Call assisting method
By obtaining and identifying voice information in VOIP calls in real time, it provides recognition and auxiliary prompts for target topics, emotions and intentions, solving the problem of single VOIP call technology functions and achieving more efficient call assistance and communication support.
Patent Information
- Application Number
- CN202411183590.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-27
- Publication Date
- 2024-12-06
AI Technical Summary
The existing VOIP calling technology only provides the functions of making and receiving calls, and cannot meet the auxiliary needs of sales staff, etc., and cannot provide functions such as auxiliary information and emotional recognition related to calls in real time.
By obtaining voice information for calls in real time, text recognition, estimating target topics and displaying relevant information, identifying the caller's emotions and intentions, providing functions such as intelligent voice response and voice translation, combining the database's flexible response and voice removal technology to improve calls assistivity and efficiency.
Provide real-time auxiliary information and emotional perception to sales staff, help them better communicate and promote, improve sales efficiency and customer satisfaction, and improve the quality and audibility of calls.
Smart Images

Figure CN119094657A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the technical field of communication, and in particular to a call assistance method. Background Art
[0002] With the development of Internet technology, VOIP calling technology has emerged. VoIP (Voice over Internet Protocol) is a telephone technology that allows calls to be made and received using the Internet instead of traditional telephone lines.
[0003] However, the current VOIP call technology can only provide the functions of making and receiving calls, which is relatively simple and cannot meet the needs of users. For example, sales personnel generally use VOIP call technology to communicate with customers. The current VOIP call technology can only provide the call function and cannot provide other auxiliary functions for sales personnel, which does not meet the actual work needs of sales personnel. Summary of the invention
[0004] Based on this, it is necessary to provide a method for assisting calls in response to the above technical problems.
[0005] The call assistance method provided by this application includes: Get call voice information in real time; Performing real-time text recognition on the call voice information to obtain the call content; Based on the content of the call, estimating a target topic; wherein the target topic is a topic involved in the call; Displaying relevant information of the target topic; The relevant information of the target topic is pre-stored in a database; the database is used to obtain a data modification request and respond to it, so as to modify the relevant information corresponding to different target topics; Based on the call content, identifying the caller's emotions in real time; When the caller's emotion is the preset emotion, a prompt is issued; Based on the content of the call, determining whether the caller has a target intention; If there is a target intention, a prompt is issued; wherein, when the call content is sales, the target intention includes the caller's intention to receive sales and the topics involved: When a video call is made, a preset recognition model is input into the content of the video call to identify the target topic, whether there is a target intention, user intention and emotion.
[0006] In some embodiments, when a preset condition is met, the intelligent voice mode is entered and the intelligent voice answers: Among them, the preset conditions are determined in advance based on the target topic and the emotions of the caller.
[0007] In some embodiments, during a call, if a new agent joins the call, the previous call content is summarized and displayed.
[0008] In some embodiments, it also includes: identifying and removing background noise and far-field technology in calls through sound removal and enhancement technology, thereby improving voice clarity and audibility, and eliminating echoes and adjusting volume.
[0009] The method according to claim 1, further comprising: When callers use different voices, real-time voice translation is performed.
[0010] In some embodiments, the method further includes: determining whether preset illegal content is obtained in the call content; When the call content contains preset illegal content, a prompt is issued.
[0011] In some embodiments, the real-time acquisition of call voice information includes: The call voice is collected and stored based on a preset recording module to obtain the call voice information; The preset recording module is specifically used to obtain uplink audio data and at least one downlink audio data after the call connection is established; wherein the uplink audio data and the downlink audio data are the call voice information; Storing the call voice information; The uplink audio data and the downlink audio data in the call voice information are matched to obtain a section of uplink audio data and a section of downlink audio data corresponding to each other as the target call voice information.
[0012] In some embodiments, it also includes: By analyzing the caller's voice features, identity recognition and authentication are performed.
[0013] In some embodiments, it also includes: The call data is encrypted and anonymized.
[0014] In some embodiments, it also includes: When a preset voice command is detected during a call: Based on the preset voice command, the device is controlled to execute a preset task.
[0015] In the method provided by the present application, call voice information is obtained in real time; a call assistance method performs real-time text recognition on the call voice information to obtain the call content; a call assistance method estimates a target topic based on the call content; wherein the target topic is a topic involved in the call; a call assistance method displays relevant information of the target topic; a call assistance method wherein the relevant information of the target topic is pre-stored in a database; the database is used to obtain data modification requests and respond to facilitate the modification of relevant information corresponding to different target topics; a call assistance method identifies the caller's emotions in real time based on the call content; a call assistance method issues a prompt when the caller's emotions are preset emotions; a call assistance method determines whether the caller has a target intention based on the call content; a call assistance method issues a prompt if there is a target intention; wherein, when the call content is sales, the target intention includes the caller's intention to receive sales promotion and the topics involved: a call assistance method inputs a preset recognition model into the content of the video call when making a video call to identify the target topic, whether there is a target intention, user intention and emotion. Call assistance method The call assistance method has the following beneficial effects: Providing auxiliary information: By acquiring call voice information in real time and performing text recognition, estimating the target topic and displaying relevant information, it can provide the caller with auxiliary information related to the call topic, which helps to communicate better. Call assistance method Emotional perception and prompting: Real-time recognition of the caller's emotions, and issuing prompts when preset emotions appear, which helps the caller to understand the other party's emotional state in a timely manner, adjust the communication method, and avoid conflicts or misunderstandings. Call assistance method Target intention recognition and prompting: Determine whether the caller has a target intention and issue a prompt, which can help sales staff grasp the customer's intention in a timely manner in the sales scenario and adopt appropriate marketing strategies. Call assistance method Video call content recognition: During a video call, it can recognize the target topic, target intention, user intention and emotion, and provide more comprehensive call analysis and assistance. Call assistance method The database can be modified: The database can obtain and respond to data modification requests, and can flexibly modify the relevant information corresponding to different target topics, so that the auxiliary information is more in line with actual needs. Call assistance method. BRIEF DESCRIPTION OF THE DRAWINGS
[0016] Figure 1 A flowchart of a call assistance method in one embodiment; DETAILED DESCRIPTION In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.
[0017] The call assistance method provided in this application can be applied to a VoIP (Voice over Internet Protocol, voice transmission based on IP) voice call application environment. Figure 1 The method provided in this application includes: Step S110, acquiring call voice information in real time; It should be noted that call voice information refers to the voice information generated by both parties during the call. The real-time acquisition method can be: based on a preset recording module, the audio in the call is collected and stored to obtain the call voice information; the specific preset recording module can be pre-embedded in the call device or system. During the user's call, the preset recording module can detect that the user has started the call, and collect all the uplink audio data and downlink audio data during the user's call to obtain the call voice information.
[0018] Step S120, performing real-time text recognition on the call voice information to obtain the call content; Specifically, voice and text recognition technology can be used to perform real-time text recognition on call voice information. Voice and text recognition technology is an existing technology for text recognition of audio. The solution provided by this application uses voice and text recognition technology to convert call voice information into text information. Among them, text information and call content are obtained through voice and text recognition technology.
[0019] Step S130, estimating a target topic based on the call content; wherein the target topic is a topic involved in the call; It should be noted that the target topic is set differently based on different actual application scenarios. For example, when the actual application scenario is a sales scenario, the salesperson may be more concerned about the product-related issues that the customer is consulting. Based on this, the target topic can be topics related to various characteristics of various products. For example: based on the content of the call, it is estimated that both parties may discuss the reliability of product a, so the reliability of product a is the target topic.
[0020] For example, when the actual application scenario is a medical consultation scenario, the call may involve certain aspects of certain diseases. Specifically, based on the content of the call, it may be estimated that the call involves the later recovery of disease C. At this time, the target topic is the later recovery of disease C.
[0021] Furthermore, the specific method of estimating the target topic is to perform semantic recognition on the call content and determine the topic involved in the call content as the target topic. Of course, the call content can also be input into a preset deep learning model to obtain the target topic output by the deep learning model.
[0022] Specifically, in the solution provided by the present application, the preset deep learning model can also be differentiated according to the actual needs of the user, so that the output target topic is the target topic that the user actually cares about. For example, for sales personnel, the deep learning model can be trained so that the output topic is sales-related topics, rather than life-related topics or medical-related topics.
[0023] Step S140: display relevant information of the target topic.
[0024] Specifically, the relevant information of the target topic can be displayed through a preset display device. How to set the caller to view the relevant information of the target topic in time to answer questions related to the target topic generated during the call.
[0025] The relevant information of the target topic is pre-stored in a database; the database is used to obtain a data modification request and respond to it so as to modify the relevant information corresponding to different target topics, or automatically modify the database based on the call content; Specifically, the relevant information of the target topic will be stored in the database in advance. This relevant information may include various data, knowledge, background information, etc. related to the target topic, aiming to provide callers with detailed and accurate information about the target topic to assist them in better communication. The database has the ability to obtain data modification requests. This means that users or related systems can send requests to the database to modify the relevant information corresponding to a specific target topic. When the database receives the data modification request, it will respond and perform the corresponding modification operation. This ensures that different target topics can update their relevant information in a timely manner to adapt to changes in actual conditions.
[0026] For example, in a sales scenario, relevant information about a product's target topic may include product features, prices, preferential policies, etc. When product information changes, such as price adjustments or new preferential activities, users can send data modification requests to the database, and the database will respond and modify the relevant information corresponding to the product's target topic, so that the call assistance system can provide the latest and accurate information in subsequent calls, helping sales staff to communicate better with customers; of course, the content of the database can also be automatically adjusted according to the call content, so that the data in the database is more in line with the actual situation.
[0027] In general, this solution ensures the timeliness and accuracy of information related to the target topic through the flexible response and modification functions of the database, thereby improving the quality and effectiveness of call assistance.
[0028] It should be noted that the solution provided by the present application can be applied to the call equipment of any party in the call. In practical applications, the party that generally needs voice call assistance often sets up its own call equipment in advance so that its own call equipment can use the call assistance method provided by the present application. Taking the call between the salesperson and the salesperson (customer) as an example, the salesperson can set up his own call equipment so that his own call equipment can run the call assistance method provided by the present application. In this way, during the sales promotion process, the corresponding content of the questions that the salesperson will ask during the call can be displayed through the preset display device, and the salesperson's inquiries can be answered in a timely and accurate manner. In summary, the solution provided by the present application can provide auxiliary functions for the caller during the call, displaying relevant information about the topic involved in the call, so that the caller can communicate better.
[0029] Step S150, identifying the emotion of the caller based on the call content; Step S160, when the caller's emotion is a preset emotion, a prompt is issued; Specifically, in the solution provided by the present application, based on the content of the call, the specific method of identifying the caller's emotions is to use the existing emotion recognition technology based on the call text to identify and obtain the caller's emotions. If the caller's emotion is a preset emotion, a prompt is issued. Specifically, the preset emotion can be set based on demand. For example, the preset emotion can be anger or irritation, and a prompt is issued when it is identified that the caller's emotion is anger or irritation. In actual applications, when the prompted emotion is one's own emotion, the caller can adjust his or her emotions based on the content of the prompt after receiving the prompt; when the prompted emotion is the other party's emotion, the caller can adjust his or her communication strategy based on the other party's emotion in the prompt.
[0030] Step S170, based on the call content, determining whether the caller has a target intention; The system will analyze and understand the content of the call to determine whether the caller has a specific target intention in the current call. The target intention here refers to a certain intention or tendency shown by the caller in the communication.
[0031] Step S180: If there is a target intention, a prompt is issued; wherein, when the call content is sales, the target intention includes the caller's intention to receive sales promotion and the topics involved: If the system determines that the caller has the target intention, it will issue a prompt. This prompt can be through sound, pop-up window, etc., so that the relevant people in the call can know it in time.
[0032] In particular, when the content of the call is related to sales, the target intention includes the caller's intention to receive sales promotion and the specific topics involved. For example, if the caller expresses interest in a product, asks about the price or function of the product, etc., these can be regarded as the intention to receive sales promotion and the topics involved.
[0033] Through such a solution, in sales and other scenarios, the caller's intention can be captured in time, which helps sales staff better grasp opportunities, provide targeted services and promotions, and improve sales efficiency and success rate. At the same time, for other types of calls, it can also help relevant personnel better understand the caller's intentions and communicate more effectively.
[0034] Step S190, when making a video call, input a preset recognition model into the content of the video call to identify the target topic, whether there is a target intention, user intention and emotion.
[0035] When making a video call, the system will input the content of the video call into a preset recognition model. This recognition model has multiple functions: Identify target topics: Able to analyze language, images and other information in video calls to determine the main topics involved in the call.
[0036] Determine whether there is a target intention: It can determine whether the caller shows a specific intention in the communication. For example, in a sales scenario, it can determine whether the caller has the intention to buy the product.
[0037] Decoding user intent: Gain a deep understanding of the caller’s intent, including their needs, expectations, and goals.
[0038] Identify emotions: Identify the caller's emotional state, such as happiness, anger, anxiety, etc., by analyzing their language, facial expressions, tone, etc.
[0039] In this way, during a video call, you can have a more comprehensive understanding of all aspects of the call, provide more accurate assistance and support to the callers, and improve the quality and effectiveness of the video call.
[0040] The preset recognition model is a pre-set model that is used to analyze and understand the content of the video call.
[0041] This model is usually built based on technologies such as machine learning and artificial intelligence, and can process and analyze various information such as voice, images, text, etc. in video calls.
[0042] In terms of identifying the target topic, the model will extract key information through semantic understanding of the voice content to determine the topic involved in the call.
[0043] When determining whether a target intention exists and interpreting user intent, the model takes into account multiple factors such as the caller's language expression, behavior, and facial expressions to infer their intent and intention.
[0044] In terms of emotion recognition, the model will judge the emotional state of the caller by analyzing their voice tone, facial expression changes, etc.
[0045] In short, the preset recognition model is designed to achieve a comprehensive understanding and analysis of the content of video calls and provide strong support for call assistance.
[0046] In some embodiments, the solution provided by the present application also includes: determining whether preset illegal content is obtained in the call content; and issuing a prompt when preset illegal content exists in the call content.
[0047] It should be noted that the illegal content can be set by the user based on his or her actual situation. Specifically, the illegal content can be some commercial secrets, or some sensitive content or insulting content, etc. A specific way to determine whether the preset illegal content is obtained in the call content can be to perform semantic recognition on the call content to determine whether there is illegal content in the call content.
[0048] It should be noted that determining whether there is preset illegal content in the text is an existing technology, which is mostly used in the review of novels or other texts.
[0049] In some embodiments, the solution provided by the present application further includes: storing the call voice information. In this way, the call voice information is stored to facilitate other processing of the call voice information.
[0050] Specifically, in actual applications, during a call, there is generally only one uplink audio data, and there are generally multiple downlink audio data (multiple downlink audio data include: downlink audio data transmitted from the call peer, downlink audio data played by a local player, and downlink audio data brought by a prompt tone). Based on this, in order to better determine the actual call content of the two parties in the call voice information, the solution provided by this application also includes: The uplink audio data and the downlink audio data in the call voice information are matched to obtain a section of uplink audio data and a section of downlink audio data corresponding to each other as the target call voice information.
[0051] Specifically, the matching method may be: determining the start time and duration of the uplink audio data and the downlink audio data. Selecting downlink audio data whose difference between the start time of the uplink audio data and the downlink audio data is less than a first preset time, and whose difference between the duration is less than a second predicted time. Specifically, the first preset time may be 3 to 5 seconds, and the second preset time may also be 3 to 5 seconds.
[0052] For example, the call voice information includes one uplink audio data and two downlink audio data; The uplink audio data (target uplink audio data) starts at 5 o'clock and lasts for 20 minutes. There is a downlink audio data (1st downlink audio data) that starts at 5 o'clock and lasts for 20 minutes and 1 second. There is another downlink audio data (2nd downlink audio data) that starts at 5 o'clock and lasts for 21 minutes. Based on the above rules, it can be determined that the 1st downlink audio data corresponds to the target uplink audio data. That is, the target call voice information includes: the target uplink audio data and the 1st downlink audio data.
[0053] In some embodiments, the solution provided by the present application also includes: performing text recognition and semantic understanding on the target call voice information to extract preset target information.
[0054] During the communication process, the two parties on the call often say some important information (preset target information). For example: the caller's name, address, and contact information. If the call content is related to express delivery, the important information may also be the express delivery order number and other information. In the solution provided by the present application, the target call voice information is subjected to text recognition and semantic understanding to extract the preset target information. This avoids the caller from actively recording such important information, which brings great convenience to the caller.
[0055] In some embodiments, the solution provided by the present application also includes: when preset conditions are met, entering the intelligent voice mode and answering by intelligent voice: wherein the preset conditions are determined in advance based on the target topic and the emotion of the caller.
[0056] In some embodiments, the solution of the present application adds a function, that is, when certain preset conditions are met, it will automatically enter the intelligent voice mode and answer by intelligent voice.
[0057] The preset conditions are determined in advance based on the target topic and the caller's emotions. This means that the system will determine whether the conditions for entering the intelligent voice mode are met based on the specific content of the call and the caller's emotional state.
[0058] For example, if the target topic is a consultation on common problems and the caller is emotionally stable, the system may determine that the preset conditions are met and enter the intelligent voice mode, with the intelligent voice providing relevant answers and explanations.
[0059] Such settings can improve the efficiency and convenience of calls to a certain extent. For example, when dealing with some routine questions, intelligent voice can quickly give answers and reduce the need for manual intervention. At the same time, determining the preset conditions based on the caller's emotions can also better adapt to different call situations and provide more appropriate services.
[0060] Furthermore, the above scheme can also realize virtual voice companionship. When the preset conditions are met, the child can be accompanied by the synthesized voice and the prescribed content can be read.
[0061] Specifically, the virtual voice companion uses speech synthesis technology to generate human-like speech. Then, it converts these texts into speech and reads them aloud according to pre-set rules or content. For example, it can tell stories, read poems, explain knowledge, etc.
[0062] For children, virtual voice companionship can provide a way to interact and entertain when they need company. It can help children relax, learn new knowledge, and give them a sense of security and companionship.
[0063] In addition, the virtual voice companion can be personalized according to the child's needs and interests, such as selecting the child's favorite story type, voice style, etc., to provide a more intimate companionship experience.
[0064] In order to use speech synthesis technology to provide a better companionship experience for children, the following aspects can be considered: Personalized settings: Customize the content and style of speech synthesis according to the child's age, interests and preferences. For example, choose stories, nursery rhymes, popular science knowledge, etc. that are suitable for the child's age group, and read them in the voice style that the child likes.
[0065] Emotional expression: Enable speech synthesis technology to express rich emotions and make the voice more appealing. By adjusting parameters such as intonation, speech speed, and volume, the voice can convey the emotions in the story more vividly, making it easier for children to resonate.
[0066] Interactivity: Increase the interactivity of voice accompaniment, such as setting up question-and-answer sessions, encouraging children to participate in storytelling or discussions, and letting children feel the joy of participation and communication.
[0067] Multi-language support: Provides speech synthesis in multiple languages to help children learn different languages and broaden their horizons.
[0068] Quality and clarity: Ensure that the quality of speech synthesis is high and the sound is clear and natural, avoiding harshness or incoherence, to enhance the child's listening experience.
[0069] Continuously updated content: The content of the voice accompaniment is constantly updated to provide fresh and interesting stories and knowledge to maintain children's interest.
[0070] Adaptive adjustment: According to the child’s feedback and needs, the settings and content of voice accompaniment will be adjusted in a timely manner to better meet the child’s personalized needs.
[0071] Combine with other functions: Speech synthesis can be combined with other functions, such as music, animation, etc., to create a richer multimedia experience and enhance the fun of companionship.
[0072] Security and privacy protection: In the process of providing voice accompaniment, ensure that the safety and privacy of children are protected and avoid exposure to bad information.
[0073] Through the above methods, speech synthesis technology can be used to provide children with a better, more personalized and interesting companionship experience to help them grow and learn.
[0074] In some embodiments, it also includes: During a call, if a new agent joins the call, the previous call content will be summarized and displayed.
[0075] With this setting, when a new agent joins the call, the system will automatically summarize the previous call content. This includes text recognition of the voice information of both parties in the previous call, sorting out the main points and key information of the call. Then, the system will display the summarized call content for the newly joined agent to view. In this way, the new agent can quickly understand the situation of the previous call, keep up with the progress of the call, and better participate in the subsequent communication. The benefit of this function is that it can improve the efficiency and continuity of the call, avoid the omission of information or repeated communication due to the replacement of personnel, and make the entire call process smoother and more efficient.
[0076] For example, in a video conference, the topic of discussion is the progress of the project. Initially, the participants in the meeting are project manager A, team members B and C. During the meeting, team member D joins the meeting. At this time, the system automatically summarizes the previous call content of A, B, and C, which may include key information such as the current stage of the project, problems encountered, and solutions taken. These summarized contents will be displayed on D's screen, allowing D to quickly understand the discussion before the meeting and better participate in the subsequent discussion.
[0077] For example, salesperson E is communicating with customer F over the phone, introducing the features and advantages of the product. During the communication, customer F raised some questions and concerns. At this time, sales manager G newly joins the call. The system will immediately summarize the content of the previous call between E and F, including the key points of the product introduction, customer questions, etc., and display these contents to G. Based on this information, G can give salesperson E more accurate guidance and suggestions to help E better meet customer needs and promote sales.
[0078] From the above examples, we can see that during a call, if a new agent joins the call, the function of summarizing and displaying the previous call content can enable the new agent to quickly understand the previous communication situation and improve communication efficiency and effectiveness. It has important application value in various scenarios such as video conferencing and sales scenarios.
[0079] In some embodiments, the solution provided by the present application also includes: identifying and removing background noise and far-field technology in calls through sound removal and enhancement technology, thereby improving the clarity and audibility of the voice, and eliminating echoes and adjusting the volume.
[0080] The specific implementation process of this technology is as follows: Sound removal: Sound removal technology is used to analyze the audio signal in the call and identify the background noise. Then, the background noise is removed from the audio signal through relevant algorithms to reduce the interference of noise on the call voice.
[0081] Voice enhancement: Use enhancement technology to process the voice signal in the call to make it clearer and more audible. This may include adjusting the frequency, amplitude and other parameters of the audio signal to enhance the clarity and loudness of the voice.
[0082] Far-field technology processing: To address the issues of sound attenuation and distortion in far-field calls, far-field technology can compensate and optimize audio signals so that voice transmitted over long distances can maintain good quality.
[0083] Echo cancellation: The echo cancellation algorithm detects and eliminates possible echoes during calls. This helps improve call quality and avoids interference from echoes.
[0084] Adjust volume: Automatically adjust the volume according to the call environment and the voice volume of both parties so that both parties can hear clear and appropriate voices.
[0085] Through the comprehensive application of these sound removal and enhancement technologies, the clarity and audibility of speech can be effectively improved, the call quality can be improved, and both parties of the call can communicate more smoothly.
[0086] In some embodiments, the solution provided by the present application also includes: performing real-time voice translation when callers use different voices.
[0087] When the callers use different voices, the specific process of real-time voice translation is as follows: Voice recognition: The system first recognizes the caller's voice and converts it into text form. Language detection: Detect the language type of the recognized text. Translation conversion: Translate the text into the target language based on the detected language type. Voice synthesis: Convert the translated text into the voice of the target language through voice synthesis technology. Real-time transmission: Transmit the synthesized voice to the other party of the call in real time to realize the function of real-time voice translation. Through such real-time voice translation, language barriers can be eliminated, allowing callers to communicate smoothly, especially in cross-language call scenarios, which has important application value.
[0088] In some embodiments, the solution provided by the present application also includes: performing identity recognition and authentication by analyzing the voice features of the caller. The process of performing identity recognition and authentication by analyzing the voice features of the caller is as follows: Voice collection: During the call, the voice signal of the caller is collected. Feature extraction: Analyze the collected voice signal and extract parameters that can reflect the voice features of the caller, such as pitch, frequency, voiceprint, etc. Establish a model: Use machine learning or deep learning algorithms to establish a corresponding model between voice features and identity information. The model can be learned and optimized through a large amount of training data to improve the accuracy of identity recognition and verification. Identity recognition: When a new caller makes a call, the system will extract its voice features and compare them with the established model to identify the caller's identity. Identity verification: In a scenario where identity verification is required, the system will match the caller's voice features with the pre-stored identity information to verify whether the caller's identity is legitimate. In this way, the uniqueness of voice features can be used to achieve identity recognition and verification, thereby improving the security and reliability of calls.
[0089] In some embodiments, the solution provided by the present application further includes: encrypting and anonymizing the call data. The steps of encrypting and anonymizing the call data are as follows: Data encryption: Select an appropriate encryption algorithm, such as a symmetric encryption algorithm (such as AES) or an asymmetric encryption algorithm (such as RSA). During call data transmission and storage, the selected encryption algorithm is used to encrypt the data. Only authorized parties with the correct key can decrypt and access the encrypted call data, ensuring the confidentiality of the data.
[0090] Anonymization: Remove or replace information in call data that can directly or indirectly identify an individual, such as name, phone number, ID number, etc. Anonymous identifiers or codes can be used to replace original personal identification information so that the data cannot be associated with a specific individual during processing and analysis. Other sensitive information in call data is obscured or desensitized to protect personal privacy.
[0091] By encrypting and anonymizing call data, users' privacy and data security can be effectively protected, and call data can be prevented from being accessed, leaked or abused without authorization. This can enhance users' trust in call services and comply with relevant data protection regulations and requirements.
[0092] In some embodiments, the solution provided by the present application further includes: when a preset voice command is recognized in the call voice: based on the preset voice command, controlling the device to perform a preset task. The preset voice command can be: turn on the camera, mute, turn off mute, etc. Of course, it can also be other commands such as turning on the light.
[0093] In these embodiments, the specific content of the scheme is: when the system recognizes that there are preset voice instructions such as "turn on camera", "mute", "turn off mute" in the call voice, the device will be controlled to perform the corresponding preset tasks according to these instructions. For example, when the "turn on camera" instruction is recognized, the system will control the device to turn on the camera; when the "mute" instruction is recognized, the device will enter the mute state; when the "turn off mute" instruction is recognized, the mute state of the device will be canceled. In addition, the preset voice instructions can also include other instructions such as "turn on the lights", and the system will control the corresponding device to perform the corresponding task according to the specific instructions.
[0094] This design can provide users with a more convenient operation method during a call. There is no need to manually operate the device. The device can be controlled through voice commands, which improves the user experience.
[0095] In a video call scenario, such as a remote medical consultation, when a patient is having a video call with a doctor, the patient may not be able to manually operate the device due to hand injuries or other reasons. At this time, the patient can say a preset voice command, such as "turn on the camera". After the system recognizes the command, it will automatically turn on the camera so that the doctor can clearly see the patient's condition.
[0096] During a call, the patient may need to have a brief exchange with the person around him or her and may not want the sound from the video call to disturb them. At this time, the patient can say "mute" and the system will mute the device and suspend the transmission of sound.
[0097] When the patient finishes communicating with the people around him and wants to continue talking to the doctor, he can say "turn off mute" and the system will cancel the mute state and resume sound transmission.
[0098] In addition, if the video call is conducted in a darker environment, the patient can say "turn on the lights". After the system recognizes the command, it will control the relevant equipment to turn on the lights to improve the lighting conditions for the video call.
[0099] By setting these preset voice commands, users can control the device more conveniently in video call scenarios, improving the efficiency and convenience of video calls.
[0100] In some embodiments, the solution provided by the present application also includes: automatic recording and classification, automatic storage and access of call recordings and recognized texts, and automatic identification of customer intentions and concerns through machine learning algorithms, classification and priority sorting: Automatic storage and access of call recordings and recognized texts: The system can automatically record call recordings and automatically store them with recognized texts. This ensures the complete preservation of call information and facilitates subsequent queries and analysis. Automatic identification of customer intentions and concerns: Through machine learning algorithms, the system can conduct in-depth analysis of call content and automatically identify customer intentions and concerns. For example, in a sales scenario, the system can determine whether a customer is interested in a product, what specific needs they have, etc. Classification and priority sorting: Based on the identified customer intentions and concerns, the system will classify calls, such as dividing customers into potential customers, intended customers, important customers, etc. At the same time, these calls will be prioritized so that relevant personnel can give priority to important customers and matters. For example, in a customer service center, the system automatically records calls between customers and customer service. Through machine learning algorithm analysis, it was found that a certain customer showed a strong interest in a certain product and raised some specific questions. The system will mark this call as a potential customer and give it a higher priority. Customer service staff can follow up with the customer in a timely manner based on this information, provide more targeted services, and improve sales conversion rates. In general, the automatic recording and classification functions can help companies better manage customer information, improve work efficiency and customer satisfaction.
[0101] Furthermore, the dialogue management engine can be opened to allow users to independently configure and modify the robot's speech and knowledge base. This makes the voice call assistance system more flexible and personalized, and can better meet the needs of different users.
[0102] The following is the specific implementation of the open dialogue management engine, which allows users to independently configure and modify the robot's speech and knowledge base: User interface design: Create an intuitive and easy-to-use user interface that users can access through a web page or a dedicated application. The interface should clearly display the structure and content of the robot's scripts and knowledge base.
[0103] Script editing area: provides a text editing box where users can directly enter and modify the robot's reply script.
[0104] Knowledge base classification and management: Display the knowledge base in the form of a tree structure or tag classification to facilitate users to add, delete or modify knowledge points.
[0105] Permission management: Establish different levels of user permissions. For example, administrators can make comprehensive configurations and modifications, while ordinary users may only be able to make adjustments within a certain range.
[0106] Registration and Login: Users need to register and log in to obtain corresponding permissions.
[0107] Speech templates and examples: Provide users with some common speech templates and examples to help them quickly get started with configuration.
[0108] Preset scenarios: sample scripts for common customer service Q&A, sales communication, and other scenarios.
[0109] Knowledge base building tools: Knowledge adding wizard: guides users to add new knowledge according to certain formats and rules.
[0110] Association and reference: Supports users to set associations and references between knowledge points to form a knowledge network.
[0111] Real-time preview and testing: After modifying the script and knowledge base, users can preview the robot's response in real time. It also provides a testing function to simulate dialogue scenarios and verify the accuracy and rationality of the configuration.
[0112] Version control: Record each modification operation of the user to form different versions, so that users can roll back to the previous configuration or compare the differences between different versions.
[0113] Data validation and rule constraints: When users input and modify data, data validation is performed to ensure that the input content conforms to certain formats and logical rules. At the same time, some constraints are set to prevent users from making mistakes or entering unreasonable content.
[0114] Import and export function: supports users to export their configured words and knowledge base in a specific format, and can also easily import existing relevant data.
[0115] Training and help documents: Provide detailed training materials and online help documents to guide users on how to configure and modify effectively.
[0116] Through the above methods, an open dialogue management engine can be implemented, allowing users to independently configure and modify the robot's dialogue and knowledge base, making the voice call assistance system more flexible and personalized.
[0117] In some embodiments, the solution provided by the present application also includes: using 5G network technology and IMS data channels to provide enhanced voice / video services. 5G new calls not only support high-definition and immersive audio and video communications, but also have innovative functions such as real-time translation, personalized avatars, and shared content, which greatly enhances the fun and practicality of calls.
[0118] The high speed, low latency and large capacity of 5G networks provide a solid foundation for high-definition and immersive audio and video communications. This means that both parties on the call can enjoy an extremely clear, smooth and realistic audio and video experience, as if the other party is right in front of them.
[0119] The real-time translation function of 5G New Call breaks the language barrier. Whether it is an international business meeting or communication with foreign friends, instant language conversion can be achieved, making communication smoother and unimpeded. The personalized avatar function adds more fun and personalized elements to the call. Users can choose a virtual image to represent themselves according to their preferences or specific scenarios, bringing more creativity and fun to the call. The innovation of shared content makes calls no longer limited to verbal communication. Both parties can share various information such as pictures, documents, videos, etc. in real time during the call, greatly improving the efficiency and richness of communication. For example, in cross-border business negotiations, the high-definition video communication of 5G New Call allows both parties to clearly see each other's expressions and body language, real-time translation ensures correct language understanding, personalized avatars add some relaxed atmosphere, and the shared content function makes it convenient for both parties to display and discuss relevant materials at any time. In the distance education scenario, teachers and students seem to be in the same classroom through high-definition immersive communication, real-time translation helps students from different language backgrounds understand the course, personalized avatars can make teaching more interesting, and shared content facilitates the display of courseware and homework. In summary, these innovative features have greatly enhanced the fun and practicality of calls, making 5G new calls show great application potential and value in many fields.
[0120] In some embodiments, the solution provided by the present application also includes: cross-modal voice interaction. Voice and visual multimodal fusion technology. Combining voice control and gesture recognition, users can issue commands through voice and use gestures to complete more complex interactions. Combining multiple input modes such as voice, gestures, expressions, and eye contact can more comprehensively understand user intentions.
[0121] Cross-modal voice interaction is a cutting-edge interactive method that fully integrates information from multiple modes such as voice and vision to achieve more natural, efficient and accurate human-computer interaction. When multiple input modes such as voice, gestures, expressions, and eye contact are integrated, the system's understanding of user intent will be more comprehensive and accurate. In the scenario of sales voice calls, cross-modal voice interaction and multimodal fusion technology can bring significant advantages and convenience.
[0122] In the scenario of sales voice calls, cross-modal voice interaction and multimodal fusion technology can bring significant advantages and convenience. For example, when a salesperson is talking to a customer, the customer can make a clear demand through voice, such as "I want to know the price of the product", and use gestures to gesture specific numbers or ranges on the screen to further explain the expected price range. This enables salespeople to understand the specific needs of customers more quickly and accurately. For example, when a customer has questions about the characteristics of a product, in addition to expressing "I don't quite understand this function" through voice, he can also frown and show a confused expression. With this multimodal information input, salespeople can more keenly perceive the customer's confusion points and provide more targeted and detailed explanations and instructions. Or, when a customer expresses interest in a product, he not only says "This product looks good", but also looks at the introduction materials of the relevant product with a focused look, and nods and other affirmative gestures. Salespeople can more accurately grasp the customer's purchasing intention from these multimodal signals, and then adjust sales strategies in time to provide services and suggestions that better meet customer needs. By integrating multiple input modalities such as voice, gestures, expressions, and eye contact, sales personnel can understand customers' intentions and emotional states more comprehensively and deeply, thereby better meeting customer needs and improving sales success rates and customer satisfaction.
[0123] In short, cross-modal voice interaction integrates multiple input modalities, enabling computers to be closer to the natural way humans communicate, thereby providing better quality, smarter and more personalized services, bringing new interactive experiences and possibilities to many fields.
[0124] In some embodiments, the solution provided by the present application also includes: edge computing and low power consumption. Some computing tasks are transferred to terminal devices to reduce the burden on the cloud and improve response speed. Algorithms and hardware are optimized to meet the low power consumption requirements of mobile devices and IoT devices.
[0125] The core of edge computing is to transfer some computing tasks from the traditional cloud to terminal devices that are closer to the data source. By sinking computing tasks to terminal devices, it brings many significant advantages. On the one hand, it greatly reduces the processing pressure on the cloud. In large-scale data processing scenarios, the cloud needs to process requests from many terminals at the same time. If all calculations rely on the cloud, it is easy to cause excessive server load, and even system congestion and delays. Completing some tasks locally on the terminal effectively shares the work of the cloud and improves the stability and scalability of the entire system.
[0126] At the sales site, sales staff often use mobile devices to show customers product information, process orders, etc. Through edge computing, some computing tasks can be completed directly on the sales staff's mobile devices. For example, when the sales staff shows the customer a 3D model of the product, the rendering and preliminary data processing of the model are performed on the local device without uploading a large amount of data to the cloud, which greatly reduces the burden on the cloud and improves the response speed of the display. Customers can see the details of the product more smoothly and quickly, improving the experience. At the same time, in order to meet the needs of sales staff to use mobile devices for a long time outside, the device has been optimized in algorithms and hardware to reduce power consumption. For example, the device will intelligently adjust the screen brightness, shut down unnecessary background processes, etc., so that the battery life is longer, ensuring that sales staff can use the device normally throughout the day's work, and will not affect sales work due to power exhaustion.
[0127] In video conferences, especially large-scale video conferences with multiple participants, a large amount of audio and video data needs to be processed. Using edge computing, the terminal devices of participants can pre-process the local audio and video, such as optimizing image quality and reducing noise, and then upload the processed data to the cloud for integration and distribution. This can not only reduce the pressure on cloud servers, but also speed up local data processing, reduce video freezes and delays, and make conference communication smoother. In addition, in order to ensure that participants can participate in video conferences anytime and anywhere, mobile devices and IoT devices (such as smart conference tablets) have adopted low-power designs. For example, the chip adopts an energy-saving architecture and automatically enters low-power mode when the device is not in use, thereby extending battery life and allowing users to successfully complete important meetings without a power outlet.
[0128] Those of ordinary skill in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be implemented by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in this application may include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.
[0129] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.
[0130] The above-mentioned embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the invention patent. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the patent of the present application shall be subject to the attached claims.
Claims
1. A call assistance method, characterized in that: The method comprises: Get call voice information in real time; Performing real-time text recognition on the call voice information to obtain the call content; Based on the content of the call, estimating a target topic; wherein the target topic is a topic involved in the call; Displaying relevant information of the target topic; The relevant information of the target topic is pre-stored in a database; the database is used to obtain a data modification request and respond to it, so as to modify the relevant information corresponding to different target topics, or automatically modify the database based on the call content; Based on the call content, identifying the caller's emotions in real time; When the caller's emotion is the preset emotion, a prompt is issued; Based on the content of the call, determining whether the caller has a target intention; If there is a target intention, a prompt is issued; wherein, when the call content is sales, the target intention includes the caller's intention to receive sales and the topics involved: When a video call is made, a preset recognition model is input into the content of the video call to identify the target topic, whether there is a target intention, user intention and emotion.
2. The method according to claim 1, characterized in that Also includes: When the preset conditions are met, the system enters the intelligent voice mode and responds with intelligent voice: Among them, the preset conditions are determined in advance based on the target topic and the emotions of the caller.
3. The method according to claim 1, characterized in that: Also includes: During a call, if a new agent joins the call, the previous call content will be summarized and displayed.
4. The method according to claim 1, characterized in that: Also includes: Through sound removal and enhancement technology, background noise and far-field technology in calls are identified and removed to improve voice clarity and audibility, eliminate echoes and adjust volume.
5. The method according to claim 1, characterized in that Also includes: Real-time voice translation when callers use different voices.
6. The method according to claim 1, characterized in that Also includes: Determining whether preset illegal content is obtained in the call content; When the call content contains preset illegal content, a prompt is issued.
7. The method according to claim 1, characterized in that The real-time acquisition of call voice information includes: The call voice is collected and stored based on a preset recording module to obtain the call voice information; The preset recording module is specifically used to obtain uplink audio data and at least one downlink audio data after the call connection is established; wherein the uplink audio data and the downlink audio data are the call voice information; Storing the call voice information; The uplink audio data and the downlink audio data in the call voice information are matched to obtain a section of uplink audio data and a section of downlink audio data corresponding to each other as the target call voice information.
8. The method according to claim 1, characterized in that Also includes: By analyzing the caller's voice features, identity recognition and authentication are performed.
9. The method according to claim 1, characterized in that: Also includes: The call data is encrypted and anonymized.
10. The method according to claim 1, characterized in that Also includes: When a preset voice command is detected during a call: Based on the preset voice command, the device is controlled to execute a preset task.
Citation Information
Cited By
Hot news real-time pushing method and system based on new video call
CN120785943A
