Emotion evaluation method and device, equipment and medium

Abstract

The invention provides an emotion evaluation method and device, equipment and a medium. The method comprises the following steps: receiving interaction feedback information of a user for an emotion evaluation problem of a current round; generating an interactive input text of the current round based on the interactive feedback information, historical information interacting with the user and prompt information of emotion evaluation; inputting the interaction input text of the current round into the large language model to obtain an emotion evaluation problem of the next round output by the large language model, and returning and receiving interaction feedback information of the user for the emotion evaluation problem of the next round until interaction is finished; and performing emotion evaluation based on the interaction feedback information of each round. According to the emotion evaluation method provided by the invention, the emotion evaluation problem is automatically generated through the large language model, man-machine interaction is carried out, and the dependence on professional evaluation personnel is reduced. And through an interactive feedback mechanism, the questioning content can be adjusted in real time according to the emotion of the user, and personalized requirements are met.

Description

Technical Field

[0001] The present invention relates to the field of artificial intelligence technology, and in particular to an emotion assessment method, device, equipment and medium. Background Art

[0002] Emotional assessment is widely used in the fields of mental health, education and social services. It is an effective means to help understand an individual's emotional state and identify psychological needs. Through accurate emotional assessment, professionals can better provide emotional support and behavioral intervention to help individuals manage their emotions and improve their mental state.

[0003] However, existing emotion assessment methods mostly rely on standardized questions and templated dialogues, which lack flexibility and make it difficult to provide personalized feedback based on individual needs. This not only reduces the accuracy of the assessment, but also makes it difficult for the assessed to gain understanding and support. At the same time, since the assessment process relies on the operation of professional assessors, but the number of professional assessors is tight and the workload is high, the assessment efficiency is limited and it is difficult to meet a wide range of needs. Summary of the invention

[0004] The present invention provides an emotion assessment method, device, equipment and medium to solve the defects of the related technology such as lack of personalized feedback and suggestions, low assessment efficiency and large occupation of medical resources.

[0005] The present invention provides a method for evaluating emotions, comprising the following steps: Receive interactive feedback from users regarding the current round of emotion assessment questions; Generate the interactive input text of the current round based on the interactive feedback information, the historical information of the interaction with the user, and the prompt information of the emotion evaluation; Inputting the interactive input text of the current round into the large language model, obtaining the emotion assessment question of the next round output by the large language model, and returning and receiving the interactive feedback information of the user for the emotion assessment question of the next round until the interaction ends; Emotional assessment is performed based on the interactive feedback information of each round.

[0006] According to an emotion assessment method provided by the present invention, the prompt information includes at least one of scene description information of the emotion assessment, identity limitation information of the emotion assessment, and output requirement information of the emotion assessment.

[0007] According to an emotion assessment method provided by the present invention, the output requirement information includes requirement information of at least one of inquiry output, empathy output, guidance output, and suggestion output for the user.

[0008] According to a method for evaluating emotions provided by the present invention, the interactive feedback information includes interactive text, interactive voice and / or interactive video.

[0009] According to a method for emotion assessment provided by the present invention, the emotion assessment is performed based on the interactive feedback information of each round, including: For each round of interactive feedback information, performing text emotion scoring on the interactive text in the interactive feedback information to obtain a text emotion score, and performing emotion recognition on the interactive voice and / or interactive video in the interactive feedback information to obtain a voice emotion recognition result and / or a video emotion recognition result; Determining an emotion evaluation result of the interactive feedback information based on the emotion consistency between the speech emotion recognition result and the text emotion score, and / or the emotion consistency between the video emotion recognition result and the text emotion score; Emotional evaluation is performed based on the emotional evaluation results of the interactive feedback information in each round.

[0010] According to an emotion assessment method provided by the present invention, the emotion assessment result of the interactive feedback information is determined based on the emotion consistency between the speech emotion recognition result and the text emotion score, and / or the emotion consistency between the video emotion recognition result and the text emotion score, including: In the case where the emotional consistency is consistent, determining an emotional evaluation result of the interactive feedback information based on the text emotional score; In a case where the emotion consistency is inconsistent, an emotion evaluation result of the interactive feedback information is determined based on the speech emotion recognition result and / or the video emotion recognition result.

[0011] According to a method for emotion assessment provided by the present invention, emotion assessment is performed based on the interactive feedback information of each round, and then the method further includes: The emotional state obtained from the emotional assessment and the interaction feedback information of each round are sent to the audit end, and the audit result for the emotional state returned by the audit end is received.

[0012] The present invention also provides an emotion assessment device, comprising the following modules: An input text generating unit, configured to generate the interactive input text of the current round based on the interactive feedback information, the historical information of the interaction with the user, and the prompt information of the emotion evaluation; An evaluation question generating unit, configured to input the interactive input text of the current round into a large language model, obtain the emotion evaluation question of the next round output by the large language model, and return and receive the interactive feedback information of the user for the emotion evaluation question of the next round until the interaction ends; The emotion evaluation unit is used to perform emotion evaluation based on the interactive feedback information of each round.

[0013] The present invention also provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, any of the above-mentioned emotion assessment methods is implemented.

[0014] The present invention also provides a non-transitory computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the emotion assessment method described above is implemented.

[0015] The present invention also provides a computer program product, comprising a computer program, wherein when the computer program is executed by a processor, the emotion assessment method described above is implemented.

[0016] The emotion assessment method, device, equipment and medium provided by the present invention automatically generate emotion assessment questions through a large language model and conduct human-computer interaction, thereby reducing the reliance on professional assessors, alleviating resource shortages and improving assessment efficiency. Through the interactive feedback mechanism, the flexibility of emotion assessment is improved, so that the content of questions can be adjusted in real time according to the user's emotions, meet personalized needs, and increase the accuracy of assessment. At the same time, personalized empathy feedback is provided to make users feel understood and supported, improving the assessment experience. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the technical solutions in the present invention or related technologies, the following briefly introduces the drawings required for use in the embodiments or related technical descriptions. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0018] Figure 1 This is one of the flow charts of the emotion assessment method provided by the present invention.

[0019] Figure 2 It is a flowchart of the multimodal emotion assessment provided by the present invention.

[0020] Figure 3 This is the second flow chart of the emotion assessment method provided by the present invention.

[0021] Figure 4 It is a structural schematic diagram of the emotion assessment device provided by the present invention.

[0022] Figure 5 It is a structural schematic diagram of the electronic device provided by the present invention. DETAILED DESCRIPTION

[0023] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with the drawings of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0024] Emotional assessment plays an important role in the fields of mental health, education and social services. It not only helps professionals identify the emotional state of individuals, but also provides a scientific basis for targeted intervention and support. Through effective emotion assessment, professionals can have a deeper understanding of the psychological needs of individuals, and then develop personalized intervention measures to help people better manage their emotions and improve their mental health. Emotional assessment is an indispensable tool in psychotherapy, emotional support and interpersonal communication. Accurate emotion assessment can promote better communication and interaction, help build trusting relationships, and thus enhance the effectiveness of psychological support.

[0025] However, existing emotion assessment methods have many problems in practical applications. Many assessment methods rely on fixed questions and templated dialogues, lack flexibility, and cannot be adjusted individually according to the actual emotional state of the person being assessed. This lack of personalized feedback often makes it difficult for the person being assessed to fully express their inner emotions, which in turn affects the accuracy and effectiveness of the assessment. In addition, the implementation of emotion assessment usually relies on the judgment and operation of professional assessors. The shortage of professional assessors and the high workload also restrict the efficiency of the assessment, making it difficult to meet the growing demand. These factors together result in existing emotion assessment methods being insufficiently comprehensive and inefficient when dealing with complex emotions.

[0026] In view of the above problems, an embodiment of the present invention provides a method for emotion assessment. Figure 1 FIG. 1 is one of the flow charts of the emotion assessment method provided by the present invention. Figure 1 As shown, the method includes: Step 110: receiving interactive feedback information from the user regarding the emotion assessment question of the current round.

[0027] Specifically, in the emotion assessment process, a "round" refers to a specific question and answer, that is, a complete process of asking a user an emotion assessment question and the user responding to it. In each round of interaction, different interactive feedback information from the user will be generated. Through this interactive feedback information, the subsequent emotion assessment questions can be gradually adjusted to better capture the user's emotional state.

[0028] The interactive feedback information of each round includes various feedback data of the user in that round, such as data in the form of voice, text and video. Voice feedback refers to the emotions and reactions expressed by the user through voice, in which the tone, speaking speed and volume changes can convey emotional characteristics. Text feedback includes the text content converted from voice and the user's answer or choice to the emotion assessment question. These text information provide a direct reference for emotion assessment. Video feedback uses hardware such as cameras to capture non-verbal signals such as facial expressions, eyes and postures of users to identify subtle changes in emotional state. These interactive feedback information constitutes a multimodal data source, providing a comprehensive foundation for emotion assessment and helping to improve the accuracy and reliability of assessment.

[0029] Step 120: Generate the interactive input text of the current round based on the interactive feedback information, the historical information of the interaction with the user, and the prompt information of the emotion evaluation.

[0030] Specifically, the generation of the next round of emotion assessment questions needs to comprehensively consider the interaction feedback information of the current round, the historical information of user interaction, and the prompt information of emotion assessment. Historical interaction information refers to the feedback provided by users in previous rounds, including emotional expression, voice intonation, facial expression reaction, and answer content. This information helps to understand the changing trend of user emotions and the continuity of the current state, thereby providing contextual support for the generation of subsequent emotion assessment questions. For example, if the user shows anxiety in the previous rounds, the next questions can be more in line with this emotional background to help the user continue to express his true feelings. The prompt information of emotion assessment is a set of guided content that has been trained and set. The prompt information contains elements such as specific interaction scenarios, identity restrictions, historical dialogue data, and output templates, so that the question content is more in line with humanized interaction needs. In specific situations, the prompt information will dynamically generate different degrees of inquiry, empathy, guidance, and suggestions according to the user's emotional state to support a personalized assessment experience, help users express emotions more naturally, and create a trusting and comfortable communication atmosphere.

[0031] The interactive input text is a set of text content used to input into the large language model, including the guide words generated based on user feedback in the current round. The interactive input text not only contains key information about the current user's emotional state, but also combines historical interaction data to provide the necessary context for the next round of evaluation questions generated, making the conversation more coherent and targeted. This text will be dynamically adjusted based on the user's emotional feedback while maintaining the evaluation framework. For example, according to the user's expressed anxiety or depression, a more appropriate way of asking questions and words will be selected. The interactive input text also contains appropriate empathetic language or encouraging sentences to guide users to share their emotional experiences more deeply. After the large language model receives the interactive input text, it will automatically generate the next round of questions that match the situation, further promoting the in-depth emotional evaluation, thereby continuously adjusting the interactive content throughout the process to ensure the comprehensiveness of the evaluation and the comfort of the user experience.

[0032] Step 130, input the interactive input text of the current round into the large language model, obtain the emotion assessment questions of the next round output by the large language model, and return to receive the interactive feedback information of the user for the emotion assessment questions of the next round until the interaction ends.

[0033] Specifically, based on the interactive input text of the current round, the large language model uses its natural language processing capabilities to analyze and generate sentiment assessment questions. The large language model is a deep learning model trained with massive text data, with billions to hundreds of billions of parameters. These parameters enable the model to understand complex language structures, semantic relationships, and various language patterns, and has strong language understanding, context association, and generation capabilities.

[0034] In emotion assessment, the large language model first uses its language understanding ability to decode the user's emotional information in the interactive input text and identify key emotional vocabulary. For example, when a user mentions "I'm a little frustrated recently" in feedback, the large voice model is able to identify "frustrated" as an emotional keyword and use it as the core information for generating the next round of questions. At the same time, the context association ability of the large model enables it to associate the interactive input text of the current round with the previous round to generate coherent and logical questions. If the user mentioned "pressure" before, the model can associate this content and generate questions like "Can you tell me what caused these pressures?" to ensure that the conversation is smooth and natural. The generation ability of the large language model enables it to automatically generate appropriate emotion assessment questions based on user feedback, guiding users to gradually unfold their emotions and express their emotions in depth.

[0035] In multiple rounds of emotion assessment cycles, the language understanding, context association, and generation capabilities of the large language model are combined to enable it to gradually generate accurate emotion assessment questions. After each user answer, the model will process the new feedback again and continue to generate targeted emotion assessment questions, iterating continuously until enough information is collected to complete the emotion assessment. Through these capabilities of the large model, the depth and flexibility of emotion assessment are significantly enhanced.

[0036] Step 140: Perform emotion assessment based on the interactive feedback information of each round.

[0037] By comprehensively analyzing the interactive feedback information of each round in the entire evaluation process, including multimodal information such as voice, text and video, the overall emotional state of the user can be deeply portrayed, and the emotional evaluation result can be obtained. The emotional evaluation result here can be an evaluation result of depression or an evaluation result of anxiety, which is not limited in the embodiment of the present invention. For example, after obtaining the evaluation result of depression, it can be synchronized to a professional evaluator to assist the professional evaluator in diagnosing the user's emotional state.

[0038] After integrating the feedback information from each round of interaction, the user's overall emotional state can be derived and quantified or classified for analysis. For example, the analysis process can identify the user's emotional trends throughout the assessment (such as persistent anxiety, gradually improving emotional state, etc.), as well as subtle changes in emotions during each round of interaction. The resulting emotional assessment results provide professionals with objective, quantitative reference data, and also lay a data foundation for subsequent emotional management and psychological support.

[0039] The emotion assessment method provided by the present invention automatically generates emotion assessment questions through a large language model and conducts human-computer interaction, thereby reducing the reliance on professional assessors, alleviating resource shortages, and improving assessment efficiency. Through the interactive feedback mechanism, the flexibility of emotion assessment is improved, so that the content of questions can be adjusted in real time according to the user's emotions, meeting personalized needs, and increasing the accuracy of assessment.

[0040] Based on the above embodiment, the prompt information includes at least one of the scene description information of the emotion assessment, the identity limitation information of the emotion assessment, and the output requirement information of the emotion assessment.

[0041] Specifically, scenario description information refers to providing corresponding settings according to the specific situation of the user during the emotion assessment process. For example, the large language model will flexibly adjust the content of the questions according to the user's current situation (such as stressful situations in the workplace, interpersonal interactions in family life, or emotional states in social activities). In this way, the assessment questions can be more targeted, allowing users to express emotions closely related to the current situation more naturally and realistically, thereby improving the comprehensiveness and authenticity of the assessment.

[0042] Identity-restricted information sets restrictions on emotion assessment questions based on the user's role or specific background information. Different identity backgrounds will bring different emotional experiences. For example, students, professionals, or family members face different sources of emotional stress, psychological needs, and ways of expression. By introducing identity-restricted information, the emotion assessment process can be closer to the user's personal characteristics and generate personalized questions that are more in line with their identity characteristics, thereby effectively improving the accuracy of the assessment and the user's sense of trust.

[0043] Output requirement information sets requirements for the output types in the evaluation process, including requirements for interactive types such as inquiry output, empathy output, guidance output, and suggestion output. For example, depending on the evaluation needs, the evaluation process may focus on generating more empathy output to express understanding and support, or emphasize guidance output to guide users to further express their emotions. In addition, output requirements can also set suggestion outputs to provide supportive suggestions in appropriate situations. Through the reasonable configuration of these output types, prompt information can optimize the evaluation interaction and allow users to experience a more personalized and caring evaluation process.

[0044] Through the mutual coordination of scene description information, identity limitation information and output requirement information, the emotion assessment process has higher personalization and flexible adaptability, thereby significantly improving the accuracy of emotion assessment and user experience.

[0045] Based on the above embodiment, the output requirement information includes requirement information of at least one of inquiry output, empathy output, guidance output, and suggestion output to the user.

[0046] Specifically, inquiry output guides users to express their inner emotional state through appropriate questions during the emotion assessment process. It can generate targeted questions based on the user's current emotions, help tap into emotions and psychological needs, and make the emotion assessment more in-depth. For example, further inquiries can allow users to describe their feelings in depth, providing richer information for the assessment. Empathy output expresses understanding and care in the assessment interaction, making users feel understood and supported. Through language, tone or text responses, users can feel emotional responses, which not only makes users feel comfortable in the interaction, but also encourages them to express their emotions more openly and authentically. Guidance output helps users explore and express their inner emotions through guiding discourse, encouraging users to further explore the emotional level and make the expression more comprehensive. For example, guide users to review events related to emotions, or describe possible emotional causes to obtain more complete information. Suggestion output provides users with supportive suggestions in appropriate situations, providing direction for emotion management and regulation, which may include comforting language, simple emotion management techniques or relaxation suggestions. This output allows users to get practical psychological support in the assessment and enhances the practicality and experience of emotion assessment. Through the reasonable configuration of inquiry, empathy, guidance and suggestion output, the output requirement information significantly improves the effect of emotion assessment and interactive experience, making the emotion assessment process more personalized and humane.

[0047] In the traditional emotion assessment process, emotion assessment relies solely on text feedback, and the person being assessed may conceal or cover up their true emotions, resulting in distorted assessment results. Therefore, it is necessary to combine non-text information such as voice and video to fully capture the user's true emotional state and improve the accuracy and objectivity of the assessment.

[0048] Based on the above embodiment, the interactive feedback information includes interactive text, interactive voice and / or interactive video.

[0049] First, interactive text includes text content directly input by users and text content generated by speech-to-text technology. These texts not only contain the user's subjective emotional expression, but also record the specific responses to emotional assessment questions, which facilitates the analysis of emotional information. For example, the user directly inputs "I feel very tired today" or the text converted by speech "I have always been a little depressed recently" can provide direct data support for subsequent emotional assessment. These two text channels help capture the user's emotional content, especially when the user tends to use verbal expression, speech-to-text further enriches the source of text feedback.

[0050] Secondly, interactive speech is the feedback information input by the user through voice, which not only records the voice content, but also includes voice features such as pitch, speaking speed, and volume. By analyzing these features, we can further understand the intensity and fluctuation of users' emotions. For example, a slightly trembling voice may indicate tension or anxiety, while a slowed speech speed may indicate depression. Voice data provides a more delicate emotional level for emotional assessment, supplementing the shortcomings of pure text feedback.

[0051] Interactive videos use cameras and other devices to record users' facial expressions, eye contact, body language and other non-verbal signals as an important supplement to emotional feedback. Video information can capture subtle features such as micro-expressions and posture changes of users during the interaction. For example, facial expressions such as frowning or sighing may indicate negative emotions, while relaxed expressions and natural postures may indicate positive emotions. Non-verbal signals are often a true reflection of emotions and help to more fully understand the user's inner emotional state.

[0052] By combining multimodal information such as interactive text, interactive voice and interactive video, interactive feedback information can provide richer emotional data and provide multi-level support for the emotional analysis and evaluation process, thereby improving the accuracy and depth of the evaluation.

[0053] Based on the above embodiment, the emotion assessment based on the interactive feedback information of each round includes: For each round of interactive feedback information, performing text emotion scoring on the interactive text in the interactive feedback information to obtain a text emotion score, and performing emotion recognition on the interactive voice and / or interactive video in the interactive feedback information to obtain a voice emotion recognition result and / or a video emotion recognition result; Determining an emotion evaluation result of the interactive feedback information based on the emotion consistency between the speech emotion recognition result and the text emotion score, and / or the emotion consistency between the video emotion recognition result and the text emotion score; Emotional evaluation is performed based on the emotional evaluation results of the interactive feedback information in each round.

[0054] Specifically, text sentiment scoring mainly relies on natural language processing technology, which extracts emotional features by analyzing the text content provided by users during the interaction or the text content obtained by voice-to-text conversion. These features include the emotional tendency (such as positive, negative, neutral) and emotional intensity (such as the degree of anxiety, anger or joy) expressed by users in the text. The text sentiment score provides a preliminary quantitative result of the emotional state in each round of interaction.

[0055] Voice emotion recognition captures the voice emotion information of users when they answer questions by analyzing their voice features, such as pitch, speaking speed, volume, and voice pauses. Video emotion recognition extracts visual features related to emotions by analyzing non-verbal signals such as facial expressions, eye contact, posture, and body movements. The results of voice and video emotion recognition provide a basis for emotion analysis for users' voice and video feedback, further enriching the dimension of emotion assessment.

[0056] After obtaining the text emotion score, speech emotion recognition results, and / or video emotion recognition results, emotion consistency analysis is required to determine the emotion evaluation results of the interactive feedback information. Emotion consistency analysis refers to comparing the emotion evaluation results of different modalities to evaluate whether they are consistent. For example, when the text emotion score indicates that the user feels "calm", and the speech and video emotion recognition results respectively show that the user's voice intonation and facial expressions also support the "calm" emotional state, it can be considered that there is a high consistency between these modalities, thereby enhancing the reliability of the emotion evaluation results. If there is inconsistency between modalities, such as the text expression is "happy", but the video shows an obvious "sad" expression, further processing is required to give priority to the emotion evaluation results of certain modalities or make comprehensive adjustments to obtain the final emotion evaluation results of the interactive feedback information.

[0057] By integrating the emotion evaluation results of each round of interactive feedback information, the overall emotion evaluation result of the user is obtained. This process comprehensively considers the user's multiple rounds of feedback during the entire evaluation process, which can reflect the changing trend and overall state of emotions and provide a scientific basis for emotion management and intervention. The fusion of multimodal data not only improves the accuracy of emotion evaluation, but also enhances the objectivity and comprehensiveness of the evaluation.

[0058] Based on the above embodiment, determining the emotion evaluation result of the interactive feedback information based on the emotion consistency between the speech emotion recognition result and the text emotion score, and / or the emotion consistency between the video emotion recognition result and the text emotion score, includes: In the case where the emotional consistency is consistent, determining an emotional evaluation result of the interactive feedback information based on the text emotional score; In a case where the emotion consistency is inconsistent, an emotion evaluation result of the interactive feedback information is determined based on the speech emotion recognition result and / or the video emotion recognition result.

[0059] Figure 2 is a flow chart of the multimodal emotion assessment provided by the present invention, such as Figure 2As shown, when the consistency between the speech emotion recognition result and / or the video emotion recognition result and the text emotion score is high, it means that the emotional state expressed by the user in different modes is unified. At this time, the emotion evaluation result of the interactive feedback information can be determined directly based on the text emotion score. Since the text content is usually a direct way for users to actively express their emotions, the consistent voice and video signals further verify the accuracy of the emotional expression in the text. Therefore, the emotion evaluation result in this case has a high degree of credibility.

[0060] When there is a clear inconsistency between the speech emotion recognition results or video emotion recognition results and the text emotion score, it indicates that there are differences in the user's emotional expression in different modalities. For example, the user may say "I feel good" in the text, but the tone of voice shows a low mood, or the facial expression shows a sad expression in the video. In this case, the text emotion score may not fully reflect the user's true emotional state, so it is necessary to rely more on the speech emotion recognition results and / or video emotion recognition results to determine the emotion evaluation results. By giving priority to the use of more reliable speech emotion recognition results and / or video emotion recognition results, the user's true emotions can be captured more accurately, avoiding the emotion evaluation results being misled by a single modality information.

[0061] Through multimodal consistency analysis, the accuracy and comprehensiveness of emotion assessment are improved. User emotions are captured from three different angles: text, voice, and video, ensuring that the results are more comprehensive and reliable. The evaluation steps are simplified for cases with high consistency, and more reliable modalities are analyzed first when the consistency is low, optimizing efficiency. In addition, multimodal fusion reduces misjudgments, improves user trust and objectivity of evaluation, and enhances the application value of the results.

[0062] Based on the above embodiment, the emotion evaluation is performed based on the interactive feedback information of each round, and then further includes: The emotional state obtained from the emotional assessment and the interaction feedback information of each round are sent to the audit end, and the audit result for the emotional state returned by the audit end is received.

[0063] Specifically, on the review side, reviewers can conduct a comprehensive review of the emotion assessment results, including reviewing each round of interactive feedback information (such as user feedback text, feedback voice, and feedback video) and the generated emotion assessment results. By comparing multimodal information with emotion assessment results, reviewers can confirm the accuracy of emotion assessments or make necessary adjustments to emotion assessment results. For example, in some cases, users may show strong emotional contrast in a specific modality, and reviewers can make corrections based on deeper emotion analysis to ensure that the final evaluation results are more objective and accurate.

[0064] The audit results returned by the audit end will be used as the basis for confirming the final emotion assessment results and fed back to the emotion assessment system. This not only improves the credibility of the assessment process, but also provides more reliable data support for subsequent emotion intervention or psychological support. This audit mechanism combines the advantages of automated analysis and manual review, effectively balancing efficiency and accuracy, and ensuring that the emotion assessment results are both scientific and meet the needs of actual situations.

[0065] Based on the above embodiments, Figure 3 This is a second flow chart of the emotion assessment method provided by the present invention. Figure 3 As shown in the figure, the doctor first issues an evaluation order to the patient, which may include depression assessment or anxiety assessment. On the patient side, the evaluation is conducted in a human-computer interaction mode, and a question-and-answer dialogue is conducted with the virtual doctor to complete various emotional assessments by himself. During the process, a large model is used to generate intelligent speech, and the content of the questions is dynamically adjusted according to the patient's answer, making the dialogue more humane and close to the actual situation. At the same time, to ensure the safety of the evaluation process, the patient's real-time video screen is synchronized to the administrator port, and professionals monitor and deal with possible unexpected situations in a timely manner. During the evaluation process, the patient's feedback text, feedback voice and feedback video information are collected and analyzed through multimodal data fusion technology to more comprehensively evaluate the emotional state. If the scores of the three modalities are consistent, the text score will be used; if there is any inconsistency, the emotional score results of the feedback video or feedback voice will be used according to the preset priority rules to ensure the accuracy of the emotional assessment results. After the evaluation is completed, the results are transmitted to the doctor side, and the doctor can view the detailed records of the evaluation process and make necessary reviews and modifications to the emotional assessment results. The final confirmed emotion assessment results are transmitted back to the medical order platform, forming a complete medical order execution closed loop.

[0066] The emotion assessment device provided by the present invention is described below. The emotion assessment device described below and the emotion assessment method described above can be referenced to each other.

[0067] Figure 4 is a schematic diagram of the structure of the emotion assessment device provided by the present invention, such as Figure 4 As shown, the device comprises: A feedback receiving unit 410 is used to receive interactive feedback information from users regarding the emotion assessment question of the current round; An input text generating unit 420, configured to generate the interactive input text of the current round based on the interactive feedback information, the historical information of the interaction with the user, and the prompt information of the emotion evaluation; An evaluation question generating unit 430 is used to input the interactive input text of the current round into the large language model, obtain the emotion evaluation question of the next round output by the large language model, and return the interactive feedback information of the user for the emotion evaluation question of the next round until the interaction ends; The emotion evaluation unit 440 is used to perform emotion evaluation based on the interaction feedback information of each round.

[0068] The device provided by the embodiment of the present invention automatically generates emotion assessment questions through a large language model and conducts human-computer interaction, thereby reducing the reliance on professional assessors, alleviating resource shortages, and improving assessment efficiency. Through the interactive feedback mechanism, the flexibility of emotion assessment is improved, so that the content of questions can be adjusted in real time according to the user's emotions, meet personalized needs, and increase the accuracy of assessment. At the same time, personalized empathy feedback is provided to make users feel understood and supported, improving the assessment experience.

[0069] Based on any of the above embodiments, the feedback receiving unit is specifically used for: The interactive feedback information includes interactive text, interactive voice and / or interactive video.

[0070] Based on any of the above embodiments, the input text generating unit is specifically used for: The prompt information includes at least one of scene description information of the emotion assessment, identity limitation information of the emotion assessment, and output requirement information of the emotion assessment.

[0071] The output requirement information includes at least one of inquiry output, empathy output, guidance output, and suggestion output for the user.

[0072] Based on any of the above embodiments, the emotion assessment unit is specifically used for: For each round of interactive feedback information, performing text emotion scoring on the interactive text in the interactive feedback information to obtain a text emotion score, and performing emotion recognition on the interactive voice and / or interactive video in the interactive feedback information to obtain a voice emotion recognition result and / or a video emotion recognition result; Determining an emotion evaluation result of the interactive feedback information based on the emotion consistency between the speech emotion recognition result and the text emotion score, and / or the emotion consistency between the video emotion recognition result and the text emotion score; Emotional evaluation is performed based on the emotional evaluation results of the interactive feedback information in each round.

[0073] In the case where the emotional consistency is consistent, determining an emotional evaluation result of the interactive feedback information based on the text emotional score; In a case where the emotion consistency is inconsistent, an emotion evaluation result of the interactive feedback information is determined based on the speech emotion recognition result and / or the video emotion recognition result.

[0074] The emotional state obtained from the emotional assessment and the interaction feedback information of each round are sent to the audit end, and the audit result for the emotional state returned by the audit end is received.

[0075] Figure 5 An example of a physical structure diagram of an electronic device is shown in FIG. Figure 5 As shown, the electronic device may include: a processor 510, a communication interface 520, a memory 530 and a communication bus 540, wherein the processor 510, the communication interface 520 and the memory 530 communicate with each other through the communication bus 540. The processor 510 may call the logic instructions in the memory 530 to execute the emotion assessment method, which includes: Receive interactive feedback from users regarding the current round of emotion assessment questions; Generate the interactive input text of the current round based on the interactive feedback information, the historical information of the interaction with the user, and the prompt information of the emotion evaluation; Inputting the interactive input text of the current round into the large language model, obtaining the emotion assessment question of the next round output by the large language model, and returning and receiving the interactive feedback information of the user for the emotion assessment question of the next round until the interaction ends; Emotional assessment is performed based on the interactive feedback information of each round.

[0076] In addition, the logic instructions in the above-mentioned memory 530 can be implemented in the form of a software functional unit and can be stored in a computer-readable storage medium when it is sold or used as an independent product. Based on such an understanding, the technical solution of the present invention, or the part that contributes to the relevant technology or the part of the technical solution, can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions to enable a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk and other media that can store program codes.

[0077] On the other hand, the present invention further provides a computer program product, the computer program product comprising a computer program, the computer program can be stored on a non-transitory computer-readable storage medium, when the computer program is executed by a processor, the computer can execute the emotion assessment method provided by the above methods, the method comprising: Receive interactive feedback from users regarding the current round of emotion assessment questions; Generate the interactive input text of the current round based on the interactive feedback information, the historical information of the interaction with the user, and the prompt information of the emotion evaluation; Inputting the interactive input text of the current round into the large language model, obtaining the emotion assessment question of the next round output by the large language model, and returning and receiving the interactive feedback information of the user for the emotion assessment question of the next round until the interaction ends; Emotional assessment is performed based on the interactive feedback information of each round.

[0078] In another aspect, the present invention further provides a non-transitory computer-readable storage medium having a computer program stored thereon, which is implemented when the computer program is executed by a processor to perform the emotion assessment method provided by the above methods, the method comprising: Receive interactive feedback from users regarding the current round of emotion assessment questions; Generate the interactive input text of the current round based on the interactive feedback information, the historical information of the interaction with the user, and the prompt information of the emotion evaluation; Inputting the interactive input text of the current round into the large language model, obtaining the emotion assessment question of the next round output by the large language model, and returning and receiving the interactive feedback information of the user for the emotion assessment question of the next round until the interaction ends; Emotional assessment is performed based on the interactive feedback information of each round.

[0079] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without paying creative labor.

[0080] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the relevant technology can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiment.

[0081] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for evaluating emotion, characterized in that: include: Receive interactive feedback from users regarding the current round of emotion assessment questions; Generate the interactive input text of the current round based on the interactive feedback information, the historical information of the interaction with the user, and the prompt information of the emotion evaluation; Inputting the interactive input text of the current round into the large language model, obtaining the emotion assessment question of the next round output by the large language model, and returning and receiving the interactive feedback information of the user for the emotion assessment question of the next round until the interaction ends; Emotional assessment is performed based on the interactive feedback information of each round.

2. The emotion assessment method according to claim 1, characterized in that: The prompt information includes at least one of scene description information of the emotion assessment, identity limitation information of the emotion assessment, and output requirement information of the emotion assessment.

3. The emotion assessment method according to claim 2, characterized in that: The output requirement information includes at least one of inquiry output, empathy output, guidance output, and suggestion output for the user.

4. The emotion assessment method according to any one of claims 1 to 3, characterized in that: The interactive feedback information includes interactive text, interactive voice and / or interactive video.

5. The emotion assessment method according to claim 4, characterized in that: The emotion assessment is performed based on the interactive feedback information of each round, including: For each round of interactive feedback information, performing text emotion scoring on the interactive text in the interactive feedback information to obtain a text emotion score, and performing emotion recognition on the interactive voice and / or interactive video in the interactive feedback information to obtain a voice emotion recognition result and / or a video emotion recognition result; Determining an emotion evaluation result of the interactive feedback information based on the emotion consistency between the speech emotion recognition result and the text emotion score, and / or the emotion consistency between the video emotion recognition result and the text emotion score; Emotional evaluation is performed based on the emotional evaluation results of the interactive feedback information in each round.

6. The emotion assessment method according to claim 5, characterized in that: The determining the emotion evaluation result of the interactive feedback information based on the emotion consistency between the speech emotion recognition result and the text emotion score, and / or the emotion consistency between the video emotion recognition result and the text emotion score, comprises: In the case where the emotional consistency is consistent, determining an emotional evaluation result of the interactive feedback information based on the text emotional score; In a case where the emotion consistency is inconsistent, an emotion evaluation result of the interactive feedback information is determined based on the speech emotion recognition result and / or the video emotion recognition result.

7. The emotion assessment method according to any one of claims 1 to 3, characterized in that: The emotional evaluation is performed based on the interactive feedback information of each round, and then further includes: The emotional state obtained from the emotional assessment and the interaction feedback information of each round are sent to the audit end, and the audit result for the emotional state returned by the audit end is received.

8. An emotion assessment device, characterized in that: include: A feedback receiving unit, used to receive interactive feedback information from users regarding the current round of emotion assessment questions; An input text generating unit, configured to generate the interactive input text of the current round based on the interactive feedback information, the historical information of the interaction with the user, and the prompt information of the emotion evaluation; An evaluation question generating unit, configured to input the interactive input text of the current round into a large language model, obtain the emotion evaluation question of the next round output by the large language model, and return and receive the interactive feedback information of the user for the emotion evaluation question of the next round until the interaction ends; The emotion evaluation unit is used to perform emotion evaluation based on the interactive feedback information of each round.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the program, the emotion assessment method according to any one of claims 1 to 7 is implemented.

10. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the emotion assessment method according to any one of claims 1 to 7 is implemented.

11. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the emotion assessment method according to any one of claims 1 to 7 is implemented.

Images