Control method and device of recording management system, and storage medium
The recording files are marked through speech recognition and semantic understanding technology, which solves the problem of low retrieval efficiency in recording review, and realizes rapid retrieval and efficient detection of agent services.
Patent Information
- Application Number
- CN202411506995.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-10-28
- Publication Date
- 2024-11-26
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
In the prior art, key information cannot be retrieved quickly during the recording and review of the call between the agent and the customer, resulting in low quality inspection efficiency and long inspection cycle.
By obtaining recording files and speech text, based on speech recognition technology and semantic understanding technology, target audio clips are determined and audio marks are generated, and the display is associated in the display area to achieve rapid retrieval of recorded content.
It improves the inspection efficiency of quality inspectors, reduces unnecessary time overhead, and improves the real-time inspection efficiency and quality of agent services.
Smart Images

Figure CN119025702A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the field of speech recognition technology, and in particular to a control method, device and storage medium for a recording management system. Background Art
[0002] At present, based on the requirements for agent service specifications, it is necessary to control the call recordings between agents and customers and check the agent's service process. In related technologies, after the customer service agent and the customer finish the telephone communication service, when checking the agent's recording, the text of the speech-to-text is identified by keywords in the keyword library. When the quality inspector reviews the speech text, he cannot determine the specific location corresponding to the key information and needs to re-test the recording file. Therefore, when the number of recording files of the agent service is large and the number of recording files is limited, it takes a lot of time to check the complete process of each recording file, resulting in a long inspection cycle for the agent's service and low inspection efficiency.
[0003] The above contents are only used to assist in understanding the technical solution of the present application and do not constitute an admission that the above contents are prior art. Summary of the invention
[0004] The main purpose of this application is to provide a control method, device and storage medium for a recording management system, aiming to solve the technical problems in the related technology that the call recordings between agents and users need to be re-tested, key information cannot be quickly retrieved, and the efficiency of retrieving recordings is low.
[0005] To achieve the above-mentioned purpose, the present application provides a control method for a recording management system, which includes the following steps: obtaining a recording file and a voice text corresponding to the recording file; determining a target audio segment based on the recording file and the voice text, and generating an audio tag corresponding to the target audio segment; and upon receiving a recording file display instruction, displaying the audio tag of the recording file in an associated manner in a display area.
[0006] In one embodiment, the control method of the recording management system comprises the following steps: Obtaining a recording file and the voice text corresponding to the recording file; Determine a target audio segment based on the recording file and the speech text, and generate an audio tag corresponding to the target audio segment; When a recording file display instruction is received, the audio tag of the recording file is displayed in association in the display area.
[0007] In one embodiment, the step of determining a target audio segment based on the recording file and the speech text, and generating an audio tag corresponding to the target audio segment includes at least one of the following: Traversing the word segmentation of the speech text according to a preset keyword library, marking the keywords in the speech text and determining the keyword audio segments corresponding to the keywords; The audio features of the recording file are obtained, the target audio segment with abnormal audio features and abnormal semantic features in the recording file is determined, and the text content corresponding to the target audio segment is marked.
[0008] In one embodiment, the step of obtaining the audio features of the recording file, determining the target audio segment with abnormal audio features and abnormal semantic features in the recording file, and marking the text content corresponding to the target audio segment includes: Performing audio recognition and semantic recognition on the recording file based on natural language processing technology to extract the audio features and the semantic features; According to a preset audio fluctuation range and a preset semantic specification, the text content whose audio features do not conform to the preset audio fluctuation range and whose semantic features conform to the preset semantic specification is marked.
[0009] In one embodiment, the step of marking the text content whose audio features do not conform to the preset audio fluctuation range and whose semantic features conform to the preset semantic specification according to the preset audio fluctuation range and the preset semantic specification includes: The voice of the recording file is traversed based on the preset audio fluctuation range, and the sentences of the voice text are traversed based on the preset semantic norms, wherein the preset audio fluctuation range includes volume and tone, and the preset semantic norms include operation norms, service awareness, service skills and service etiquette; The text content corresponding to the voice in the recording file that does not conform to the preset audio fluctuation range and the sentences that conform to the preset semantic specifications are highlighted and marked.
[0010] In one embodiment, the step of obtaining a recording file and a voice text corresponding to the recording file includes: Capturing an original speech signal and preprocessing the original speech signal; Extracting feature information of the preprocessed original speech signal; The feature information is decoded based on a speech model to determine that the target word sequence is the speech text, wherein the speech model includes an acoustic model and a language model.
[0011] In one embodiment, when receiving the recording file display instruction, the step of displaying the audio tag of the recording file in the display area in association with the audio tag of the recording file includes: associating the target audio segment with the audio tag; The display instruction is received, and the audio mark of the recording file is displayed in the display area.
[0012] In one embodiment, after the step of receiving the display instruction and displaying the audio mark of the recording file in the display area, the method further includes: Obtaining the tag corresponding to the triggering operation of the audio tag, and determining the duration range of the voice segment corresponding to the tag; The voice clip is played according to the start time point and the end time point of the duration range.
[0013] In one embodiment, after the step of displaying the audio tag of the recording file in the display area in association with the recording file upon receiving the recording file display instruction, the method further includes: Determining a processing priority of the voice text based on the customer evaluation corresponding to the voice text; If the processing priority is an emergency state, the recording file is pushed to manual customer service for processing; Reply to the customer evaluation based on the result of the manual customer service processing the recording file.
[0014] In addition, to achieve the above-mentioned purpose, the present application also provides a control device for a recording management system, the control device for the recording management system comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, the computer program being configured to implement the steps of the control method for the recording management system as described above.
[0015] In addition, to achieve the above-mentioned purpose, the present application also provides a storage medium, which is a computer-readable storage medium, and a computer program is stored on the storage medium. When the computer program is executed by the processor, the steps of the control method of the recording management system as described above are implemented.
[0016] The present application provides a control method for a recording management system, which obtains a recording file and a voice text corresponding to the recording file; determines a target audio segment based on the recording file and the voice text, and generates an audio tag corresponding to the target audio segment; and when a recording file display instruction is received, displays the audio tag of the recording file in a display area in association with the audio tag of the recording file.
[0017] In summary, this application combines speech recognition technology and semantic understanding technology to mark voice text, achieve rapid retrieval of recorded content, effectively improve the quality of agent service, and improve the efficiency of real-time detection of agent service. BRIEF DESCRIPTION OF THE DRAWINGS
[0018] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present application and, together with the description, serve to explain the principles of the present application.
[0019] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.
[0020] Figure 1 A flowchart of a control method for a recording management system according to an embodiment of the present invention; Figure 2 This is a flow chart of customer evaluation processing in Embodiment 1 of the control method of the recording management system of this application; Figure 3 A flowchart diagram of a second embodiment of the control method of the recording management system of the present application; Figure 4 This is a flowchart of audio tag processing in Embodiment 2 of the control method of the recording management system of the present application; Figure 5 A flowchart diagram of a third embodiment of the control method of the recording management system of the present application; Figure 6 A flowchart diagram of a fourth embodiment of the control method of the recording management system of the present application is provided; Figure 7 Schematic diagram of the device structure of the hardware operating environment involved in the control method of the recording management system in the embodiment of the present application.
[0021] The realization of the purpose, functional features and advantages of this application will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION
[0022] 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.
[0023] In order to better understand the above technical solution, the above technical solution will be described in detail below in conjunction with the accompanying drawings and specific implementation methods.
[0024] The main solution of the embodiment of the present application is: obtaining a recording file and a voice text corresponding to the recording file; determining a target audio segment based on the recording file and the voice text, and generating an audio tag corresponding to the target audio segment; and upon receiving a recording file display instruction, displaying the audio tag of the recording file in association in a display area.
[0025] Since the relevant technology only searches for keywords in the agent recordings and does not determine the location information of the keywords, when reviewing the voice content, the quality inspector needs to search again to determine the key information. With limited quality inspectors, when the number of agent service recording files is large, it takes a lot of time to complete the process inspection of each recording file, resulting in a long inspection cycle for the agent's service and low inspection efficiency.
[0026] The present application provides a control method for a recording management system, which combines speech recognition technology with semantic understanding technology, checks the speech text obtained by speech conversion based on a preset keyword library, and marks and associates the checked target audio segments, thereby realizing rapid retrieval of key information, solving the tediousness of re-retrieval during review, reducing unnecessary time expenditure, and improving the inspection efficiency of quality inspectors.
[0027] In order to better understand the above technical solution, exemplary embodiments of the present application will be described in more detail below with reference to the accompanying drawings. Although exemplary embodiments of the present application are shown in the accompanying drawings, it should be understood that the present application can be implemented in various forms and should not be limited by the embodiments described herein. On the contrary, these embodiments are provided to enable a more thorough understanding of the present application and to fully convey the scope of the present application to those skilled in the art.
[0028] It should be noted that the execution subject of this embodiment can be a voice recognition device, or a computing service device with data processing, network communication and program running functions, such as a tablet computer, a personal computer, a mobile phone, etc., or an electronic device capable of realizing the above functions, etc., and this embodiment does not specifically limit this. The following takes a voice recognition device as an example to illustrate this embodiment and the following embodiments.
[0029] Based on this, the first embodiment of the present application provides a control method for a recording management system, referring to Figure 1 , Figure 1 This is a flow chart of Embodiment 1 of the control method of the recording management system of the present application.
[0030] In this embodiment, the control method of the recording management system includes steps S100 to S300: Step S100, obtaining a recording file and a voice text corresponding to the recording file.
[0031] It should be noted that the recording file refers to the original voice record of the conversation between the agent and the customer during the service process, which is saved in the form of a file. Speech text refers to the conversion of the voice content in the recording file into written text data through the Automatic Speech Recognition (ASR) technology, including the steps of sound signal collection, preprocessing, feature extraction, acoustic model recognition and language model decoding.
[0032] As an optional implementation, after acquiring the recording file, natural language processing technology is used to perform speech-to-text operations on the recording file to obtain a speech text corresponding to the recording file.
[0033] Exemplarily, after receiving the recording file of the agent service, the recording file is converted into text, the voice signal is preprocessed, and then the voice features are extracted. Based on the acoustic model and language model, the extracted features are decoded to obtain the corresponding word sequence as voice text.
[0034] Step S200: determining a target audio segment based on the recording file and the speech text, and generating an audio tag corresponding to the target audio segment.
[0035] It should be noted that audio tags refer to tags added to the identified target audio clips, which are used to mark the importance of this text. Tags highlight the compared keywords, which can be underlined or bolded. The highlighting operation is not limited here. The main purpose is to make the keywords conspicuous and easy to identify. The target audio clip refers to one or more continuous or intermittent audio areas identified in the original recording file according to specific needs. These clips may contain key information, such as tags for keywords. Keywords can be sensitive words, including words involving privacy, controversy or inappropriate content, etc., or they can be banned words. The scope of keywords is not limited here, and different requirements are set based on specific scenarios.
[0036] As an optional implementation, audio features are extracted from the recording file, which may be time domain features, frequency domain features (such as Mel Frequency Cepstrum Coefficient (MFCC)), time-frequency features, etc. Then, the audio is converted into text through speech recognition technology and compared with the known speech text to determine the target audio segment. The target segment can then be marked based on the anomaly detection of the audio features and text content. For example, automatic prosodic boundary annotation can be used to combine the silence duration features and text features for prosodic boundary annotation.
[0037] Exemplarily, audio features are extracted from the agent recording file, and then the audio is converted into text using speech recognition technology. The text is semantically analyzed to find segments that are not related to the agent service. Finally, the audio features and text content are combined to mark abnormal segments.
[0038] Step S300, when receiving a recording file display instruction, displaying the audio tag of the recording file in a display area in association with the audio tag of the recording file.
[0039] It should be noted that the recording file display instruction is to perform the "open file" operation on the recording file on the front-end page, and manually process the recording file by operating the peripheral. The display area can be the computer desktop or a small window in a specific client page. The display area is not limited here. The association is to associate the recording file with the audio tag for display. When the recording file is opened, the tag will also be attached to the recording file.
[0040] As an optional implementation, when receiving a display instruction, the system first queries the database to obtain metadata of the recording file, including audio tags. These audio tags may include keywords, speaker information, etc. After the system retrieves this information, it associates it with the recording file and displays it in the display area.
[0041] For example, in a recording clip of an agent service, a quality inspector opens an automatically marked recording file, and then displays the recording file on the client, and when reading, audio tags corresponding to keywords are left on the recording file and displayed.
[0042] Optionally, in this embodiment, after step S300, steps S400 to S600 are further included, referring to Figure 2 , Figure 2 The flowchart of steps S400 to S600 is as follows: Step S400: determining a processing priority of the voice text based on the customer evaluation corresponding to the voice text.
[0043] It should be noted that customer evaluation refers to the real evaluation of the agent service collected by customers in the user feedback community, which can be used to adjust the service quality of the agent, such as (very good service attitude), (very enthusiastic service), (poor service quality). Priority refers to the different importance of an event, so the order of processing can be changed accordingly.
[0044] As an optional implementation, after the agent service corresponding to the voice text is completed, the customer evaluates the agent service quality, and the processing priority corresponding to the language text is determined based on the evaluation result and the voice text.
[0045] For example, after collecting customers' evaluations of a certain agent's service quality, the administrator will make corresponding adjustments to the agent based on the service quality evaluation. For example, if the agent has a very good service attitude, the administrator may give appropriate praise and issue a small bonus, which is a medium priority; if the agent has a very enthusiastic service, the administrator may increase the salary and promote the agent, which is a low priority; if the agent has a poor service quality, the administrator may reduce the salary or have a conversation to improve the working status, which is a high priority.
[0046] Step S500: If the processing priority is an emergency state, the recording file is pushed to manual customer service for processing.
[0047] It should be noted that in this embodiment, the emergency status is the result of a comprehensive analysis of customer reviews and voice texts, and the priority of processing the voice texts is urgent and needs to be processed immediately. Human customer service is handled by humans rather than conversational robots, which plays a role in attaching importance to feedback.
[0048] As an optional implementation, the processing priority of the voice text is determined based on the customer's evaluation and the corresponding voice text. When the processing priority is urgent, it needs to be pushed to manual customer service for processing.
[0049] For example, when receiving negative feedback from a customer, the quality inspector combines the recording file and finds that the agent has made a major service error during the service process. Therefore, the processing time is rated as an emergency event, and the recording clip will be pushed to the manual customer service for direct processing without entering the automatic analysis sequence queue for processing.
[0050] Step S600: reply to the customer evaluation based on the result of the manual customer service processing the recording file.
[0051] It should be noted that when the result of the recording file is received, the customer evaluation should be responded to in a timely manner.
[0052] As an optional implementation, the result of receiving the recording file sent to the manual customer service is used to reply to the customer evaluation and follow up the customer evaluation with a negative review in a timely manner.
[0053] Exemplarily, after the recording segment is pushed to the manual customer service for direct processing, the manual customer service listens to the entire recording file, manually checks the agent service quality in the recording file, and promptly generates a corresponding response plan to reply to the customer's evaluation. The response process can be based on the phone number reserved by the customer, and promptly call back to feedback the processing results, or the processing results can be notified in the form of text messages.
[0054] In this embodiment, the recording file is transcribed into text to obtain the corresponding voice text, the target audio segment is extracted from the recording file, the corresponding audio tags are generated, and when the user requests to display the recording file, these tags are associated with the recording file and displayed on the interface, so that the user can quickly identify and access specific audio content, facilitate the quality inspector's quick review, and improve the quality inspection efficiency of the quality inspector.
[0055] Based on the first embodiment, the second embodiment of the present application provides a control method for a recording management system, referring to Figure 3 , Figure 3 This is a flow chart of the second embodiment of the control method of the recording management system of the present application, step S200 includes steps S210 to S220: Step S210, traversing the word segments of the speech text according to a preset keyword library, marking the keywords in the speech text and determining the keyword audio segments corresponding to the keywords.
[0056] It should be noted that the keyword library refers to a set of pre-defined keywords in the voice text, which is used to match and identify important words in the search process in the voice text. It can be sensitive words, obscene words, banned words, and generally words whose lexical nature is determined based on the meaning of the phrase itself. The keyword voice segment refers to the audio segment corresponding to the keyword identified in the voice text. Each keyword corresponds to the time period in the recording file where the keyword appears.
[0057] As an optional implementation, based on a pre-built preset keyword library containing all sensitive words and banned words, the voice text is then traversed using a regular expression to identify specific word patterns. It should be noted that the regular expression is used to identify and obtain text fragments that meet specific patterns. Whenever the regular expression matches a word in the preset keyword library, the word is marked in the voice text, and the matched keyword is marked, which can be highlighted.
[0058] For example, after the text of the recorded content after the agent service is completed is converted, the preset keyword library contains swear words, obscene words, etc., and regular expressions are used to match the keywords. When obscene words are traversed, the words will be automatically marked and displayed in bold. After the traversal is completed, all voice clips containing sensitive words will be marked.
[0059] As another optional implementation, the speech text is traversed by word segmentation traversal, and a preset keyword library is matched during the traversal process. If a word matching the preset keyword library appears, the word is marked and highlighted.
[0060] Exemplarily, the speech text is first segmented into separate words or phrases by word segmentation and traversal, and then the segmented speech text is traversed to determine matching key words based on the standards of a preset keyword library, which may be banned words that indicate that the agent service does not meet the specified requirements, such as politically sensitive words.
[0061] Step S220, obtaining audio features of the recording file, determining a target segment with abnormal audio features and abnormal semantic features in the recording file, and marking text content corresponding to the target audio segment.
[0062] It should be noted that audio features refer to parameters extracted from audio signals to represent sound characteristics, such as Mel-frequency cepstral coefficients, frequency domain features, etc. Audio tagging is the process of marking specific segments in audio signals. It can be a tag for keywords or a tag for sentences, which is determined by audio features and semantic features.
[0063] As an optional implementation method, the acoustic features and semantic features of the recording file are input into the acoustic model and language model for analysis respectively. If a speech segment with abnormal audio features or semantic features appears, the text content corresponding to the speech segment is marked. The audio features may be volume, pitch, etc., while the semantic features are a comprehensive analysis of the tone and context of the speech context.
[0064] Exemplarily, the voice signal of the recording file is identified, the acoustic features and semantic features of the recording file are identified, and the acoustic features and language features are input into the corresponding model for analysis. For example, if the volume of the agent suddenly increases, or the volume has been in a low state, and the tone of the context, operating standards, service awareness, service skills, workplace etiquette, etc. in the context, such as "Thank you...", "Please...", determine the audio features and semantic features of the agent's service, and mark out abnormal fragments.
[0065] Optionally, step S220 includes steps S221-S222: Step S221, performing audio recognition and semantic recognition on the recording file based on natural language processing technology to extract the audio features and the semantic features.
[0066] It should be noted that natural language processing technology is a tool for artificial intelligence to understand human language.
[0067] As an optional implementation method, semantic preprocessing including text cleaning, part-of-speech tagging, stem extraction, feature extraction, etc. is first required to perform simple processing on the second language text.
[0068] For example, an audio processing library is used to load the recording file and perform preprocessing, including sampling and denoising. Then, audio features, such as Mel frequency cepstral coefficients, are extracted. Then, speech recognition technology is used to convert the audio into text and identify the semantic information in the speech. Natural language processing technology is used to perform part-of-speech tagging and syntactic analysis to extract the semantic features of the text.
[0069] Step S222: According to a preset audio fluctuation range and a preset semantic specification, the text content whose audio features do not conform to the preset audio fluctuation range and whose semantic features conform to the preset semantic specification is marked.
[0070] It should be noted that the audio fluctuation range refers to the decibel difference between the quietest and loudest sounds in an audio file, and in this embodiment, the preset audio fluctuation range refers to setting the audio fluctuation within a certain range, which can detect and mark abnormal situations that do not meet the preset audio fluctuation range. The preset semantic specification is to analyze the text content according to the context, thereby determining the semantic features of the text content and marking the text content that does not meet the semantic specification.
[0071] As an optional implementation, firstly, audio features are extracted from the audio file, such as time domain features, frequency domain features, time-frequency features, etc. Then, the audio is converted into text by speech recognition technology, and compared with the known speech text to determine the target audio segment. Then, the target segment can be marked based on the semantic analysis of the audio features and text content. For example, an automatic prosodic boundary marking system can be used, which combines the silence duration features and text features to mark the prosodic boundaries.
[0072] For example, it is necessary to identify the speech segments of the speakers in a conference recording, use the librosa library to load the recording file, perform preprocessing, extract the MFCC features of the recording, and use the speech recognition model to convert the audio into text. Through the Neuro-Linguistic Programming (NLP) technology, perform semantic analysis on the converted text to identify key semantic content. Combine audio features and semantic features to mark the audio segments containing key speeches.
[0073] As another optional implementation, audio features are extracted from the recording file, including silence duration and spectral features, the audio is converted into text using speech recognition technology, and the semantic features of the text are extracted using natural language processing technology. Then, according to the preset audio fluctuation range and preset semantic specifications, the audio features and semantic features are matched and analyzed, and audio clips that do not meet the preset audio fluctuation range and text content that meets the semantic specifications will be marked.
[0074] For example, when detecting the recording of the agent service call, the service target segment is the segment of customer dissatisfaction. Therefore, the audio processing tool is first used to extract the audio features of the recording file, which can be the length of silence. The audio is converted into text through speech recognition, and the NLP tool is used to analyze the speech content of the text. According to the preset audio fluctuation range, for example, the segment with a sudden increase in volume, semantic norms, and text containing negative emotions, the audio segment of customer dissatisfaction is identified, and then the text content corresponding to the segment is marked.
[0075] Optionally, in this embodiment, after step S222, steps S223 to S224 are also included, referring to Figure 4 , Figure 4 The flowchart of steps S223 to S224 is as follows: Step S223, traversing the voice of the recording file based on the preset audio fluctuation range, and traversing the sentences of the voice text based on the preset semantic specifications, the preset audio fluctuation range includes volume and tone, and the preset semantic specifications include operating specifications, service awareness, service skills and service etiquette.
[0076] As an optional implementation, define the audio fluctuation range, such as the volume level and frequency range of the tone, and traverse the recording file according to semantic specifications, including keywords and phrases such as operating specifications, service awareness, and service skills, analyze the characteristics and semantic features of each audio, detect the volume of the sentence, and whether the audio is within the range, and mark the sentences that do not meet the preset conditions.
[0077] For example, in a telephone recording of a seat service, it is necessary to check whether the customer service representative complies with the service specifications. The preset range of the preset audio is the normal call volume, and the preset range of the tone is the natural voice conversation tone. Define semantic specifications, including "operation specifications", "service awareness", "service skills", and "service etiquette", traverse the sentences of the recording file, and analyze the audio features and semantic features of each sentence.
[0078] Step S224, highlighting and marking the text content corresponding to the voice in the recording file that does not conform to the preset audio fluctuation range and the sentences that conform to the preset semantic specification.
[0079] It should be noted that highlighting is to mark out the audio content in the recording file that meets the preset audio fluctuation range and preset semantic specifications.
[0080] As an optional implementation, if the set audio features and semantic features meet the inspection specifications, the text is annotated by a text annotator.
[0081] For example, words that meet the semantic feature inspection standards, such as professional terminology and contextual logical coherence, are analyzed and highlighted.
[0082] In this embodiment, feature information is extracted from the recording file and evaluated, word segmentation and traversal are performed through a preset keyword library and the converted speech text, keywords in the text are identified and marked, and the exact location of these keywords in the recording file is determined. Then, natural language processing technology is used to perform in-depth audio recognition and semantic recognition on the recording, and audio features and semantic features are extracted. According to the preset audio fluctuation range and semantic specifications, text content whose audio features do not meet the standards but whose semantic features meet the preset specifications is screened out and marked, thereby achieving rapid positioning of the recording and improving the efficiency and quality of speech quality inspection.
[0083] Based on any of the above embodiments, the third embodiment of the present application provides a control method for a recording management system, referring to Figure 5 , Figure 5 This is a flow chart of the third embodiment of the control method of the recording management system of the present application, step S100 includes steps S110 to S130: Step S110: capturing an original speech signal and preprocessing the original speech signal.
[0084] It should be noted that the original speech signal refers to the initial content of the recording file, which has not been processed in any way. Preprocessing is to remove background noise from the original speech signal, adjust the volume level for normalization, apply pre-emphasis to enhance the high-frequency part, improve the quality of the speech signal, reduce noise, and eliminate unnecessary information.
[0085] Step S120: extracting feature information of the preprocessed original speech signal.
[0086] It should be noted that after the preprocessing operation is completed, the speech signal is subjected to preliminary feature extraction of key information for speech recognition.
[0087] Step S130, decoding the feature information based on a speech model to determine that the target word sequence is the speech text, wherein the speech model includes an acoustic model and a language model.
[0088] It should be noted that the acoustic model is responsible for converting the characteristics of the speech signal into the probability distribution of acoustic units. According to the physical properties of speech, such as frequency, amplitude, etc., these properties are mapped to specific pronunciation units, while the language model is used to evaluate the linguistic rationality of word sequences and predict the possibility of word sequences based on the grammar and statistical laws of the language. The target word sequence refers to the word sequence that best conforms to the statistical probability identified from the speech model after the speech signal is input, and is used as the speech text. Decoding is the process of converting the acoustic features obtained by the acoustic model into text.
[0089] As an optional implementation, key acoustic features, such as Mel frequency cepstral coefficients, are extracted from the preprocessed original speech signal. The acoustic features represent important attributes of the speech signal, such as pitch, timbre, etc. Then, these features are decoded using the acoustic model and language model. The acoustic model can be a traditional model based on the Hidden Markov Model (HMM) or a model based on deep learning, such as a recurrent neural network (RNN), a long short-term memory network (LSTM) or a convolutional neural network (CNN), which is responsible for mapping acoustic features to possible phonemes or words. Then the language model evaluates the possibility of different word sequences based on linguistic rules and statistical information, and uses the word sequence that best conforms to grammatical and semantic rules as the output of the decoder combined with the acoustic model and language model to determine the target word sequence and form a speech text.
[0090] For example, a call of an agent service is preprocessed, including noise reduction and volume normalization. It is feasible that when a customer says "I want to cancel my subscription to this service", the acoustic model may recognize the phoneme "wanttocancemysubscription", and then the language model determines the correct word sequence as "Want to cancel my subscription" based on the context and grammatical rules, thereby converting the call content into language text.
[0091] As another optional implementation, firstly, speech signals of the speech to be recognized are collected, and minimal processing is performed on the collected speech signals. Subsequently, the speech signal waveform is input into a deep learning model, which may be a recurrent neural network or an attention mechanism enhanced Transformer model, to realize automatic learning of the mapping of speech waveform to text. In this implementation, the method of automatically learning the mapping of speech waveform to text is not limited. Thus, the recognized text is trained on a large amount of annotated speech data and output as speech text.
[0092] For example, after the agent service recording is completed, the audio file does not need to be preprocessed, and the original waveform is directly processed. The deep learning model automatically extracts the necessary acoustic and language features and generates text output to obtain the speech text.
[0093] In this embodiment, pre-processing, denoising and segmentation are first performed, and then the key feature information in the speech signal is extracted, and then the acoustic model and language model are used to decode these feature information, and the two are combined to determine the final target word sequence, that is, the speech text. This embodiment is a specific operation of speech recognition, which converts the captured original speech signal into understandable text information.
[0094] Based on any of the above embodiments, the fourth embodiment of the present application provides a control method for a recording management system, referring to Figure 6 , Figure 6 This is a flow chart of Embodiment 4 of the control method of the recording management system of the present application.
[0095] In this embodiment, step S300 includes steps S310 to S320: Step S310: Associating the target audio segment with the audio tag.
[0096] As an optional implementation, each audio segment and its corresponding audio tag may be stored together in a database or metadata format, and then a programming interface or query system may be used to quickly access a specific audio segment and retrieve its tag information.
[0097] Step S320: receiving the display instruction, and displaying the audio mark of the recording file in the display area.
[0098] As an optional implementation, a request to display audio tags is received in the front-end user interface, and then the audio tags related to the recording file are displayed in a specific display area, which can be a screen or a specific window.
[0099] For example, in an audio editing software, various marks are added, such as "start speaking" and "important comments". When the mark is added, it is displayed on the timeline progress bar.
[0100] Optionally, step S320 includes steps S321-S322: Step S321, obtaining the tag corresponding to the triggering operation of the audio tag, and determining the duration range of the voice segment corresponding to the tag.
[0101] As an optional implementation, when the event detector detects a user trigger operation, the system captures the trigger operation and identifies the triggered tag, and quickly retrieves the duration range associated with the tag stored in the database based on the tag content.
[0102] Step S322: Play the voice clip according to the start time point and the end time point of the duration range.
[0103] As an optional implementation, after determining the duration range of the corresponding voice segment, the voice segment is outputted starting from the linked start time point.
[0104] In this embodiment, by associating the target audio segment with the audio tags, the search display instruction is displayed, and these tags are displayed in the display area. An intuitive and easy-to-navigate interface is provided, allowing users to quickly locate and identify specific content in the recording file, efficiently manage and retrieve key moments in the recording, and improve work efficiency and content quality.
[0105] It should be noted that the above examples are only used to understand the present application and do not constitute a limitation on the control method of the recording management system of the present application. More simple transformations based on this technical concept are all within the protection scope of the present application.
[0106] The present application provides a control device for a recording management system, the device comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the control method of the recording management system in the above-mentioned embodiment 1.
[0107] Reference below Figure 7 , which shows a schematic diagram of the structure of a control device suitable for implementing the recording management system of the embodiment of the present application. The control device of the recording management system in the embodiment of the present application may include but is not limited to mobile terminals such as mobile phones, laptop computers, digital broadcast receivers, PDAs (Personal Digital Assistants), PADs (Portable Application Descriptions), PMPs (Portable Media Players), vehicle-mounted terminals (such as vehicle-mounted navigation terminals), etc., and fixed terminals such as digital TVs, desktop computers, etc. Figure 7 The control device of the recording management system shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present application.
[0108] like Figure 7As shown, the control device of the recording management system may include a processing device 1001 (e.g., a core processor, a graphics processor, etc.), which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM: Read Only Memory) 1002 or a program loaded from a storage device 1003 to a random access memory (RAM: Random Access Memory) 1004. In RAM1004, various programs and data required for the operation of the control device of the recording management system are also stored. The processing device 1001, ROM1002, and RAM1004 are connected to each other through a bus 1005. An input / output (I / O) interface 1006 is also connected to the bus. Generally, the following systems can be connected to the I / O interface 1006: an input device 1007 including, for example, a touch screen, a touch pad, a keyboard, a mouse, an image sensor, a microphone, an accelerometer, a gyroscope, etc.; an output device 1008 including, for example, a liquid crystal display (LCD: Liquid Crystal Display), a speaker, a vibrator, etc.; a storage device 1003 including, for example, a magnetic tape, a hard disk, etc.; and a communication device 1009. The communication device 1009 can allow the control device of the recording management system to communicate with other devices wirelessly or by wire to exchange data. Although the control device of the recording management system with various systems is shown in the figure, it should be understood that it is not required to implement or have all the systems shown. More or fewer systems can be implemented or have alternatively.
[0109] In particular, according to the embodiments disclosed in the present application, the process described above with reference to the flowchart can be implemented as a computer software program. For example, the embodiments disclosed in the present application include a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program includes a program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from a network through a communication device, or installed from a storage device 1003, or installed from a ROM 1002. When the computer program is executed by the processing device 1001, the above-mentioned functions defined in the method of the embodiment disclosed in the present application are executed.
[0110] The control device of the recording management system provided by the present application adopts the control method of the recording management system in the above embodiment, which can solve the technical problem of the control of the recording management system. Compared with the prior art, the beneficial effects of the control device of the recording management system provided by the present application are the same as the beneficial effects of the control method of the recording management system provided by the above embodiment, and other technical features in the control device of the recording management system are the same as the features disclosed in the method of the previous embodiment, which will not be repeated here.
[0111] It should be understood that the various parts disclosed in this application can be implemented by hardware, software, firmware or a combination thereof. In the description of the above embodiments, specific features, structures, materials or characteristics can be combined in any one or more embodiments or examples in a suitable manner.
[0112] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.
[0113] The present application provides a computer-readable storage medium having computer-readable program instructions (ie, computer programs) stored thereon, and the computer-readable program instructions are used to execute the control method of the recording management system in the above-mentioned embodiment.
[0114] The computer-readable storage medium provided in the present application may be, for example, a USB flash drive, but is not limited to electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems, systems or devices, or any combination of the above. More specific examples of computer-readable storage media may include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM: Random Access Memory), a read-only memory (ROM: Read Only Memory), an erasable programmable read-only memory (EPROM: Erasable Programmable Read Only Memory or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM: CD-Read Only Memory), an optical storage device, a magnetic storage device, or any suitable combination of the above. In this embodiment, the computer-readable storage medium may be any tangible medium containing or storing a program, which may be used by or in combination with an instruction execution system, system or device. The program code contained on the computer-readable storage medium may be transmitted using any appropriate medium, including but not limited to: wires, optical cables, RF (Radio Frequency: Radio Frequency), etc., or any suitable combination of the above.
[0115] The computer-readable storage medium may be included in the control device of the recording management system; or may exist independently without being assembled into the control device of the recording management system.
[0116] The above-mentioned computer-readable storage medium carries one or more programs. When the above-mentioned one or more programs are executed by the control device of the recording management system, the control device of the recording management system: obtains the recording file and the voice text corresponding to the recording file; determines the target audio segment based on the recording file and the voice text, and generates an audio tag corresponding to the target audio segment; and when receiving a recording file display instruction, displays the audio tag of the recording file in an associated manner in the display area.
[0117] Computer program code for performing the operations of the present application may be written in one or more programming languages or a combination thereof, including object-oriented programming languages such as Java, Smalltalk, C++, and conventional procedural programming languages such as "C" or similar programming languages. The program code may be executed entirely on the user's computer, partially on the user's computer, as a separate software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., via the Internet using an Internet service provider).
[0118] The flow chart and block diagram in the accompanying drawings illustrate the possible architecture, function and operation of the system, method and computer program product according to various embodiments of the present application. In this regard, each square box in the flow chart or block diagram can represent a module, a program segment or a part of a code, and the module, the program segment or a part of the code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the square box can also occur in a sequence different from that marked in the accompanying drawings. For example, two square boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each square box in the block diagram and / or flow chart, and the combination of the square boxes in the block diagram and / or flow chart can be implemented with a dedicated hardware-based system that performs a specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.
[0119] The modules involved in the embodiments described in this application may be implemented by software or hardware, wherein the name of the module does not constitute a limitation on the unit itself in some cases.
[0120] The readable storage medium provided in the present application is a computer-readable storage medium, which stores computer-readable program instructions (i.e., computer programs) for executing the control method of the recording management system, and can solve the technical problem of controlling the recording management system. Compared with the prior art, the beneficial effects of the computer-readable storage medium provided in the present application are the same as the beneficial effects of the control method of the recording management system provided in the above embodiment, and will not be described in detail here.
[0121] The above descriptions are only some embodiments of the present application, and are not intended to limit the patent scope of the present application. All equivalent structural changes made using the contents of the present application specification and drawings under the technical concept of the present application, or direct / indirect applications in other related technical fields are included in the patent protection scope of the present application.
Claims
1. A control method for a recording management system, characterized in that: The control method of the recording management system comprises the following steps: Obtaining a recording file and the voice text corresponding to the recording file; Determine a target audio segment based on the recording file and the speech text, and generate an audio tag corresponding to the target audio segment; When a recording file display instruction is received, the audio tag of the recording file is displayed in association in the display area.
2. The control method of the recording management system according to claim 1, characterized in that: The step of determining a target audio segment based on the recording file and the speech text, and generating an audio tag corresponding to the target audio segment comprises at least one of the following: Traversing the word segmentation of the speech text according to a preset keyword library, marking the keywords in the speech text and determining the keyword audio segments corresponding to the keywords; The audio features of the recording file are obtained, the target audio segment with abnormal audio features and abnormal semantic features in the recording file is determined, and the text content corresponding to the target audio segment is marked.
3. The control method of the recording management system according to claim 2, characterized in that: The step of obtaining the audio features of the recording file, determining the target audio segment with abnormal audio features and abnormal semantic features in the recording file, and marking the text content corresponding to the target audio segment includes: Performing audio recognition and semantic recognition on the recording file based on natural language processing technology to extract the audio features and the semantic features; According to a preset audio fluctuation range and a preset semantic specification, the text content whose audio features do not conform to the preset audio fluctuation range and whose semantic features conform to the preset semantic specification is marked.
4. The control method of the recording management system according to claim 3, characterized in that: The step of marking the text content whose audio features do not conform to the preset audio fluctuation range and whose semantic features conform to the preset semantic specification according to the preset audio fluctuation range and the preset semantic specification comprises: The voice of the recording file is traversed based on the preset audio fluctuation range, and the sentences of the voice text are traversed based on the preset semantic norms, wherein the preset audio fluctuation range includes volume and tone, and the preset semantic norms include operation norms, service awareness, service skills and service etiquette; The text content corresponding to the voice in the recording file that does not conform to the preset audio fluctuation range and the sentences that conform to the preset semantic specifications are highlighted and marked.
5. The control method of the recording management system according to claim 1, characterized in that: The step of obtaining the recording file and the voice text corresponding to the recording file includes: Capturing an original speech signal and preprocessing the original speech signal; Extracting feature information of the preprocessed original speech signal; The feature information is decoded based on a speech model to determine that the target word sequence is the speech text, wherein the speech model includes an acoustic model and a language model.
6. The control method of the recording management system according to claim 1, characterized in that: The step of displaying the audio mark of the recording file in the display area in association with the recording file upon receiving the recording file display instruction comprises: associating the target audio segment with the audio tag; The display instruction is received, and the audio mark of the recording file is displayed in the display area.
7. The control method of the recording management system according to claim 6, characterized in that: After the step of receiving the display instruction and displaying the audio mark of the recording file in the display area, the method further includes: Obtaining the tag corresponding to the triggering operation of the audio tag, and determining the duration range of the voice segment corresponding to the tag; The voice clip is played according to the start time point and the end time point of the duration range.
8. The control method of the recording management system according to claim 1, characterized in that: After the step of displaying the audio mark of the recording file in the display area in association with the recording file display instruction upon receiving the recording file display instruction, the method further includes: Determining a processing priority of the voice text based on the customer evaluation corresponding to the voice text; If the processing priority is an emergency state, the recording file is pushed to manual customer service for processing; Reply to the customer evaluation based on the result of the manual customer service processing the recording file.
9. A control device for a recording management system, characterized in that: The control device of the recording management system includes: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the computer program is configured to implement the steps of the control method of the recording management system as described in any one of claims 1 to 8.
10. A storage medium, characterized in that: The storage medium is a computer-readable storage medium, and a computer program is stored on the storage medium. When the computer program is executed by a processor, the steps of the control method of the recording management system according to any one of claims 1 to 8 are implemented.
Citation Information
Patent Citations
Method and device for playing speech on demand
CN101329867A
Method and device for examining service quality
CN107093431A
Telephone recording quality inspection method and device, equipment and medium
CN111263009A
Audio annotation method and device, computer readable medium and electronic equipment
CN113903364A
Video dotting method and device, equipment and storage medium
CN115866290A
