generating instant empathetic responses for robots based on micro-expression triggers according to claim 1, characterized in that, In step S6, the instant empathy response content generation process is performed based on the empathy response trigger signal and the emotional state classification data. The generation of empathy response text data includes the following steps: S61. Obtain the empathy response trigger signal and the emotional state classification data; S62. Based on the emotional state classification data, retrieve matching similar response template data from the response template library; S63. Use a natural language generation model to personalize the empathy response template data and generate empathy response text data, wherein the personalization includes adding the name of the interaction object and emotional resonance phrases.

8. A method for generating instant empathetic responses for robots based on micro-expression triggers according to claim 1, characterized in that, The step S7, which involves speech synthesis and output processing of the empathetic response text data to generate the robot's instant empathetic response, includes the following steps: S71. Obtain the similar reply text data; S72. Perform speech synthesis processing on the similar response text data to generate speech audio data; S73. Output the voice audio data through the robot's speaker and simultaneously display the text reply to generate the robot's instant empathetic reply.

9. A method for generating instant empathetic responses for robots based on micro-expression triggers according to claim 1, characterized in that, The step S8, which optimizes the model based on the interactive feedback data of the immediate empathetic response to generate an updated micro-expression analysis model, includes the following steps: S81. Collect the interactive feedback data of the instant empathetic response, including the facial reaction and voice response data of the interactive object; S82. Based on the interactive feedback data, perform effect evaluation processing to generate model optimization index data; S83. Using a reinforcement learning algorithm, update the parameters of the micro-expression analysis model based on the model optimization index data to generate the updated micro-expression analysis model.

10. A method for generating instant empathetic responses for robots based on micro-expression triggers according to claim 1, characterized in that, The method further includes the following steps: S91. Monitor the performance data of the updated micro-expression analysis model in real time; S92. Based on performance data, adaptive adjustment processing is performed to optimize the efficiency of generating similar responses and ensure real-time responsiveness.

Images