Telephone customer service dialect accent conversion system and method

Abstract

The invention discloses a telephone customer service dialect accent conversion system and method, and the system comprises a real-time interaction module which is used for the real-time translation of voices and characters; the dialect voice library training module is used for training and model storage of customized feature voices; and the voice real-time conversion module is used for performing voice real-time conversion according to the characters translated in real time and the trained model. Mandarin can be converted into dialects, and the language of the telephone customer service is converted into the local dialect or accent of the target customer in real time through the voice synthesis technology and acoustic training. After a customer answers a call, the customer is kept in a concerned sound environment in a dialogue process, rejection and semantic ambiguity are reduced, and efficient communication is realized.

Description

Technical Field

[0001] The invention belongs to the technical field of voice conversion, and in particular to a telephone customer service dialect accent conversion system and method. Background Art

[0002] At present, domestic telephone marketing and telephone customer service have become the main means of communication besides online communication on the Internet. However, due to the geographical location of the service provider and the dialect ability of the customer service staff, it is impossible to provide very humanized dialect services. In particular, some more professional and written words in the financial industry cannot be conveyed to customers in a faithful and elegant manner, resulting in communication efficiency problems.

[0003] The existing technical solutions are:

[0004] 1. Speech recognition: This technology is used to convert speech into text.

[0005] 2.TTS (Text To Speech): This technology is used to convert text into speech.

[0006] 3. Voice training model: It consists of a text front-end, an acoustic model, and a vocoder. This technology enables training based on the voice of customer service personnel to achieve the authenticity and uniqueness of the generated accent.

[0007] 4. Natural semantic large model: This model is used to convert Mandarin into dialect semantics.

[0008] Currently, telephone customer service and intelligent customer service systems are limited by geographical factors and the dialects of customer service personnel, and are unable to provide perfect service to customers. Due to regional and dialect differences, customers' emotions after receiving a call and their aversion to telephone conversations are greatly reduced due to differences in dialects, which greatly reduces communication efficiency. Summary of the invention

[0009] In view of the above problems, the present invention is proposed to provide a telephone customer service dialect accent conversion system and method that overcomes the above problems or at least partially solves the above problems.

[0010] To achieve the above object, the present invention adopts the following technical solutions:

[0011] A telephone customer service dialect accent conversion system, the system comprising:

[0012] Real-time interaction module, used for real-time translation of voice and text;

[0013] Dialect speech library training module, used for training and model preservation of customized characteristic speech;

[0014] The real-time speech conversion module is used to perform real-time speech conversion based on the real-time translated text and the trained model.

[0015] Optionally, the system further comprises a decision analysis server module, and the decision analysis server module is used to determine whether to use a dialect conversion service.

[0016] Optionally, the system further includes a dialect feature training server module, and the dialect feature training server module is used to generate a natural language model.

[0017] Optionally, the system further comprises a dialect synthesis server module, and the dialect synthesis server module is used to connect to speech synthesis software.

[0018] Optionally, the system further includes a communication terminal module, which is used to dispatch manual agent customer service and load decision service AI.

[0019] Optionally, the real-time speech conversion module includes a speech synthesis unit, and the speech synthesis unit includes loading a dialect model, performing speech conversion and generating speech.

[0020] The present invention also provides a method for converting telephone customer service dialect accents, the method being based on any of the aforementioned systems, the method comprising:

[0021] The human agent receives the customer's voice call;

[0022] Prompt the user to use the dialect description and obtain the dialect type;

[0023] The language answered by the human agent is converted into a dialect;

[0024] Communicate with customers in real time, completing the language conversion cycle until the communication ends.

[0025] In summary, due to the adoption of the above technical solution, the beneficial effects of the present invention are:

[0026] The present invention can convert Mandarin into dialects. Through speech synthesis technology and acoustic training, the language of telephone customer service can be converted into the local dialect or accent of the target customer in real time. After the customer answers the call, they can maintain a caring voice environment during the conversation, reduce rejection and semantic ambiguity, and achieve efficient communication. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] Figure 1 A flowchart of a telephone customer service dialect accent conversion system provided in an embodiment of the present application. DETAILED DESCRIPTION

[0028] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments in the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0029] Embodiment 1:

[0030] See also Figure 1 This embodiment provides a telephone customer service dialect accent conversion system, the system comprising:

[0031] Real-time interaction module, used for real-time translation of voice and text.

[0032] Dialect speech library training module, used for training and model preservation of customized characteristic speech.

[0033] Decision analysis server module, used to determine whether to use the dialect conversion service.

[0034] Dialect feature training server module, used to generate natural language models.

[0035] Dialect synthesis server module, used to connect to speech synthesis software.

[0036] The communication terminal module is used to dispatch human customer service agents and load decision-making service AI.

[0037] The real-time speech conversion module is used to perform real-time speech conversion based on the real-time translated text and the trained model.

[0038] The real-time speech conversion module includes a speech synthesis unit, which includes loading a dialect model, performing speech conversion and generating speech.

[0039] Specifically, this embodiment also includes the following two parts:

[0040] Part 1: Basic speech model training

[0041] 1. Telephone customer service personnel generate dialects through the training software built by the voice training server. For example, Xiao Zhang, a human customer service representative, reads Sichuan dialect and Cantonese dialect according to the dialect dictionary. The training software generates the Sichuan dialect and Cantonese of Xiao Zhang's voice. The model will be saved in the voice database of the training server. It provides a voice basis for future conversion.

[0042] Part II: Dialect Implementation Conversion

[0043] 1. Customers connect to the communication terminal through telephone or various app voice access methods. After the communication terminal is connected, an artificial customer service agent is assigned and the decision service AI is loaded. The artificial agent waits for the decision service instruction before intervening in the communication.

[0044] 2. AI uses standard Mandarin to ask whether to use dialect services. If the customer responds that they do not want to use dialect services, the manual customer service will intervene directly and use standard Mandarin for subsequent business communications, and this process ends. For example, if customer Wang wants to communicate in Mandarin, Wang can say "no" or "no", and Mandarin will be used throughout the subsequent communication.

[0045] 3. If the customer feedback is used, the decision robot will start the dialect conversion service, notify the dialect synthesis server to start the relevant service, and establish a communication link. At the same time, the human agent will be notified to start intervening in the call. For example, if customer Zhao wants to use Sichuan dialect, then Zhao can answer: "Yes, I want to continue to communicate in Sichuan dialect", then the device will use Sichuan dialect throughout the process, and convert the language of the human customer service into Sichuan dialect to communicate with Zhao.

[0046] 4. When the human customer service greets or asks a question in Mandarin, the voice is first transmitted to the voice conversion server for conversion and then sent to the customer, who can now hear the predetermined dialect. The entire communication process is circulated in this mode until the communication is completed and the customer hangs up.

[0047] This embodiment can convert Mandarin into dialects, and through speech synthesis technology and acoustic training, the language of telephone customer service can be converted into the local dialect or accent of the target customer in real time. After the customer answers the phone, they can maintain a caring voice environment during the conversation, reduce rejection and semantic ambiguity, and achieve efficient communication.

[0048] Embodiment 2:

[0049] Embodiment 2 discloses a method for converting a telephone customer service dialect and accent. The method is based on the above-mentioned telephone customer service dialect and accent conversion system, and the method comprises the following steps:

[0050] S1. The human agent receives a voice call from the customer.

[0051] S2. Prompt the user to use the dialect description and obtain the dialect type.

[0052] S3. The language of the agent's answer is converted into a dialect.

[0053] S4. Communicate with customers in real time, and complete the language conversion cycle until the communication ends.

[0054] The above contents are further detailed descriptions of the present invention in combination with specific preferred embodiments, and it cannot be determined that the specific implementation of the present invention is limited to these descriptions. For ordinary technicians in the technical field to which the present invention belongs, several simple deductions or substitutions can be made without departing from the concept of the present invention, which should be regarded as falling within the protection scope of the present invention.

Claims

1. A telephone customer service dialect accent conversion system, characterized in that: The system comprises: Real-time interaction module, used for real-time translation of voice and text; Dialect speech library training module, used for training and model preservation of customized characteristic speech; The real-time speech conversion module is used to perform real-time speech conversion based on the real-time translated text and the trained model.

2. A telephone customer service dialect accent conversion system as claimed in claim 1, characterized in that: The system further comprises a decision analysis server module, and the decision analysis server module is used to determine whether to use the dialect conversion service.

3. A telephone customer service dialect accent conversion system as claimed in claim 1, characterized in that: The system also includes a dialect feature training server module, which is used to generate a natural language model.

4. A telephone customer service dialect accent conversion system as claimed in claim 1, characterized in that: The system also includes a dialect synthesis server module, and the dialect synthesis server module is used for connecting to speech synthesis software.

5. A telephone customer service dialect accent conversion system as claimed in claim 1, characterized in that: The system also includes a communication terminal module, which is used to assign manual agent customer service and load decision-making service AI.

6. A telephone customer service dialect accent conversion system as claimed in claim 1, characterized in that: The real-time speech conversion module includes a speech synthesis unit, which includes loading a dialect model, performing speech conversion and generating speech.

7. A telephone customer service dialect accent conversion method, the method is based on a telephone customer service dialect accent conversion system according to any one of claims 1 to 6, characterized in that: The method comprises: The human agent receives the customer's voice call; Prompt the user to use the dialect description and obtain the dialect type; The language answered by the human agent is converted into a dialect; Communicate with customers in real time, completing the language conversion cycle until the communication ends.

Images