CN111263009A

CN111263009A - Quality inspection method, device, equipment and medium for telephone recording

Info

Publication number: CN111263009A
Application number: CN202010054848.8A
Authority: CN
Inventors: 王照; 许培东; 李燕; 吴珊羽; 李祥军; 毛俊丰; 易经纬; 张琦
Original assignee: Beijing Sankuai Online Technology Co Ltd
Current assignee: Beijing Sankuai Online Technology Co Ltd; Qiandai Beijing Information Technology Co Ltd
Priority date: 2020-01-17
Filing date: 2020-01-17
Publication date: 2020-06-09
Anticipated expiration: 2040-01-17
Also published as: CN111263009B

Abstract

The application discloses a quality inspection method, a quality inspection device, equipment and a medium for telephone recording, and belongs to the field of computers. The method comprises the following steps: acquiring a telephone record to be quality-checked; inputting the telephone recording into a recording quality inspection model, and identifying abnormal time periods and abnormal information corresponding to abnormal recording segments; displaying the abnormal time period and the abnormal information on a visual quality inspection interface; and responding to a playback instruction on the visual quality inspection interface, and playing back the abnormal recording segment. The quality testing personnel can quickly position the abnormal recording fragments, the quality testing personnel can conveniently perform quality detection on the telephone recording according to the time period of the abnormal recording fragments, and the detection efficiency of the quality testing personnel is improved.

Description

Quality inspection method, device, equipment and medium for telephone recording

Technical Field

The embodiment of the application relates to the field of computers, in particular to a quality inspection method, device, equipment and medium for telephone recording.

Background

The telephone service is a service mode for communicating with the client by telephone, the telephone service and the client carry out one-to-one conversation to solve the problem of the client, and the telephone service can also handle the service according to the information provided by the client, for example, the telephone service can inquire the residual telephone charge of the number according to the mobile phone number provided by the client.

The quality inspection of the telephone customer service refers to that quality inspection personnel detect the service quality of the telephone customer service when the telephone customer service is in telephone communication with a customer. In the related technology, when the telephone customer service and the client communicate with each other, the call is recorded, the quality testing personnel need to extract the call records for many times from a batch of call records, and the quality testing personnel need to play back and listen to a complete call record to test the service quality of the telephone customer service.

Based on the above situation, a large amount of call records need to be listened through a mode of manually detecting the service quality, and the efficiency is low.

Disclosure of Invention

The embodiment of the application provides a quality inspection method, a quality inspection device, a quality inspection equipment and a quality inspection medium for telephone recording, and provides a mode for displaying abnormal recording segments detected by a recording quality inspection model, so that quality inspection personnel can be quickly positioned to the abnormal recording segments through operation on a visual quality inspection interface, and the working efficiency of the quality inspection personnel is improved. The technical scheme is as follows:

according to one aspect of the present application, there is provided a quality inspection method of a telephone recording, the method comprising:

acquiring a telephone record to be quality-checked;

inputting the telephone recording into a recording quality inspection model, and identifying abnormal time periods and abnormal information corresponding to abnormal recording segments;

displaying the abnormal time period and the abnormal information on a visual quality inspection interface;

and responding to a playback instruction on the visual quality inspection interface, and playing back the abnormal recording segment.

According to another aspect of the present application, there is provided a quality inspection apparatus for a telephone recording, the apparatus comprising:

the acquisition module is used for acquiring the telephone record to be subjected to quality inspection;

the recording quality inspection model is used for carrying out quality inspection on the telephone recording and identifying abnormal time periods and abnormal information corresponding to abnormal recording fragments;

the display module is used for displaying the abnormal time period and the abnormal information on a visual quality inspection interface;

and the receiving module is used for responding to a playback instruction on the visual quality inspection interface and playing back the abnormal recording segment.

According to another aspect of the present application, there is provided a computer device comprising a processor and a memory, the memory having stored therein at least one instruction, at least one program, set of codes, or set of instructions, which is loaded and executed by the processor to implement the quality control method of telephone recordings as described above.

According to another aspect of the present application, there is provided a computer readable storage medium having stored therein at least one instruction, at least one program, set of codes, or set of instructions that is loaded and executed by a processor to implement a method of quality inspection of a telephone recording as described above.

The beneficial effects brought by the technical scheme provided by the embodiment of the application at least comprise:

the abnormal recording segment is identified to correspond to the abnormal time period and the abnormal information, the abnormal recording segment and the abnormal time period are displayed on the visual quality inspection interface, when the visual quality inspection interface receives a playback instruction, the corresponding abnormal recording segment can be played back, so that quality inspection personnel can quickly position the abnormal recording segment, the quality inspection personnel can conveniently perform quality detection on the telephone recording according to the time period in which the abnormal recording segment occurs, and the detection efficiency of the quality inspection personnel is improved.

Drawings

In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings needed to be used in the description of the embodiments are briefly introduced below, and it is obvious that the drawings in the following description are only some embodiments of the present application, and it is obvious for those skilled in the art to obtain other drawings based on these drawings without creative efforts.

FIG. 1 is a block diagram of an implementation environment provided by an exemplary embodiment of the present application;

FIG. 2 is a flow chart of a method for quality testing of a recorded telephone call provided by an exemplary embodiment of the present application;

FIG. 3 is a flow chart of a method for quality testing of a recorded telephone call provided by another exemplary embodiment of the present application;

FIG. 4 is a schematic view of a visual quality inspection interface provided by an exemplary embodiment of the present application;

FIG. 5 is a schematic view of a visual quality inspection interface provided by another exemplary embodiment of the present application;

FIG. 6 is a flowchart of a method for quality control of a recording of a telephone call in conjunction with a play interface, as provided by an exemplary embodiment of the present application;

FIG. 7 is a schematic diagram of a visual quality inspection interface with a play interface provided by an exemplary embodiment of the present application;

FIG. 8 is a schematic diagram of a visual quality inspection interface with a play interface provided by another exemplary embodiment of the present application;

FIG. 9 is a flowchart of a quality control method for call records based on a collection service according to an exemplary embodiment of the present application;

FIG. 10 is a block diagram of a quality control device for telephone recordings provided in an exemplary embodiment of the present application;

fig. 11 is a schematic structural diagram of a computer device according to an exemplary embodiment of the present application.

Detailed Description

To make the objects, technical solutions and advantages of the present application more clear, embodiments of the present application will be described in further detail below with reference to the accompanying drawings.

Fig. 1 shows a block diagram of a telephone customer service system according to an exemplary embodiment of the present application. The telephone customer service system 100 includes: a first human agent 120, a second human agent 160, a server 140, a quality inspection terminal 180, a first access line 130, a second access line 150, and a local area network 170.

The first manual seat 120 and the second manual seat 170 are connected through a local area network 170, the first manual seat 120 corresponds to a first telephone service, the second manual seat corresponds to a second telephone service, and the manual seats generally refer to one or more manual seats. A telephone network is also connected to the lan 170 for connecting the first and second

human seats

120, 160 to the first and

second access lines

130, 150. The first access line 130 corresponds to a first subscriber and the second access line corresponds to a second subscriber. Optionally, the first user is connected to the first human agent 120 via a first access line 130, or the first user is connected to the second human agent 160 via a first access line 130. Optionally, the second user is connected to the first human agent 120 via a second access line 150, or the second user is connected to the second human agent 160 via a second access line 150. The access line generally refers to one or more lines, and the embodiment of the present application is described by taking the first access line 130 and the second access line 150 as an example.

The first human agent 120 and the second human agent 130 are connected to the server 140 through a wireless network or a wired network, and the quality inspection terminal 180 is connected to the server 140 through a wireless network or a wired network.

The server 140 includes at least one of a server, a plurality of servers, a cloud computing platform, and a virtualization center. Illustratively, the server 140 includes a processor 144 and a memory 142, the memory 142 in turn including an acquisition module 1421, a display module 1422, and a processing module 1423. The server 140 is used to provide background services for applications or platforms that support detecting the quality of a recording of a telephone call. The obtaining module 1421 is configured to obtain a telephone recording, the displaying module 1422 is configured to send the visual quality inspection interface to the quality inspection terminal 180, and the processing module is configured to determine an abnormal recording segment and abnormal information. Optionally, the server 140 undertakes primary computing work, and the quality inspection terminal 180 undertakes secondary computing work; or, the server 140 undertakes the secondary calculation work, and the quality inspection terminal 180 undertakes the primary calculation work; alternatively, the server 140 and the quality inspection terminal 180 perform cooperative computing by using a distributed computing architecture.

Optionally, the telephone customer service system 100 includes at least one quality inspection terminal 180, and the quality inspection application installed or the quality inspection platform running on each quality inspection terminal 180 is the same, or the application installed on each terminal is the same type of application of different control system platforms. The quality control terminal 180 may be generally referred to as one of a plurality of terminals, and the quality control terminal 180 is illustrated in this embodiment. The types of the devices of the terminals are the same or different, and the types of the devices include: at least one of a smartphone, a tablet, an e-book reader, an MP3 player, an MP4 player, a laptop portable computer, and a desktop computer, a notebook computer.

Those skilled in the art will appreciate that the number of terminals described above may be greater or fewer. For example, the number of the terminals may be only one, or several tens or hundreds of the terminals, or more. The number of terminals and the type of the device are not limited in the embodiments of the present application.

The application provides a quality inspection method for telephone recording, which can be applied to a scene of quality inspection of call recording service. Fig. 2 is a flowchart illustrating a quality inspection method for a telephone recording according to an exemplary embodiment of the present application, which can be applied to the quality inspection terminal 180 in the telephone customer service system shown in fig. 1. The method comprises the following steps:

step 201, obtaining the telephone recording to be quality checked.

The telephone recording is the recording of the first speaking object and the second speaking object in the call process, and schematically, the first speaking object and the second speaking object are the telephone customer service and the user respectively. Optionally, the telephone record to be quality-checked is obtained through a telephone customer service system, which is also named as a call center, the telephone customer service system accesses the telephone channel of the user to the telephone customer service place or to a manual seat, the telephone customer service system records the call between the telephone customer service and the user, and records the telephone number used by the accessed user, the problem to be solved by the user, the solution of the problem, the evaluation of the service quality of the telephone customer service by the user, and the like.

Alternatively, the telephone customer service is a telephone customer service of an online e-commerce platform, or a telephone customer service of an offline enterprise. Optionally, the e-commerce platform comprises at least one of a shopping platform, a social platform, a video platform, a music platform, an education platform, an information platform, a transportation platform, and a travel platform. Optionally, the offline enterprise includes at least one of a logistics enterprise, a wholesale retail enterprise, a lodging catering enterprise, a leasing enterprise, a communication enterprise, a bank, and a real estate enterprise.

Step 202, inputting the telephone recording into the recording quality inspection model, and identifying abnormal time periods and abnormal information corresponding to the abnormal recording segments.

Optionally, the recording quality control model is a machine learning model having a detection capability for voice, the recording quality control model is one or more models, and when the recording quality control model is a plurality of models, the recording quality control model is the same model, or different models of the same type, or different models.

Optionally, the abnormal time period and the abnormal information are one of machine quality inspection results obtained by the sound recording quality inspection model, and the abnormal time period is a time period corresponding to the occurrence of sensitive words (such as forbidden words, non-civilized words, and the like) and sensitive words or a time period corresponding to the occurrence of abnormal emotions.

Optionally, the machine quality inspection result of the telephone recording further includes at least one of the number of the speaking objects in the abnormal recording segment, abnormal speech content corresponding to the speaking objects in the abnormal recording segment (for example, the abnormal recording segment is generated by the telephone customer service, and the machine quality inspection result includes the abnormal speech content of the telephone customer service), emotion score, volume of the speaking objects, intonation of the speaking objects, tone of the speaking objects, and grade of service quality (for example, good service quality). Optionally, the voice recording quality inspection model extracts an emotional characteristic parameter in the telephone voice recording, the emotional characteristic parameter is a k-dimensional vector, the emotional characteristic parameter is a parameter corresponding to a characteristic for judging emotion in the voice, and the emotional characteristic parameter includes at least one of a tone, a word and a volume. Optionally, the recording quality inspection model identifies the emotional characteristic parameters by combining semantics to obtain emotional information in the telephone recording. The emotion information comprises normal emotion and abnormal emotion, and the abnormal emotion refers to negative emotion expressed by a speaking object by changing aspects of tone, intonation, volume and the like during speaking.

Illustratively, when the emotion score of the telephone recording is greater than 5 minutes (the abnormal threshold value is 10 minutes), the recording quality inspection model judges that the service quality of the telephone recording is unqualified, the telephone recording is input into the recording quality inspection model, and the emotion score of the telephone recording is 2 minutes, then the recording quality inspection model judges that the service quality of the telephone recording is qualified.

In one example, the telephone recording is a recording of a call between the user and a telephone service of a cell phone operator, and the recording content is a remaining telephone charge of a cell phone number used by the user through the telephone service inquiry. The section of the recording is input into a recording quality inspection model, the recording quality inspection model divides the section of the recording into a recording segment corresponding to the telephone customer service and a recording segment corresponding to the user, for example, the recording segment corresponding to the telephone customer service contains voice corresponding to civilized terms such as "you are good and are very happy to serve you", and the recording segment corresponding to the user contains voice corresponding to query language such as "i want to find the remaining telephone charge of the mobile phone number". The recording quality inspection model identifies the sensitive words and emotion in the recording, the sensitive words and emotion are not found, and the machine quality inspection result of the recording is qualified.

Optionally, the recording quality inspection model can store the machine quality inspection result into a database, and the database stores scoring standards of sensitive words and historical abnormal emotions.

Optionally, the record quality inspection model is trained by:

s1, obtaining a sample telephone recording and a quality inspection result of the sample telephone recording, wherein the sample telephone recording is composed of at least one group of telephone recordings between customer service and users, and the quality inspection result of the sample telephone recording comprises at least one of emotion scores and sensitive words.

The sample telephone recordings are obtained from telephone customer service systems of a plurality of platforms, and the quality inspection results of the sample telephone recordings are obtained by manual detection, or other recording quality inspection models, or emotion recognition models, or text-to-speech models.

S2, extracting emotion feature parameters from the sample telephone recording, wherein the emotion feature parameters comprise emotion feature vectors of k dimensions, and k is an integer larger than zero.

Illustratively, the emotional characteristic parameter includes at least one of voice, tone and volume.

And S3, inputting the emotional characteristic parameters into the initial recording quality inspection model to obtain the predicted quality inspection result of the sample telephone recording.

The initial sound recording quality control model is an untrained sound recording quality control model. Illustratively, the sample telephone recording is input into an initial recording quality control model, and the result of the predicted quality control of the sample telephone recording is that sensitive words appear in the sample telephone at 3 minutes, and the emotion score of the sample telephone at 4-4 minutes 05 seconds is an abnormal emotion score.

And S4, calculating the error between the predicted quality inspection result and the quality inspection result.

The error between the predicted quality test result and the quality test result can be calculated by any error function, such as a norm loss function.

And S5, training the initial recording quality inspection model by adopting an error back propagation algorithm to obtain the trained recording quality inspection model.

And the recording quality inspection model is subjected to repeated iterative training through an error back propagation algorithm, so that the detection result of the recording quality inspection model on the telephone recording is more accurate.

And step 203, displaying the abnormal time period and the abnormal information on a visual quality inspection interface.

The visual quality inspection interface is displayed according to the machine quality inspection result of the telephone recording, and comprises: and the abnormal recording section comprises at least one of a first abnormal recording section of a first speaking object and a second abnormal recording section of a second speaking object. Illustratively, the first speaking party is a telephone customer service and the second speaking party is a user.

Optionally, the visual quality inspection interface is an automatically generated interface, or the visual quality inspection interface is an interface generated by triggering a corresponding control (e.g., a quality inspection control, a review control). Optionally, the visual quality inspection interface is an interface independent from the quality inspection platform or an interface in the quality inspection platform, and the visual quality inspection interface is one or more interfaces.

Optionally, the visual quality control interface is an interface displaying the machine quality control result of the telephone recording and a time axis corresponding to the telephone recording. Illustratively, the duration of the telephone recording is 5 minutes, the sensitive vocabulary appears in 3 minutes to 3 minutes and 05 seconds of the telephone recording, and the visual quality control interface displays the recording segment of the telephone recording in 3 minutes to 3 minutes and 05 seconds, the time corresponding to the sensitive vocabulary, and the speaking object generating the sensitive vocabulary (if the speaking object generating the sensitive vocabulary is a user), the abnormal recording segment is the second abnormal recording segment of the user. Optionally, the abnormal information includes abnormal emotions, sensitive vocabulary, and information of a speaking subject. Optionally, the unit of the abnormal time period corresponding to the abnormal recording segment displayed on the visual quality control interface is hour, minute, second, millisecond, microsecond or a combination thereof. Optionally, the first abnormal recording segment and the second abnormal recording segment are within the same time period and within the same time interval.

And step 204, responding to a playback instruction on the visual quality inspection interface, and playing back the abnormal recording segment.

Optionally, the visual quality control interface is an interface for interacting with a user. Optionally, the visual quality inspection interface is displayed on a display screen of the terminal, and the terminal includes at least one of a smart phone, a tablet computer, a desktop computer, and a notebook computer. When the terminal is a terminal with a touch display screen, such as a smart phone or a tablet personal computer, the playback instruction comprises an instruction generated by at least one of single-click operation, double-click operation, long-press operation, sliding operation and dragging operation; when the terminal is a terminal including an external device, such as a desktop computer or a notebook computer, the playback instruction is an instruction generated by the external device, such as an instruction generated by a key on a keyboard, or an instruction generated by a single-click operation, a double-click operation, a drag operation, or a right-click operation of a mouse.

In summary, in the method provided in this embodiment, the call record between the first speaking object and the second speaking object is input into the record quality inspection model, the abnormal time period and the abnormal information corresponding to the abnormal record segment are identified, the abnormal record segment and the abnormal time period are displayed on the visual quality inspection interface, and when the visual quality inspection interface receives the playback instruction, the corresponding abnormal record segment can be played back, so that the quality inspection personnel can quickly locate the abnormal record segment, the quality inspection personnel can conveniently perform quality inspection on the telephone record according to the time period in which the abnormal record segment occurs, and the detection efficiency of the quality inspection personnel is improved.

Fig. 3 is a flowchart illustrating a quality testing method for a telephone recording according to another exemplary embodiment of the present application. The method can be applied to the quality inspection terminal 180 in the telephone customer service system as shown in fig. 1. The method comprises the following steps:

step 301, obtaining the telephone record to be quality checked.

Step 301 is identical to step 201 shown in fig. 2 and will not be described herein.

Step 302a, inputting the telephone recording into a recording quality inspection model, and identifying a first time period corresponding to the recording segment containing the abnormal emotion score and a first speaking object corresponding to the abnormal emotion score.

Optionally, the abnormal information includes abnormal emotion scores, the abnormal emotion scores are obtained after the recording quality inspection model detects the recording segments containing the abnormal emotion, and the abnormal emotion scores correspond to the time or the time period when the abnormal emotion occurs. Illustratively, the higher the score of the abnormal emotion score is, the higher the possibility that the abnormal emotion exists in the recording segment corresponding to the time period is, the recording segment is an abnormal recording segment, and a quality inspector needs to focus on detecting the recording segment corresponding to the time period with the higher abnormal emotion score. Optionally, the one-time telephone recording includes one or more abnormal recording segments, and the one-time telephone recording refers to a recording recorded by a telephone customer service and a user in a call process. Optionally, there are one or more abnormal emotion scores in one abnormal recording segment, and when the abnormal recording segments are included in one telephone recording, the abnormal emotion score of each abnormal recording segment in the abnormal recording segments is the same or different. Illustratively, a recording segment existing in one-time telephone recording is an abnormal recording segment, a time segment corresponding to the abnormal recording segment is a first time segment, the first time segment can be any time segment in one-time telephone recording, and the abnormal recording segment is generated by a first speaking object or a second speaking object. Illustratively, the first speaking party is a telephone customer service and the second speaking party is a user.

And the recording quality inspection model determines a time period corresponding to the recording segment containing the abnormal emotion score as a first time period. Illustratively, the duration of the first time period is 3 seconds, or the duration of the first time period is 500 milliseconds.

Step 302b, inputting the telephone recording into the recording quality inspection model, and identifying a second time period corresponding to the recording fragment containing the sensitive vocabulary and a second speaking object corresponding to the sensitive vocabulary.

Optionally, the exception information includes sensitive words. Sensitive words are words with political tendencies, violence tendencies, unhealthy colors. Optionally, the sensitive vocabulary includes at least one of forbidden vocabulary and non-civilized vocabulary, and the sensitive vocabulary corresponds to the time or time period of the sensitive vocabulary. Illustratively, a mark is arranged in a time period corresponding to a time period containing sensitive words in the sound recording segment, the mark is used for indicating that the time period contains the sensitive words, the time period is an abnormal time period, and the sound recording segment corresponding to the time period is an abnormal sound recording segment. Quality testing personnel need to focus on detecting the recording segments corresponding to the abnormal time periods with the marks. Optionally, the one-time telephone recording includes one or more abnormal recording segments with sensitive words, one abnormal recording segment includes one or more sensitive words, and when the one-time telephone recording includes multiple abnormal recording segments, the sensitive words corresponding to each abnormal recording segment in the multiple abnormal recording segments are the same or different. Illustratively, a recording segment existing in one-time telephone recording is an abnormal recording segment, the time period corresponding to the recording segment is a second time period, the second time period can be any time period in one-time telephone recording, and the abnormal recording segment is generated by the first speaking object or the second speaking object. Illustratively, the first speaking object is a user and the second speaking object is a telephone customer service.

And the recording quality inspection model determines a time period corresponding to the recording fragment with the sensitive vocabulary as a second time period. Illustratively, the duration of the second time period is 2 seconds, or the duration of the second time period is 300 milliseconds.

And step 303, displaying the abnormal time period and the abnormal information on a visual quality inspection interface.

Illustratively, the visual quality control interface comprises a rectangular plane coordinate system which takes the first time period corresponding to the abnormal recording segment as an abscissa and takes the abnormal emotion score corresponding to the abnormal recording segment as an ordinate, and the abnormal recording segment is generated by the first speaking object. Optionally, the visual quality control interface includes a graph of the mood score of the first speaking subject in a recording of the call and a first time period.

Illustratively, the visual quality inspection interface comprises a plane rectangular coordinate system which takes the second time period corresponding to the abnormal recording segment as an abscissa and marks the sensitive vocabulary corresponding to the abnormal recording segment as an ordinate, wherein the abnormal recording segment is generated by the second speaking object. Optionally, the visual quality control interface includes a graph of the sensitive vocabulary indicia of the second speaker during the one-time phone recording and the second time period.

And 304, responding to a playback instruction on the visual quality inspection interface, selecting a target abnormal recording segment from the at least one abnormal recording segment, and playing back the target abnormal recording segment.

Optionally, at least one abnormal recording segment is included on the visual quality control interface.

Optionally, the visual quality inspection interface is displayed on a display screen of the terminal, and the terminal includes at least one of a smart phone, a tablet computer, a desktop computer, and a notebook computer. When the terminal is a terminal with a touch display screen, such as a smart phone or a tablet personal computer, the playback instruction comprises an instruction generated by at least one of single-click operation, double-click operation, long-press operation, sliding operation and dragging operation; when the terminal is a terminal including an external device, such as a desktop computer or a notebook computer, the playback instruction is an instruction generated by the external device, such as an instruction generated by a key on a keyboard, or an instruction generated by a single-click operation, a double-click operation, a drag operation, or a right-click operation of a mouse. The playback instruction is used for selecting a target abnormal recording segment on the visual quality inspection interface, wherein the target abnormal recording segment can be a recording segment with higher abnormal emotion score, or a recording segment containing sensitive vocabulary marks, or a recording segment which a quality inspection person wants to listen to.

Optionally, the quality testing personnel click the target abnormal recording segment on the visual quality testing interface and directly play the target abnormal recording segment on the visual quality testing interface, or the quality testing personnel click the target abnormal recording segment on the visual quality testing interface, switch the playing interface for displaying the target abnormal recording segment and directly play the target abnormal recording segment on the playing interface, or the quality testing personnel click on the playing interface and then play the target abnormal recording segment.

It is understood that step 302a and step 302b may be implemented separately or in combination.

In summary, in the method provided in this embodiment, the recording quality inspection model is used to detect the recording segment to be quality inspected, to obtain the abnormal information (including the abnormal emotion score and the sensitive vocabulary) of the recording segment, the abnormal time period corresponding to the abnormal information, and the speaking object generating the abnormal information, and the visual quality inspection interface is displayed according to the abnormal information of the abnormal recording segment, the abnormal time period, and the speaking object generating the abnormal information. The quality testing personnel can select the target abnormal time period on the quality testing interface according to the abnormal information and play the target abnormal recording segment corresponding to the target abnormal time period, and the quality of service of the telephone recording can be judged whether to meet the standard or not without listening to the whole telephone recording by the quality testing personnel, so that the quality testing efficiency of the quality testing personnel is improved.

Fig. 6 is a flowchart illustrating a quality control method for a telephone recording in conjunction with a play interface according to an exemplary embodiment of the present application. The method can be applied to the quality inspection terminal 180 in the telephone customer service system as shown in fig. 1. The method comprises the following steps:

step 601, obtaining the telephone record to be quality checked.

Step 601 is identical to step 201 shown in fig. 2, and is not described herein again.

Step 602a, inputting the telephone recording into the emotion recognition model, and recognizing the emotion score and the speaking object of the telephone recording in different time periods.

Optionally, the quality control model of the recording comprises one or more emotion recognition models, the emotion recognition models being machine learning models having emotion recognition capabilities for the recording. When the recording quality inspection model includes a plurality of emotion recognition models, the plurality of emotion recognition models are the same model, or different models in the same type, or different models. And inputting the telephone recording into the emotion recognition model, distinguishing the telephone customer service in the telephone recording from the recording of the user by the emotion recognition model, and respectively carrying out emotion scoring on the telephone customer service and the recording fragment of the user. Optionally, the one-time recorded call has no abnormal emotion or includes one or more abnormal emotions. Optionally, the emotion recognition model evaluates the telephone recording by at least one of a score, a rating, and a positive or negative tendency, if the emotion rating of the telephone recording is S level, the emotion corresponding to the recording segment tends to be abnormal emotion, or if the emotion rating of the telephone recording is D level, the emotion corresponding to the recording segment tends to be normal emotion. As another example, the emotion recognition model identifies whether the emotion corresponding to the telephone recording is a positive emotion, or a negative emotion, or a neutral emotion (without any mood swings).

Step 603a, determining a time period when the first emotion score exceeds the abnormal threshold value as a first time period, and determining a speaking object corresponding to the first emotion score as a first speaking object; and/or determining a time period when the second emotion score exceeds the abnormal threshold value as a first time period, and determining a speaking object corresponding to the second emotion score as a first speaking object; and the speaking object corresponding to the first emotion score and the speaking object corresponding to the second emotion score are different speaking objects.

Illustratively, the abnormal threshold is 5, and the recording segment with the emotion score greater than 5 points is identified by the emotion recognition model as containing an abnormal emotion score, and the time period corresponding to the abnormal emotion score is the first time period. Optionally, the speaking subject generating the abnormal emotion score is a first speaking subject, and the speaking subject generating the abnormal emotion score includes at least one of a telephone customer service and a user. Illustratively, the first speaking object is a telephone customer service, or the first speaking object is a user, or the first speaking object is a telephone customer service and a user.

The emotion recognition model can identify a first speaking object generating abnormal emotion scores, and if the abnormal emotion scores are generated by the telephone customer service, the telephone customer service is the first speaking object; the abnormal emotion score is generated by the user, and the user is the first speaking subject.

Step 602b, inputting the telephone recording into the speech-to-text model, and identifying the speech text and the speaking object corresponding to the telephone recording in different time periods.

Optionally, the sound recording quality control model comprises one or more speech-to-text models, the speech-to-text models being machine learning models having the ability to convert sound recordings to text files. When the recording quality inspection model comprises a plurality of voice-to-text models, the plurality of voice-to-text models are the same model, or different models in the same type, or different models. And inputting the telephone record into a voice-to-text model, wherein the voice-to-text model distinguishes the telephone customer service in the telephone record from the record of the user, and identifies sensitive words and phrases of the telephone customer service and the record of the user. Optionally, a single recording of a call is free of sensitive words or includes one or more sensitive words. Optionally, the speech-to-text model marks a time period containing the sensitive words, or screens the time period containing the sensitive words, and the screened recording segments are separately displayed in the visual quality inspection interface or are marked according to the classification of the sensitive words. For example, the sensitive vocabulary can be classified into non-civilized types (dirty words, etc.), forbidden types (askew evil, etc.), illegal types (yellow gambling, etc.), a voice-to-text model uses a circular mark for marking the non-civilized sensitive vocabulary, a triangular mark for marking the forbidden sensitive vocabulary, and a rectangular mark for marking the illegal sensitive vocabulary. The shape of the sensitive vocabulary indicia is not limited in this application.

Step 603b, determining a time period of the first voice text with the sensitive words as a second time period, and determining a speaking object corresponding to the first voice text as a second speaking object; and/or determining a time period of the second voice text with the sensitive words as a second time period, and setting a speaking object corresponding to the second voice text of the user as a second speaking object.

Optionally, the spoken object that generates the sensitive vocabulary is a second spoken object, the spoken object that generates the sensitive vocabulary including at least one of a telephone customer service and a user. Illustratively, the second speaking object is a telephone customer service, or the second speaking object is a user, or the second speaking object is a telephone customer service and a user.

The voice-to-text model can identify a second speaking object corresponding to the recording fragment containing the sensitive vocabulary, and if the sensitive vocabulary is generated by the telephone customer service, the telephone customer service is determined as the second speaking object; the sensitive vocabulary is generated by the user, and the user is determined as a second speaking object.

Step 604, responding to a playback instruction on the visual quality inspection interface, and switching and displaying a playing interface of the target abnormal recording segment selected by the playback instruction, wherein the playing interface of the target abnormal recording segment comprises a time interval of a target abnormal time period corresponding to the target abnormal recording segment.

Optionally, a plurality of abnormal recording segments are included on the visual quality control interface. The playback instruction is used for switching and displaying a playing interface of the target abnormal recording segment. Optionally, the playing interface of the target abnormal recording clip includes a playing time axis, and the playing time axis includes a time interval in which a target abnormal time period corresponding to the target abnormal recording clip is located.

Optionally, the playing interface of the target abnormal recording segment is an interface which is switched and displayed after the visual quality inspection interface skips, the playing interface and the visual quality inspection interface are the same or different in size, a return control can be set on the playing interface, and the visual quality inspection interface can be switched and displayed after the return control is triggered.

Optionally, the playing interface of the target abnormal recording segment is a newly added display interface independent of the visual quality inspection interface, and the playing interface and the visual quality inspection interface are the same or different in size.

In one example, a time axis corresponding to the abnormal recording segment and abnormal information corresponding to the abnormal recording segment are displayed on the visual quality control interface. The quality testing personnel select the time interval of the target abnormal time period through the abnormal information, and the quality testing personnel determine the target abnormal time period by clicking the time axis. For example, the emotion score corresponding to the time period from 3 minutes to 30 seconds to 4 minutes of the telephone recording is higher, and the quality inspector clicks the time interval of the time period on the time axis.

In one example, when the emotion score is greater than 0.5, the recording quality inspection model determines that the emotion score is an abnormal emotion score, the time period corresponding to the abnormal emotion score is an abnormal time period, and the recording segment corresponding to the abnormal time period is an abnormal recording segment. As shown in fig. 4, a user mark 101, a telephone service mark 102, a sensitive vocabulary mark 103, an emotion mark 104, a time axis 105 (the unit of the time axis is seconds), and a speaking object which generates an abnormal recording clip are displayed on the visual quality control interface 10. The sensitive vocabulary label 103 appears in the time interval corresponding to 60 seconds to 100 seconds, the abnormal emotion score appears in the time interval corresponding to 220 seconds to 300 seconds, and the abnormal emotion score is 1. The quality testing personnel can focus on detecting the abnormal recording segments corresponding to 220 seconds to 300 seconds or 60 seconds to 100 seconds on the visual quality testing interface 10. Alternatively, the quality inspector clicks 260 seconds on the time axis 105 by a double-click operation, and may also click 300 seconds on the time axis 105. Optionally, the visual quality control interface 10 automatically plays an abnormal recording clip for a certain period of time, for example, the visual quality control interface 10 plays an abnormal recording clip for 20 seconds, and when the quality control person clicks 260 seconds on the timeline 105, the visual quality control interface 10 plays a recording clip between 260 seconds and 280 seconds, or plays a recording clip between 240 seconds and 260 seconds, or plays a recording clip between 250 seconds and 270 seconds. Optionally, an input control 106 is displayed on the visual quality control interface 10, and the quality control person can input the selected time interval numbers, such as 60 and 100, in the input control 106.

In another example, when the emotion score is greater than 4 minutes, the recording quality test model determines that the emotion score is an abnormal emotion score, the time period corresponding to the abnormal emotion score is an abnormal time period, and the recording segment corresponding to the abnormal time period is an abnormal recording segment. As shown in fig. 5, the visual quality inspection interface 11 displays a user mark 111, a telephone customer service mark 112, a sensitive vocabulary mark 113, an emotion mark 114, a time axis 115 (the unit of the time axis is seconds), and a speaking object generating an abnormal recording clip. The sensitive vocabulary flag 113 appears at 100 seconds and 220 seconds, and the abnormal emotion score is 5 points at 220 seconds. The quality testing personnel can focus on detecting the abnormal recording segment corresponding to 220 seconds on the visual quality testing interface 11, and can also detect the abnormal recording segment corresponding to 100 seconds. Optionally, a play control 116, a fast forward control 117, a fast rewind control 118, and a pause control 119 are displayed on the visual quality inspection interface 11. The quality testing personnel click the time period on the time axis 115 and click the control to play the abnormal recording clip of the corresponding time period, if the quality testing personnel click 220 seconds on the time axis 115 to determine that the playing time period is 210 seconds to 230 seconds, then click the playing control 116 to play the abnormal recording clip corresponding to 210 seconds to 230 seconds, the quality testing personnel can also click the fast forward control 117 to play the abnormal recording clip between 220 seconds and 240 seconds, and similarly, the quality testing personnel click the fast backward control 118 to realize the return play of the abnormal recording clip between 200 seconds and 220 seconds, and in the playing process, the quality testing personnel can pause the playing of the abnormal recording clip by clicking the pause control 119.

As shown in fig. 7 (a), a user mark 131, a telephone customer service mark 132, a sensitive vocabulary mark 133, an emotion score 134, and a time axis 135 (the unit of the time axis is seconds) and a speaking subject who generates an abnormal recording clip are displayed on the visual quality control interface 13. At 220 seconds on the time axis, higher emotion scores and sensitive vocabulary marks 133 are displayed, the recording clip corresponding to the time period is a target abnormal recording clip, and a quality inspector can play the target abnormal recording clip by clicking the time axis at 220 seconds. Alternatively, the quality inspector clicks the corresponding time axis at 220 seconds to display the playback interface 14 as shown in fig. 7 (b). Optionally, a play timeline at 220 seconds, a play control 136, a fast forward control 137, a fast reverse control 138, a pause control 139, and an input control 140 are displayed on the play interface 14. Illustratively, the play timeline includes a time period between 210 seconds and 230 seconds, or the play timeline includes a time period between 220 seconds and 240 seconds of quality control, or the play timeline includes a time period between 200 seconds and 220 seconds. The interval displayed on the playing time axis is a preset interval, or the preset interval can be defined by users. The playing function of the recording clip is realized through the control.

Optionally, the time interval in which the target abnormal time period is located corresponds to n interval ranges.

Responding to an interval dragging instruction on a playing interface of the target abnormal recording segment, and determining a time interval corresponding to an ith interval range from n ranges corresponding to the time interval of the target abnormal time period, wherein the time interval corresponding to the ith range comprises at least one target abnormal moment corresponding to the target abnormal recording segment, and n is greater than 0, and i is greater than 0.

And step 605, playing back the target abnormal recording segment according to the time interval.

Optionally, the playing interface is displayed on a display screen of the terminal, and the terminal includes at least one of a smart phone, a tablet computer, a desktop computer, and a notebook computer. When the terminal is a terminal with a touch display screen, such as a smart phone or a tablet personal computer, the interval dragging instruction comprises an instruction generated by at least one of single-click operation, double-click operation, long-press operation, sliding operation and dragging operation; when the terminal is a terminal including an external device, such as a desktop computer or a notebook computer, the interval dragging instruction is an instruction generated by the external device, such as an instruction generated by a key on a keyboard, or an instruction generated by a single-click operation, a double-click operation, a dragging operation, or a right-click operation of a mouse.

Illustratively, the quality testing staff changes the time interval of the target abnormal time period by dragging the playing time axis on the playing interface.

In one example, the quality control person drags the play time axis display on the play interface 14 as shown in fig. 8, and the play interface is shown in fig. 8 (a), and the time interval in which the target abnormal time period displayed on the play interface 15 is located is [207, 227], which is the time interval corresponding to the 1 st interval range among the n interval ranges. As shown in fig. 8 (b), the time interval in which the target abnormal time period displayed on the playback interface 16 is present is [217, 237], which is the time interval corresponding to the 2 nd interval range among the n interval ranges, and the unit of the playback time axis is seconds. Alternatively, the length of the time interval may be any length, such as the time interval is [200, 240 ].

Illustratively, the quality control staff clicks the playing time axis to play the target abnormal recording clip corresponding to the time interval, or clicks the playing control 136 to play the target abnormal recording clip corresponding to the time interval.

In summary, the method provided in this embodiment detects the emotion score and the sensitive vocabulary in the telephone recording through the emotion recognition model and the speech-to-text model in the recording quality inspection model, and distinguishes the speaking objects generating abnormal emotion scores and sensitive vocabularies, thereby identifying the party generating the problem. The quality testing personnel can change the time interval being played by dragging the playing time axis corresponding to the abnormal recording clip, for example, the time interval corresponding to the 2 nd minute 30 seconds of the telephone recording is changed into the time interval corresponding to the 2 nd minute 20 seconds to the 2 nd minute 50 seconds of the telephone recording, so that the quality testing personnel can judge whether the service quality of the telephone recording meets the standard without listening to the complete telephone recording, and the quality testing efficiency of the quality testing personnel is improved.

The quality inspection method of the telephone recording will be described with reference to the collection case. Fig. 9 is a flowchart illustrating a quality control method for call records of the collection service according to an exemplary embodiment of the present application. The method can be applied to the quality inspection terminal 180 in the telephone customer service system as shown in fig. 1. The method comprises the following steps:

step 901, the voice of the person who asks for collection is prompted.

Taking network loan as an example, a borrower borrows money on a loan platform, the borrower needs to pay principal and interest back to the loan platform within a specified period, and a worker of the loan platform urges the borrower to pay as soon as possible by calling, so that the business of urging to accept is generated. In order to prevent the loan platform from threatening the borrower to pay in an improper way, relevant supervision departments detect the telephone recordings of the acquirer and the borrower, and if the malicious acquirer hasten phenomenon, the loan platform is processed.

Step 902, obtain a call record of the call record.

And acquiring the telephone recording of the acquirer through a telephone customer service system of the loan lending platform, wherein optionally, the telephone recording comprises the employee number of the acquirer who dials the acquirer who calls the acquirer and the telephone number of the borrower.

Step 903a, adopting a recording quality inspection model to perform voice recognition on the call recording of the call record.

Step 904a, sensitive words in the telephone recording are obtained.

Optionally, the voice recording quality control model includes a voice-to-text model that identifies sensitive words present in the voice recording.

And 903b, performing emotion recognition on the call record of the call record by adopting a record quality inspection model.

Step 904b, the emotional questions in the recording of the call are obtained.

Optionally, the quality control model of the recording includes an emotion recognition model that identifies emotional problems that exist in the recording of the telephone. Illustratively, the emotion recognition model scores the recordings, with higher scores of recordings having a greater likelihood of having an abnormal emotion.

Step 905a, visually displaying a problem time axis corresponding to the telephone recording.

And step 905b, performing interval recording playback on the problem points corresponding to the telephone recording.

Alternatively, steps 903a to 904a and steps 903b to 904b may be implemented individually or in combination.

Illustratively, there are unusual emotions and sensitive words in the telephone recordings of urge to collect cases. The voice-to-text model in the recording quality inspection model identifies sensitive words, sensitive word marks are displayed on a visual quality inspection interface, the emotion identification model in the recording quality inspection model identifies abnormal emotion, and abnormal emotion scores are displayed on the visual quality inspection interface. The sensitive vocabulary mark is generated by the borrower, or generated by the acquirer, or generated by both the borrower and the acquirer; the abnormal emotion is generated by the borrower, the acquirer, or both the borrower and the acquirer. The recording section containing abnormal emotion is an abnormal recording section, or the recording section containing sensitive words is an abnormal recording section, or the recording section containing abnormal emotion and sensitive words is an abnormal recording section, the abnormal recording section corresponds to an abnormal time period, and the time interval of the abnormal time period corresponds to n ranges.

And step 906, manually rechecking.

Quality testing personnel recheck the telephone recording, click the time axis corresponding to the abnormal recording clip on a visual quality testing interface, display the playing interface of the abnormal recording clip, listen to the recording clip corresponding to the time interval on the playing interface, and also drag the time axis to change the current playing time interval. When quality testing personnel can not determine whether the telephone recording has the service quality problem after listening to the recording segments of a plurality of time intervals, the quality testing personnel can play the complete telephone recording to determine.

In summary, the method provided in this embodiment applies the quality inspection method for the telephone recording to the collection prompting case, so that the quality inspection efficiency of quality inspection personnel is improved, the quality of service problem of the telephone recording can be found in time, the complaint of the user is avoided, and the stability of the collection prompting service is ensured.

The following are embodiments of the apparatus of the present application, and for details that are not described in detail in the embodiments of the apparatus, reference may be made to corresponding descriptions in the above method embodiments, and details are not described herein again.

Fig. 10 is a schematic structural diagram illustrating a quality inspection apparatus for telephone recordings according to an exemplary embodiment of the present application. The device can be realized by software, hardware or the combination of the two to become all or a part of the terminal, and the device is applied to the quality control service of the telephone recording, and the device comprises:

an obtaining module 1010, configured to obtain a telephone recording to be quality-checked;

a recording quality inspection model 1020 for performing quality inspection on the telephone recording and identifying abnormal time periods and abnormal information corresponding to the abnormal recording segments;

the display module 1030 is configured to display the abnormal time period and the abnormal information on a visual quality inspection interface;

the receiving module 1040 is configured to respond to a playback instruction on the visual quality inspection interface, and play back the abnormal recording segment.

In an alternative embodiment, the anomalous information includes an anomalous sentiment score; the apparatus includes a processing module 1050;

the processing module 1050 is configured to input the telephone recording to the recording quality inspection model 1020, and identify a first time period corresponding to the recording segment with the abnormal emotion score and a first speaking object corresponding to the abnormal emotion score.

In some embodiments of the present application, the sound recording quality control model 1020 includes an emotion recognition model 10201;

the emotion recognition model 10201 is used for recognizing the telephone recording to obtain emotion scores and speaking objects of the telephone recording in different time periods;

the processing module 1050 is configured to determine a time period in which the first emotion score exceeds an abnormal threshold as a first time period, and determine a speaking object corresponding to the first emotion score as a first speaking object; and/or determining a time period when the second emotion score exceeds the abnormal threshold value as a first time period, and determining a speaking object corresponding to the second emotion score as a first speaking object; and the speaking object corresponding to the first emotion score and the speaking object corresponding to the second emotion score are different speaking objects.

In some embodiments of the present application, the exception information includes sensitive vocabulary;

the processing module 1050 is configured to input the telephone recording to the recording quality inspection model 1020, and identify a second time period corresponding to the recording segment containing the sensitive vocabulary and a second speaking object corresponding to the sensitive vocabulary.

In some embodiments of the present application, the recorded matter quality control model 1020 includes a speech-to-text model 10202;

the speech-to-text model 10202 is used for identifying a telephone recording and identifying a speech text and a speaking object corresponding to the telephone recording in different time periods;

the processing module 1050 is configured to determine a time period in which the first speech text has the sensitive vocabulary as a second time period, and determine a speaking object corresponding to the first speech text as a second speaking object; and/or determining a time period of the second voice text with the sensitive words as a second time period, and determining a speaking object corresponding to the second voice text as a second speaking object; and the speaking object corresponding to the first voice text and the speaking object corresponding to the second voice text are different speaking objects.

In some embodiments of the present application, the visual quality inspection interface includes at least one abnormal recording segment; the receiving module 1040 is configured to, in response to a playback instruction on the visual quality inspection interface, select a target abnormal recording segment from the at least one abnormal recording segment, and play back the target abnormal recording segment.

In some embodiments of the present application, the display module 1030 is configured to switch, in response to a playback instruction on a visual quality inspection interface, a play interface for displaying a target abnormal recording segment selected by the playback instruction, where the play interface for the target abnormal recording segment includes a time interval in which a target abnormal time period corresponding to the target abnormal recording segment is located;

the receiving module 1040 is configured to play back the target abnormal recording segment according to the time interval.

In some embodiments of the application, the time interval in which the target abnormal time period is located corresponds to n interval ranges, and the processing module 1050 is configured to determine, in response to a drag instruction on a play interface of the target abnormal recording segment, a time interval corresponding to an ith interval range from the n interval ranges corresponding to the time interval in which the target abnormal time period is located, where the time interval corresponding to the ith range includes at least one target abnormal time corresponding to the target abnormal recording segment, where n and i are integers greater than zero.

In some embodiments of the present application, the apparatus further comprises an extraction module 1060, a calculation module 1070, and a training module 1080;

the obtaining module 1010 is configured to obtain a sample telephone recording and a quality inspection result of the sample telephone recording, where the sample telephone recording is composed of at least one group of telephone recordings between a customer service and a user, and the quality inspection result of the sample telephone recording includes at least one of an emotion score and a sensitive vocabulary;

the extracting module 1060 is configured to extract emotional characteristic parameters from the sample telephone recording, where the emotional characteristic parameters include emotional characteristic vectors of k dimensions, and k is an integer greater than zero;

the processing module 1050 is used for inputting the emotional characteristic parameters into the initial recording quality inspection model to obtain a predicted quality inspection result of the sample telephone recording;

the calculating module 1070 is configured to calculate an error between the predicted quality inspection result and the quality inspection result;

and the training module 1080 trains the initial recording quality inspection model by adopting an error back propagation algorithm to obtain the trained recording quality inspection model.

Referring to FIG. 11, a block diagram of a computer device 1100 according to an exemplary embodiment of the present application is shown. The computer device 1100 may be a portable mobile terminal such as: smart phones, tablet computers, MP3 players (Moving Picture Experts Group Audio Layer III, motion video Experts compression standard Audio Layer 3), MP4 players (Moving Picture Experts Group Audio Layer IV, motion video Experts compression standard Audio Layer 4). Computer device 1100 may also be referred to by other names such as user equipment, portable terminals, etc.

Generally, the computer device 1100 includes: a processor 1101 and a memory 1102.

Processor 1101 may include one or more processing cores, such as a 4-core processor, an 8-core processor, or the like. The processor 1101 may be implemented in at least one hardware form of a DSP (Digital Signal Processing), an FPGA (Field-Programmable Gate Array), and a PLA (Programmable Logic Array). The processor 1101 may also include a main processor and a coprocessor, the main processor is a processor for processing data in an awake state, and is also called a Central Processing Unit (CPU); a coprocessor is a low power processor for processing data in a standby state. In some embodiments, the processor 1101 may be integrated with a GPU (Graphics Processing Unit) that is responsible for rendering and drawing the content that the display screen needs to display. In some embodiments, the processor 1101 may further include an AI (Artificial Intelligence) processor for processing computing operations related to machine learning.

Memory 1102 may include one or more computer-readable storage media, which may be tangible and non-transitory. Memory 1102 can also include high-speed random access memory, as well as non-volatile memory, such as one or more magnetic disk storage devices, flash memory storage devices. In some embodiments, a non-transitory computer readable storage medium in memory 1102 is used to store at least one instruction for execution by processor 1101 to implement the quality inspection method for telephone recordings provided herein.

In some embodiments, the computer device 1100 may also optionally include: a peripheral interface 1103 and at least one peripheral. Specifically, the peripheral device includes: at least one of radio frequency circuitry 1104, touch display screen 1105, camera 1106, audio circuitry 1107, positioning component 1108, and power supply 1109.

The peripheral interface 1103 may be used to connect at least one peripheral associated with I/O (Input/Output) to the processor 1101 and the memory 1102. In some embodiments, the processor 1101, memory 1102, and peripheral interface 1103 are integrated on the same chip or circuit board; in some other embodiments, any one or two of the processor 1101, the memory 1102 and the peripheral device interface 1103 may be implemented on separate chips or circuit boards, which is not limited by this embodiment.

The Radio Frequency circuit 1104 is used to receive and transmit RF (Radio Frequency) signals, also called electromagnetic signals. The radio frequency circuit 1104 communicates with communication networks and other communication devices via electromagnetic signals. The radio frequency circuit 1104 converts an electric signal into an electromagnetic signal to transmit, or converts a received electromagnetic signal into an electric signal. Optionally, the radio frequency circuit 1104 includes: an antenna system, an RF transceiver, one or more amplifiers, a tuner, an oscillator, a digital signal processor, a codec chipset, a subscriber identity module card, and so forth. The radio frequency circuit 1104 may communicate with other terminals via at least one wireless communication protocol. The wireless communication protocols include, but are not limited to: the world wide web, metropolitan area networks, intranets, generations of mobile communication networks (2G, 3G, 4G, and 5G), Wireless local area networks, and/or WiFi (Wireless Fidelity) networks. In some embodiments, the rf circuit 1104 may further include NFC (Near Field Communication) related circuits, which are not limited in this application.

The touch display screen 1105 is used to display a UI (User Interface). The UI may include graphics, text, icons, video, and any combination thereof. The touch display screen 1105 also has the ability to capture touch signals on or over the surface of the touch display screen 1105. The touch signal may be input to the processor 1101 as a control signal for processing. Touch display 1105 is used to provide virtual buttons and/or a virtual keyboard, also referred to as soft buttons and/or a soft keyboard. In some embodiments, the touch display screen 1105 may be one, providing the front panel of the computer device 1100; in other embodiments, the touch display screens 1105 can be at least two, each disposed on a different surface of the computer device 1100 or in a folded design; in still other embodiments, the touch display screen 1105 may be a flexible display screen disposed on a curved surface or on a folded surface of the computer device 1100. Even more, the touch display screen 1105 can be arranged in a non-rectangular irregular pattern, i.e., a shaped screen. The touch Display screen 1105 can be made of LCD (Liquid Crystal Display), OLED (organic light-Emitting Diode), and other materials.

Camera assembly 1106 is used to capture images or video. Optionally, camera assembly 1106 includes a front camera and a rear camera. Generally, a front camera is used for realizing video call or self-shooting, and a rear camera is used for realizing shooting of pictures or videos. In some embodiments, the number of the rear cameras is at least two, and each of the rear cameras is any one of a main camera, a depth-of-field camera and a wide-angle camera, so that the main camera and the depth-of-field camera are fused to realize a background blurring function, and the main camera and the wide-angle camera are fused to realize a panoramic shooting function and a VR (Virtual Reality) shooting function. In some embodiments, camera assembly 1106 may also include a flash. The flash lamp can be a monochrome temperature flash lamp or a bicolor temperature flash lamp. The double-color-temperature flash lamp is a combination of a warm-light flash lamp and a cold-light flash lamp, and can be used for light compensation at different color temperatures.

The audio circuitry 1107 is used to provide an audio interface between a user and the computer device 1100. The audio circuitry 1107 may include a microphone and a speaker. The microphone is used for collecting sound waves of a user and the environment, converting the sound waves into electric signals, and inputting the electric signals to the processor 1101 for processing or inputting the electric signals to the radio frequency circuit 1104 to achieve voice communication. The microphones may be multiple and placed at different locations on the computer device 1100 for stereo sound acquisition or noise reduction purposes. The microphone may also be an array microphone or an omni-directional pick-up microphone. The speaker is used to convert electrical signals from the processor 1101 or the radio frequency circuit 1104 into sound waves. The loudspeaker can be a traditional film loudspeaker or a piezoelectric ceramic loudspeaker. When the speaker is a piezoelectric ceramic speaker, the speaker can be used for purposes such as converting an electric signal into a sound wave audible to a human being, or converting an electric signal into a sound wave inaudible to a human being to measure a distance. In some embodiments, the audio circuitry 1107 may also include a headphone jack.

The Location component 1108 is used to locate the current geographic Location of the computer device 1100 for navigation or LBS (Location Based Service). The Positioning component 1108 may be a Positioning component based on the Global Positioning System (GPS) in the united states, the beidou System in china, or the galileo System in russia.

The power supply 1109 is used to provide power to the various components within the computer device 1100. The power supply 1109 may be alternating current, direct current, disposable or rechargeable. When the power supply 1109 includes a rechargeable battery, the rechargeable battery may be a wired rechargeable battery or a wireless rechargeable battery. The wired rechargeable battery is a battery charged through a wired line, and the wireless rechargeable battery is a battery charged through a wireless coil. The rechargeable battery may also be used to support fast charge technology.

In some embodiments, the computer device 1100 also includes one or more sensors 1110. The one or more sensors 1110 include, but are not limited to: acceleration sensor 1111, gyro sensor 1112, pressure sensor 1113, fingerprint sensor 1114, optical sensor 1115, and proximity sensor 1116.

The acceleration sensor 1111 detects the magnitude of acceleration in three coordinate axes of a coordinate system established with the computer apparatus 1100. For example, the acceleration sensor 1111 is configured to detect components of the gravitational acceleration in three coordinate axes. The processor 1101 may control the touch display screen 1105 to display a user interface in a landscape view or a portrait view according to the gravitational acceleration signal of the set of acceleration sensors 1111. The acceleration sensor 1111 may be used for acquisition of motion data of a game or a user.

The gyro sensor 1112 may detect a body direction and a rotation angle of the computer device 1100, and the gyro sensor 1112 may collect a 3D motion of the user on the computer device 1100 together with the acceleration sensor 1111. From the data collected by gyroscope sensor 1112, processor 1101 may implement the following functions: motion sensing (such as changing the UI according to a user's tilting operation), image stabilization at the time of photographing, game control, and inertial navigation.

The pressure sensors 1113 may be disposed on the side bezel of the computer device 1100 and/or on the lower layer of the touch display screen 1105. When the pressure sensor 1113 is disposed on the side frame of the computer device 1100, the holding signal of the user to the computer device 1100 can be detected, and the left-right hand recognition or the shortcut operation can be performed according to the holding signal. When the pressure sensor 1113 is disposed at the lower layer of the touch display screen 1105, the operability control on the UI interface can be controlled according to the pressure operation of the user on the touch display screen 1105. The operability control comprises at least one of a button control, a scroll bar control, an icon control and a menu control.

The fingerprint sensor 1114 is used for collecting a fingerprint of a user to identify the identity of the user according to the collected fingerprint. Upon recognizing that the user's identity is a trusted identity, the user is authorized by the processor 1101 to perform relevant sensitive operations including unlocking the screen, viewing encrypted information, downloading software, paying for and changing settings, etc. The fingerprint sensor 1114 may be disposed on the front, back, or side of the computer device 1100. When a physical key or vendor Logo is provided on the computer device 1100, the fingerprint sensor 1114 may be integrated with the physical key or vendor Logo.

Optical sensor 1115 is used to collect ambient light intensity. In one embodiment, the processor 1101 may control the display brightness of the touch display screen 1105 based on the ambient light intensity collected by the optical sensor 1115. Specifically, when the ambient light intensity is high, the display brightness of the touch display screen 1105 is increased; when the ambient light intensity is low, the display brightness of the touch display screen 1105 is turned down. In another embodiment, processor 1101 may also dynamically adjust the shooting parameters of camera assembly 1106 based on the ambient light intensity collected by optical sensor 1115.

A proximity sensor 1116, also referred to as a distance sensor, is typically provided on the front side of the computer device 1100. The proximity sensor 1116 is used to capture the distance between the user and the front of the computer device 1100. In one embodiment, the touch display screen 1105 is controlled by the processor 1101 to switch from a bright screen state to a dark screen state when the proximity sensor 1116 detects that the distance between the user and the front face of the computer device 1100 is gradually decreasing; when the proximity sensor 1116 detects that the distance between the user and the front face of the computer device 1100 becomes progressively larger, the touch display screen 1105 is controlled by the processor 1101 to switch from a breath-screen state to a light-screen state.

Those skilled in the art will appreciate that the configuration illustrated in FIG. 11 does not constitute a limitation of the computer device 1100, and may include more or fewer components than those illustrated, or may combine certain components, or may employ a different arrangement of components.

The present application further provides a computer-readable storage medium, in which at least one instruction, at least one program, a code set, or a set of instructions is stored, and the at least one instruction, the at least one program, the code set, or the set of instructions is loaded and executed by a processor to implement the quality inspection method for telephone recordings provided by the above-mentioned method embodiments.

It should be understood that reference to "a plurality" herein means two or more. "and/or" describes the association relationship of the associated objects, meaning that there may be three relationships, e.g., a and/or B, which may mean: a exists alone, A and B exist simultaneously, and B exists alone. The character "/" generally indicates that the former and latter associated objects are in an "or" relationship.

It will be understood by those skilled in the art that all or part of the steps for implementing the above embodiments may be implemented by hardware, or may be implemented by a program instructing relevant hardware, where the program may be stored in a computer-readable storage medium, and the above-mentioned storage medium may be a read-only memory, a magnetic disk or an optical disk, etc.

The above description is only exemplary of the present application and should not be taken as limiting, as any modification, equivalent replacement, or improvement made within the spirit and principle of the present application should be included in the protection scope of the present application.

Claims

1. A method for quality testing of a telephone recording, the method comprising:

acquiring a telephone record to be quality-checked;

2. The method of claim 1, wherein the anomalous information comprises an anomalous mood score; the inputting the telephone recording into the recording quality inspection model, and identifying the abnormal time period and the abnormal information corresponding to the abnormal recording segment, comprises:

and inputting the telephone recording into the recording quality inspection model, and identifying a first time period corresponding to the recording segment containing the abnormal emotion score and a first speaking object corresponding to the abnormal emotion score.

3. The method of claim 2, wherein the record quality control model comprises: a mood recognition model; the inputting the telephone recording into the recording quality inspection model, and identifying a first time period corresponding to the recording segment containing the abnormal emotion score and a first speaking object corresponding to the abnormal emotion score, includes:

inputting the telephone recording into the emotion recognition model, and recognizing emotion scores and speaking objects of the telephone recording in different time periods;

determining a time period when a first emotion score exceeds an abnormal threshold as the first time period, and determining a speaking object corresponding to the first emotion score as the first speaking object; and/or determining a time period when the second emotion score exceeds the abnormal threshold value as the first time period, and determining a speaking object corresponding to the second emotion score as the first speaking object;

and the speaking object corresponding to the first emotion score and the speaking object corresponding to the second emotion score are different speaking objects.

4. The method of claim 1, wherein the exception information comprises sensitive words; the inputting the telephone recording into the recording quality inspection model, and identifying the abnormal time period and the abnormal information corresponding to the abnormal recording segment, comprises:

and inputting the telephone recording into the recording quality inspection model, and identifying a second time period corresponding to the recording fragment containing the sensitive vocabulary and a second speaking object corresponding to the sensitive vocabulary.

5. The method of claim 4, wherein the record quality control model comprises: converting the voice into a text model;

the inputting the telephone recording into the recording quality inspection model, and identifying a second time period corresponding to the recording segment containing the sensitive vocabulary and a second speaking object corresponding to the sensitive vocabulary, includes:

inputting the telephone recording into the voice-to-text model, and identifying voice texts and speaking objects corresponding to the telephone recording in different time periods;

determining a time period of the first voice text with the sensitive vocabulary as the second time period, and determining a speaking object corresponding to the first voice text as the second speaking object; and/or determining a time period of a second voice text with the sensitive vocabulary as the second time period, and determining a speaking object corresponding to the second voice text as the second speaking object;

and the speaking object corresponding to the first voice text and the speaking object corresponding to the second voice text are different speaking objects.

6. The method of any one of claims 1 to 5, wherein the visual quality inspection interface comprises at least one abnormal recording segment;

the responding to the playback instruction on the visual quality inspection interface, the playing back the abnormal recording segment includes:

responding to a playback instruction on the visual quality inspection interface, selecting a target abnormal recording segment from the at least one abnormal recording segment, and playing back the target abnormal recording segment.

7. The method of claim 6, wherein the selecting a target abnormal recording segment from the at least one abnormal recording segment in response to a playback instruction on the visual quality inspection interface, and playing back the target abnormal recording segment comprises:

responding to a playback instruction on the visual quality inspection interface, and switching and displaying a playing interface of a target abnormal recording section selected by the playback instruction, wherein the playing interface of the target abnormal recording section comprises a time interval of the target abnormal time period corresponding to the target abnormal recording section;

and playing back the target abnormal recording segment according to the time interval.

8. The method according to claim 7, wherein the time interval in which the target abnormal time period is located corresponds to n interval ranges, the method further comprising:

responding to an interval dragging instruction on a playing interface of the target abnormal recording segment, and determining a time interval corresponding to an ith interval range from n interval ranges corresponding to the time interval of the target abnormal time period, wherein the time interval corresponding to the ith range comprises at least one target abnormal time corresponding to the target abnormal recording segment, and both n and i are integers larger than zero.

9. The method of claim 1, wherein the record quality control model is trained by:

obtaining a sample telephone recording and a quality inspection result of the sample telephone recording, wherein the sample telephone recording is composed of at least one group of telephone recordings between a telephone customer service and a user, and the quality inspection result of the sample telephone recording comprises at least one of emotion scores and sensitive vocabularies;

extracting emotional characteristic parameters from the sample telephone recording, wherein the emotional characteristic parameters comprise emotional characteristic vectors of k dimensions, and k is an integer larger than zero;

inputting the emotional characteristic parameters into an initial recording quality inspection model to obtain a prediction quality inspection result of the sample telephone recording;

calculating the error between the predicted quality inspection result and the quality inspection result;

and training the initial recording quality inspection model by adopting an error back propagation algorithm to obtain a trained recording quality inspection model.

10. A quality control apparatus for telephone recordings, the apparatus comprising:

11. A computer device comprising a processor and a memory, the memory having stored therein at least one instruction, at least one program, a set of codes, or a set of instructions, the at least one instruction, the at least one program, the set of codes, or the set of instructions being loaded and executed by the processor to implement the method of quality testing of a voice recording according to any one of claims 1 to 9.

12. A computer storage medium having stored therein at least one instruction, at least one program, set of codes, or set of instructions, which is loaded and executed by a processor to implement a method of quality testing of a telephone recording according to any one of claims 1 to 9.