Call identification method, system and device based on TWS earphone and storage medium

Abstract

The invention discloses a call recognition method, system and device based on a TWS earphone and a storage medium. The method comprises the following steps: acquiring input and output voice information; recognizing voice content, speed, definition and sound decibel based on the voice information; based on the voice content, the voice speed, the definition and the sound decibel, comparing the voice content, the voice speed, the definition and the sound decibel with corresponding standard values, and calculating input and output scores of the call; the method comprises the following steps: correspondingly outputting a current call state based on input and output scores and a score ratio, obtaining input and output voice information, carrying out corresponding parameter identification comparison on the voice information, and respectively scoring input and output of call content, so as to obtain the call content based on the score ratio. The states of the two parties of the call can be known by outputting the call state, and whether the call is abnormal or not can be analyzed conveniently.

Description

Technical Field

[0001] The present application relates to the field of headphones, and in particular to a call identification method, system, device and storage medium based on TWS headphones. Background Art

[0002] Generally, during a conversation, a user's emotional changes will be reflected in the voice output, and the user's emotional changes will also change with the person they are talking to. Therefore, in the online customer service industry, for example, how to analyze the content of the call based on the content of the conversation to determine whether the customer service staff's call is compliant is a technical problem that needs to be solved urgently.

[0003] Therefore, the existing technology still needs to be improved Summary of the invention The purpose of this application is to determine whether a person's call is compliant through call identification.

[0004] The above technical objectives of this application are achieved through the following technical solutions: A call identification method based on a TWS headset, comprising: Get input and output voice information; Based on voice information, identify voice content, speaking speed, clarity and sound decibel; Based on the speech content, speaking speed, clarity and sound decibel, the scores of the input and output of this call are calculated by comparing them with the corresponding standard values; Based on the input and output scores and the score ratio, the status of this call is output accordingly.

[0005] The above scheme of the present application obtains the input and output voice information, performs corresponding parameter recognition and comparison on the voice information, scores the input and output of the call content respectively, and then outputs the status of the call based on the score ratio. The status of both parties in the call can be known, which is convenient for analyzing whether the call is abnormal.

[0006] Optionally, in the call recognition method based on TWS headset, the step of recognizing voice content, speech speed, clarity and sound decibel based on voice information includes: Split the conversation content in the voice information to obtain multiple single interaction information; The voice content, speaking speed, clarity and sound decibel of a single interactive message are obtained respectively.

[0007] The above-mentioned scheme of the present application obtains multiple single interaction information by splitting the conversation content, and can identify the voice parameters of each interaction, so as to analyze the state changes of the conversation and achieve accurate analysis.

[0008] Optionally, in the call identification method based on TWS headset, the step of outputting the call status based on the input and output scores and the score ratio includes: When the score output for this call is lower than the standard value, the status of this call is output as unqualified.

[0009] The above scheme of the present application determines whether the output call is qualified by comparing the output call scores. If the scores are lower than the standard value, the output call is unqualified, which is convenient for analyzing the specific reasons for the failure.

[0010] Optionally, in the call identification method based on TWS headset, the step of outputting the call status based on the input and output scores and the score ratio includes: When the score input for this call is lower than the standard value, but the score output for this call is higher than the standard value, this call is marked and the status of this call is output as abnormal.

[0011] The above scheme of the present application analyzes the input and output calls. If the output score is high but the input score is low, it means that the output voice information is compliant, but the input voice is not compliant. Therefore, the call status is marked as abnormal, which is convenient for analyzing the cause of the abnormal input voice.

[0012] Optionally, the call identification method based on TWS headset, wherein the method further includes: When the call status is abnormal, the input score and output score of the single interaction information are obtained respectively; If the input scores are all lower than the standard value, the customer is marked as abnormal; If there is a single interaction information with an input score higher than the standard value, obtain the first interaction information below the standard value and mark this interaction information.

[0013] The above scheme of the present application splits and identifies the calls. If the input score is always lower than the standard value, it means that the customer is abnormal. If there is a single interaction information that is higher than the standard value, the first interaction information below the standard value is obtained. The turning point of the user's voice score can be known, which is convenient for analyzing the cause of the abnormality.

[0014] Optionally, the call identification method based on TWS headset, wherein the method further includes: The weights of speech content, speaking speed, clarity and sound decibels are pre-set, as well as the corresponding standard values ​​and corresponding scores.

[0015] The above scheme of the present application facilitates accurate scoring by setting weights.

[0016] Optionally, in the call recognition method based on TWS headset, the steps of voice content, speech speed, clarity and sound decibel include: Set the voice content keywords, including bonus keywords and deduction keywords and the corresponding bonus and deduction points.

[0017] The above solution of this application improves the accuracy of scoring by setting keywords.

[0018] On the other hand, the present application discloses a call identification system based on a TWS headset, which includes: Voice acquisition module, voice acquisition input and output voice information; A speech recognition module is used to identify speech content, speech speed, clarity and sound decibel based on speech information; The scoring module is used to compare the speech content, speech speed, clarity and sound decibel with the corresponding standard values ​​to calculate the scores of the input and output of this call; The call status output module is used to output the call status based on the input and output scores and the score ratio.

[0019] The above scheme of the present application obtains the input and output voice information, performs corresponding parameter recognition and comparison on the voice information, scores the input and output of the call content respectively, and then outputs the status of the call based on the score ratio. The status of both parties in the call can be known, which is convenient for analyzing whether the call is abnormal.

[0020] On the other hand, the present application discloses an electronic device, which includes a memory and a processor, and the memory stores a computer program that can be loaded by the processor and execute the call identification method based on the TWS headset as described above.

[0021] On the other hand, the present application discloses a storage medium, which stores a computer program that can be loaded by a processor and execute the call identification method based on TWS headset as described above.

[0022] In summary, the present application has at least one of the following beneficial effects: 1. By obtaining the input and output voice information, the corresponding parameters of the voice information are identified and compared, and the input and output of the call content are scored respectively. Based on the score ratio, the status of the call is output to know the status of both parties of the call, which is convenient for analyzing whether the call is abnormal; 2. By splitting the conversation content and obtaining multiple single interaction information, the voice parameters of each interaction can be identified, so as to analyze the state changes of the conversation and achieve accurate analysis; 3. By analyzing the input and output calls, if the output score is high but the input score is low, it means that the output voice information is compliant but the input voice is not compliant. Therefore, the call status is marked as abnormal to facilitate analysis of the cause of the abnormal input voice. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] Figure 1 This is a step flow chart of a call identification method based on TWS headphones in the present application.

[0024] Figure 2 This is a structural block diagram of a call identification system based on TWS headphones in this application. DETAILED DESCRIPTION

[0025] The present application is further described in detail below in conjunction with the accompanying drawings.

[0026] The present application embodiment discloses a call recognition method based on a TWS headset, see Figure 1 , including: S1. Get input and output voice information; After the call is established, voice information will be obtained, including the voice information input and output of the earphone, that is, the voice information of both parties in the call. By obtaining the voice information, the call can be identified; S2. Based on the voice information, identify the voice content, speaking speed, clarity and sound decibel; Recognize voice information, including the specific content of the voice, speaking speed, voice clarity and sound decibel. The specific content can be used for subsequent conversation content analysis. The speaking speed can be used for speaking speed analysis. If the speaking speed suddenly speeds up or slows down, it may correspond to changes in the user's emotions. Therefore, when the speaking speed changes, it may indicate that the user's emotions have changed. For example, when a person is excited, the speaking speed will speed up. Similarly, when a person is excited, the clarity of the speech may also be affected, and the decibel of the speaking sound may also increase. By obtaining these voice parameters, the call status can be analyzed accordingly.

[0027] S3. Based on the voice content, speaking speed, clarity and sound decibel, the scores of the input and output of this call are calculated by comparing with the corresponding standard values ​​respectively; in the embodiment of the present application, the scores of both parties of the call are calculated by comparing each parameter with the standard value. For example, if there is negative emotion in the voice content, there will be a corresponding score, or if there is positive motivation, such as words of praise, then there will be extra points, and the increase in speaking speed will also correspond to a score. Different parameters can be assigned corresponding scores, for example, 25 points for voice content, speaking speed, clarity, and decibel each. If the standard value is reached, the score of the standard value can be set. For example, if the standard value is 20 points, if the standard value is reached, 20 points will be obtained. If there is extra points, they will be added on the basis of 20 points. The maximum score for each parameter is 25 points, and the overall score is obtained based on the score.

[0028] S4. Based on the input and output scores and the score ratio, the current call status is output accordingly.

[0029] In the embodiment of the present application, after respectively obtaining the input and output scores, the score ratio is also obtained. For example, if the score of the input end is high, but the score of the output end is low, it is possible that the call on the customer service side is normal, but the client is in a bad mood, and a specific analysis can be performed later. Alternatively, if the input and output scores are both above the standard value, and the ratio is close to 1, it means that the call status is relatively good. Based on the score and the score ratio, the corresponding call status can be obtained to facilitate subsequent analysis.

[0030] In some possible embodiments of the present application, the call recognition method based on TWS headset, wherein the step of recognizing voice content, speech speed, clarity and sound decibel based on voice information includes: Split the conversation content in the voice information to obtain multiple single interaction information; The voice content, speaking speed, clarity and sound decibel of a single interactive message are obtained respectively.

[0031] In the embodiments of the present application, during a call, a person's emotions may fluctuate, that is, they may be normal at the beginning, become abnormal during the call, and then return to normal again. Therefore, in order to accurately analyze the call status, the conversation content in the voice information is split to obtain multiple single interaction information. By acquiring parameters for the single interaction information, it is possible to determine whether the status of the single interaction information is normal, thereby accurately analyzing the cause of the abnormality.

[0032] In some possible embodiments of the present application, the call identification method based on TWS headset, wherein the step of outputting the call status correspondingly based on the input and output scores and the score ratio includes: When the score output for this call is lower than the standard value, the status of this call is output as unqualified.

[0033] In the embodiment of the present application, when judging the call status, the professional ethics of the customer service is taken into consideration. Therefore, during the entire call process, the input end, that is, the customer, has emotions, and for the output end, the emotions must be stable. Therefore, when judging the call status, if the score output of this call is lower than the standard value, the call status is output as unqualified. Through the above scheme, the attitude of customer service personnel can be standardized.

[0034] In some possible embodiments of the present application, the call identification method based on TWS headset, wherein the step of outputting the call status correspondingly based on the input and output scores and the score ratio includes: When the score input for this call is lower than the standard value, but the score output for this call is higher than the standard value, this call is marked and the status of this call is output as abnormal.

[0035] In the embodiments of the present application, it is mentioned in the aforementioned embodiments that the score at the input end may be lower than the standard value, that is, the customer is in a bad mood, but the call score at the output end, that is, the customer service side, is higher than the standard value, which means that there is no abnormality at the output end, but for the input end, it is necessary to analyze the reason for the low call score at the input end. Therefore, the call status will be marked as an abnormal status to facilitate the subsequent analysis of the specific cause of the abnormality.

[0036] In some possible embodiments of the present application, the call identification method based on TWS headset further includes: When the call status is abnormal, the input score and output score of the single interaction information are obtained respectively; If the input scores are all lower than the standard value, the customer is marked as abnormal; If there is a single interaction information with an input score higher than the standard value, obtain the first interaction information below the standard value and mark this interaction information.

[0037] In the embodiment of the present application, if the status of this call is abnormal, as mentioned in the above embodiment, abnormality means that the call at the output end is normal, but the input end is lower than the standard value. Therefore, a single input score and output score are obtained respectively. If the input scores are lower than the standard values, it means that the call is abnormal from the beginning to the end. At this time, the customer is marked as abnormal, and it is necessary to focus on analyzing whether it is a historical problem in the future. If the input score is higher than the standard value, it means that the customer is in a normal call state during the call. At this time, it is necessary to analyze the turning point, obtain the interactive information from the first time it is lower than the standard value, and mark it to facilitate the analysis of the cause of the turning point. If the input end starts to be lower than and then higher than, the last call below the standard value is obtained, and it can be analyzed how the output end makes the input end call normal. Both positive feedback and negative feedback can be provided to improve the standardization of customer service.

[0038] In some possible embodiments of the present application, the call identification method based on TWS headset further includes: The weights of speech content, speaking speed, clarity and sound decibels are pre-set, as well as the corresponding standard values ​​and corresponding scores.

[0039] In the embodiments of the present application, the aforementioned embodiments mentioned that the overall score is determined based on the score of each parameter, so different parameters may have different corresponding weights. For example, voice content is the most direct feedback on the user's call status, and can have a higher weight. For example, the total score is 100 points, the voice content weight is 0.5, the speaking speed weight is 0.1, the clarity is 0.1, and the sound decibel is 0.3. The corresponding score of each parameter is 80 points for voice content, 50 points for speaking speed, 60 points for clarity, and 50 points for sound decibel. The overall score = (80*0.5+50*0.1+60*0.1+50*0.3) = 40+5+6+15=66 points; by setting weights, the call status can be analyzed more accurately and the accuracy can be improved.

[0040] In some possible embodiments of the present application, the call recognition method based on TWS headset, wherein the steps of voice content, speech speed, clarity and sound decibel include: Set the voice content keywords, including bonus keywords and deduction keywords and the corresponding bonus and deduction points.

[0041] The aforementioned embodiments of the present application mentioned that during a call, there will be some keywords that add or subtract points. For example, if the user uses words of praise or swear words during a call, points can be added or subtracted accordingly, because keywords are the most direct reflection of the user's attitude. Of course, some catchphrases can be analyzed in combination with the context. For example, if the customer uses swear words at the beginning of every sentence, but the other aspects of the speech speed, clarity, decibel level, etc. are normal, no points will be subtracted accordingly.

[0042] Another embodiment of the present application discloses a call identification system based on a TWS headset, see Figure 2 , including: Speech acquisition module, 100, speech acquisition input and output speech information; The speech recognition module 200 is used to recognize speech content, speech speed, clarity and sound decibel based on the speech information; Scoring module 300, for calculating the scores of the input and output of the call based on the speech content, speech speed, clarity and sound decibel, respectively, by comparing with the corresponding standard values; The call status output module 400 is used to output the current call status based on the input and output scores and the score ratio.

[0043] The specific implementation of each module of the above system of the present application has been described in detail in the method steps, so it will not be repeated here.

[0044] Another embodiment of the present application discloses an electronic device, which includes a memory and a processor, and the memory stores a computer program that can be loaded by the processor and executes the following call identification method based on a TWS headset: Get input and output voice information; Based on voice information, identify voice content, speaking speed, clarity and sound decibel; Based on the speech content, speaking speed, clarity and sound decibel, the scores of the input and output of this call are calculated by comparing them with the corresponding standard values; Based on the input and output scores and the score ratio, the status of this call is output accordingly.

[0045] Split the conversation content in the voice information to obtain multiple single interaction information; The voice content, speaking speed, clarity and sound decibel of a single interactive message are obtained respectively.

[0046] When the score output for this call is lower than the standard value, the status of this call is output as unqualified.

[0047] When the score input for this call is lower than the standard value, but the score output for this call is higher than the standard value, this call is marked and the status of this call is output as abnormal.

[0048] When the call status is abnormal, the input score and output score of the single interaction information are obtained respectively; If the input scores are all lower than the standard value, the customer is marked as abnormal; If there is a single interaction information with an input score higher than the standard value, obtain the first interaction information below the standard value and mark this interaction information.

[0049] The weights of speech content, speaking speed, clarity and sound decibels are pre-set, as well as the corresponding standard values ​​and corresponding scores.

[0050] Set the voice content keywords, including bonus keywords and deduction keywords and the corresponding bonus and deduction points.

[0051] Another embodiment of the present application discloses a storage medium, in which a computer program capable of being loaded by a processor and executing the following call identification method based on a TWS headset is stored: Get input and output voice information; Based on voice information, identify voice content, speaking speed, clarity and sound decibel; Based on the speech content, speaking speed, clarity and sound decibel, the scores of the input and output of this call are calculated by comparing them with the corresponding standard values; Based on the input and output scores and the score ratio, the status of this call is output accordingly.

[0052] Split the conversation content in the voice information to obtain multiple single interaction information; The voice content, speaking speed, clarity and sound decibel of a single interactive message are obtained respectively.

[0053] When the score output for this call is lower than the standard value, the status of this call is output as unqualified.

[0054] When the score input for this call is lower than the standard value, but the score output for this call is higher than the standard value, this call is marked and the status of this call is output as abnormal.

[0055] When the call status is abnormal, the input score and output score of the single interaction information are obtained respectively; If the input scores are all lower than the standard value, the customer is marked as abnormal; If there is a single interaction information with an input score higher than the standard value, obtain the first interaction information below the standard value and mark this interaction information.

[0056] The weights of speech content, speaking speed, clarity and sound decibels are pre-set, as well as the corresponding standard values ​​and corresponding scores.

[0057] Set the voice content keywords, including bonus keywords and deduction keywords and the corresponding bonus and deduction points.

[0058] The steps of the method or algorithm described in conjunction with the embodiments disclosed herein may be implemented directly using hardware, a software module executed by a processor, or a combination of the two. The software module may be placed in a random access memory (RAM), a memory, a read-only memory (ROM), an electrically programmable ROM, an electrically erasable programmable ROM, a register, a hard disk, a removable disk, a CD-ROM, or any other form of storage medium known in the art.

[0059] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "example", "specific example", or "some examples" etc. means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or more embodiments or examples in a suitable manner. In addition, those skilled in the art may combine and combine the different embodiments or examples described in this specification and the features of the different embodiments or examples, without contradiction.

[0060] In addition, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined as "first" and "second" may explicitly or implicitly include at least one of the features. In the description of the present invention, the meaning of "plurality" is at least two, such as two, three, etc., unless otherwise clearly and specifically defined.

[0061] Any process or method description in a flowchart or otherwise described herein may be understood to represent a module, segment or portion of code comprising one or more executable instructions for implementing the steps of a custom logical function or process, and the scope of the preferred embodiments of the present invention includes alternative implementations in which functions may not be performed in the order shown or discussed, including performing functions in a substantially simultaneous manner or in reverse order depending on the functions involved, which should be understood by technicians in the technical field to which the embodiments of the present invention belong.

[0062] This embodiment provides a computer-readable storage medium. Since the computer program stored in the medium can implement the various steps of the aforementioned embodiments after being run on a processor, the same technical effects as the aforementioned embodiments can be achieved. The principle analysis can refer to the aforementioned related step descriptions and will not be repeated here.

[0063] The embodiments of this specific implementation method are all preferred embodiments of the present application, and are not intended to limit the protection scope of the present application. Therefore, all equivalent changes made based on the structure and principle of the present application should be included in the protection scope of the present application.

Claims

1. A call recognition method based on TWS headset, characterized in that: include: Get input and output voice information; Based on voice information, identify voice content, speaking speed, clarity and sound decibel; Based on the speech content, speaking speed, clarity and sound decibel, the scores of the input and output of this call are calculated by comparing them with the corresponding standard values; Based on the input and output scores and the score ratio, the status of this call is output accordingly.

2. The call identification method based on TWS headset according to claim 1, characterized in that: Based on the voice information, the steps of identifying the voice content, speech speed, clarity and sound decibel include: Split the conversation content in the voice information to obtain multiple single interaction information; The voice content, speaking speed, clarity and sound decibel of a single interactive message are obtained respectively.

3. The call identification method based on TWS headset according to claim 2, characterized in that: Based on the input and output scores and the score ratio, the steps of outputting the call status include: When the score output for this call is lower than the standard value, the status of this call is output as unqualified.

4. The call identification method based on TWS headset according to claim 2, characterized in that: Based on the input and output scores and the score ratio, the steps of outputting the call status include: When the score input for this call is lower than the standard value, but the score output for this call is higher than the standard value, this call is marked and the status of this call is output as abnormal.

5. The call identification method based on TWS headset according to claim 4 is characterized in that: The method further comprises: When the call status is abnormal, the input score and output score of the single interaction information are obtained respectively; If the input scores are all lower than the standard value, the customer is marked as abnormal; If there is a single interaction information with an input score higher than the standard value, obtain the first interaction information below the standard value and mark this interaction information.

6. The call identification method based on TWS headset according to claim, characterized in that: The method further comprises: The weights of speech content, speaking speed, clarity and sound decibels are pre-set, as well as the corresponding standard values ​​and corresponding scores.

7. The call identification method based on TWS headset according to claim, characterized in that: The steps for speech content, speech speed, clarity and sound decibels include: Set the voice content keywords, including bonus keywords and deduction keywords and the corresponding bonus and deduction points.

8. A call identification system based on TWS headset, characterized in that: include: Voice acquisition module, voice acquisition input and output voice information; A speech recognition module is used to identify speech content, speech speed, clarity and sound decibel based on speech information; The scoring module is used to compare the speech content, speech speed, clarity and sound decibel with the corresponding standard values ​​to calculate the scores of the input and output of this call; The call status output module is used to output the call status based on the input and output scores and the score ratio.

9. An electronic device, characterized in that: It includes a memory and a processor, and the memory stores a computer program that can be loaded by the processor and executes the call identification method based on the TWS headset as described in any one of claims 1-7.

10. A storage medium, characterized in that: A computer program is stored which can be loaded by a processor and executes a call identification method based on a TWS headset as described in any one of claims 1 to 7.

Images