Audio resume generation and page access method and device, storage medium and program product

Abstract

The embodiment of the invention provides an audio resume generation and page access method and device, a storage medium and a program product. In the embodiment of the invention, according to the identification information of the target service personnel meeting the audio resume generation condition, the call record of the broker of the target service platform related to the target service personnel is obtained, and the audio information only containing the call content of the target service personnel is extracted through the sound channel of the call equipment; generating clone based on the audio information; furthermore, a target audio clip meeting audio quality and duration requirements can be selected from the audio information, the target audio clip and the text introduction information of the target service personnel are input into a sound clone model based on deep learning, and audio resume information with sound features of the target service personnel is generated. The richness of the resume can be improved, the user can quickly know useful information related to the target service personnel by listening to the audio resume information, and the conversion rate of the resume landing page is improved.

Description

Technical Field

[0001] The present application relates to the field of computer processing technology, and in particular to an audio resume generation and page access method, device, storage medium and program product. Background Art

[0002] With the development of Internet technology, it has become mainstream for job seekers to use electronic resumes. Recruiters can fully understand the personal identity, work experience, areas of expertise and other relevant information of job seekers through electronic resumes. In some scenarios, such as housekeeping service scenarios, the housekeeping platform can query the electronic resumes of housekeeping staff. Users can get a deep understanding of housekeeping staff based on the electronic resumes queried on the housekeeping platform and choose housekeeping staff that suits their needs.

[0003] Although electronic resumes are now widely used, the content format is relatively simple and contains a lot of text information, making it difficult for users to quickly obtain useful information, resulting in a low conversion rate of resume landing pages. Summary of the invention

[0004] Multiple aspects of the present application provide an audio resume generation and page access method, device, storage medium and program product to improve the richness of the resume, and can quickly obtain useful information related to the target service personnel, thereby improving the conversion rate of the resume landing page.

[0005] An embodiment of the present application provides an audio resume generation method, which is applied to a target service platform, and includes: determining a target service personnel who meets the conditions for generating an audio resume from service personnel who have signed a contract with the target service platform; obtaining call record information between a broker of the target service platform and the target service personnel from a call record database according to the identification information of the target service personnel, and the two channels of the target call device correspond to the broker and the target service personnel respectively; when the call record information includes two channels, parsing the call record information into audio information on the two channels, and obtaining the audio information on the channels corresponding to the target service personnel as the target audio information; extracting multiple audio clips from the target audio information, and selecting a target audio clip that meets the audio quality and duration requirements from the multiple audio clips; inputting the target audio clip and the text introduction information of the target service personnel into a sound cloning model based on deep learning for sound cloning, so as to obtain the audio introduction information corresponding to the text introduction information; adding the audio introduction information to the details page of the target service personnel, so as to obtain the audio resume page of the target service personnel.

[0006] The embodiment of the present application also provides a page access method, which is applied to a terminal device, and the terminal device runs a target application. The method includes: displaying a service personnel list page provided by the target application, the service personnel list page including introduction information of multiple service personnel; in response to a trigger operation on the introduction information of any service personnel, sending an access request for the details page of any service personnel to the target service platform; upon receiving an audio resume page of any service personnel returned by the target service platform, jumping from the service personnel list page to the audio resume page, on which audio introduction information is displayed; upon receiving a details page of any service personnel returned by the target service platform, jumping from the service personnel list page to the details page; wherein, when the target service platform includes the storage address of the audio introduction information in the details page of any service personnel, the audio introduction information is added to the details page to obtain the audio resume page and returned to the terminal device; or, when the target service platform does not include the storage address of the audio introduction information in the details page of any service personnel, the details page is returned to the terminal device.

[0007] An embodiment of the present application also provides an electronic device, including: a memory and a processor; the memory is used to store a computer program; the processor is coupled to the memory and is used to execute the computer program to implement the steps in the above method.

[0008] An embodiment of the present application also provides a computer-readable storage medium storing a computer program. When the computer program is executed by a processor, the processor implements the steps in the above method.

[0009] An embodiment of the present application also provides a computer program product, which includes a computer program / instructions. When the computer program / instructions are executed by a processor, the processor is enabled to implement the steps in the above method.

[0010] In an embodiment of the present application, based on the identification information of the target service personnel who meets the conditions for generating an audio resume, the call records between the agent of the target service platform and the target service personnel are obtained, and the audio information containing only the call content of the target service personnel is extracted through the sound channel of the call device. Cloning is performed based on the audio information to avoid voice interference from the target service personnel. Furthermore, a target audio clip that meets the audio quality and duration requirements can be selected from the audio information, and the target audio clip and the text introduction information of the target service personnel are input into a sound cloning model based on deep learning. Based on the target sound clip, information such as the sound characteristics of the target service personnel is extracted, and based on the sound characteristics and text introduction of the target service personnel, audio resume information with the sound characteristics of the target service personnel is generated. As a result, the richness of the resume can be improved, and the user can quickly understand the useful information related to the target service personnel by listening to the audio resume information, thereby improving the conversion rate of the resume landing page. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] The drawings described herein are used to provide a further understanding of the present application and constitute a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0012] Figure 1 A structural schematic diagram of an audio resume generation system provided for an exemplary embodiment of the present application;

[0013] Figure 2 A flowchart of a method for generating an audio resume provided as an exemplary embodiment of the present application;

[0014] Figure 3 A flowchart of a page access method provided by an exemplary embodiment of the present application;

[0015] Figure 4 A schematic diagram of the structure of an audio resume generating device provided by an exemplary embodiment of the present application;

[0016] Figure 5 A schematic diagram of the structure of a page access device provided by an exemplary embodiment of the present application;

[0017] Figure 6 A schematic structural diagram of an electronic device provided as an exemplary embodiment of the present application. DETAILED DESCRIPTION

[0018] In order to make the purpose, technical solution and advantages of the present application clearer, the technical solution of the present application will be clearly and completely described below in combination with the specific embodiments of the present application and the corresponding drawings. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present application.

[0019] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with the relevant laws, regulations and standards of the relevant countries and regions, and provide corresponding operation entrances for users to choose to authorize or refuse.

[0020] In view of the technical problem that the content of the above-mentioned resume is relatively simple in form and the electronic resume contains a lot of information, it is difficult for users to quickly obtain useful information, resulting in a low conversion rate of the resume landing page. In an embodiment of the present application, based on the identification information of the target service personnel who meets the conditions for generating an audio resume, the call records between the agent of the target service platform and the target service personnel are obtained, and the audio information containing only the call content of the target service personnel is extracted through the sound channel of the call device. Cloning is performed based on the audio information to avoid voice interference from the target service personnel; further, a target audio clip that meets the audio quality and duration requirements can be selected from the audio information, and the target audio clip and the text introduction information of the target service personnel are input into a sound cloning model based on deep learning, and the sound characteristics and other information of the target service personnel are extracted based on the target sound clip. Based on the sound characteristics and text introduction of the target service personnel, audio resume information with the sound characteristics of the target service personnel is generated. In this way, the richness of the resume can be improved, and the user can quickly understand the useful information related to the target service personnel by listening to the audio resume information, thereby improving the conversion rate of the resume landing page.

[0021] A solution provided by an embodiment of the present application is described in detail below in conjunction with the accompanying drawings.

[0022] Figure 1 The following is a schematic diagram of the structure of an audio resume generation system provided by an exemplary embodiment of the present application. Figure 1 As shown, the system includes: a target service platform 101 and a terminal device 102, and the target service platform 101 and the terminal device 102 can be connected to each other in communication.

[0023] In this embodiment, the communication connection between the target service platform 101 and the terminal device 102 can be a wireless communication connection mode or a wired connection mode. For the wireless communication connection mode, optionally, the communication connection can be achieved through a mobile network, and accordingly, the network standard of the mobile network can be 2G (GSM), 2.5G (GPRS), 3G (WCDMA, TD-SCDMA, CDMA2000, UTMS), 4G (LTE), 4G + (LTE +), 5G, WiMax or any of the new network standards that will appear in the future. In addition, when the target service platform 101 and the terminal device 102 are located in the same local area network, wireless communication connection can also be achieved through Bluetooth, WiFi, infrared, Zigbee or NFC. For the wired communication connection mode, optionally, the communication connection can be achieved through Ethernet, serial communication, Universal Serial Bus (USB), High Definition Multimedia Interface (HDMI) and the like. Among them, Ethernet is usually used for local area network (LAN) connection, which supports multiple transmission rates such as 10Mbps, 100Mbps, 1Gbps, 10Gbps, etc. Serial communication includes standards such as RS-232 and RS-485; it is used to connect computers and external devices such as keyboards, mice, printers, etc. USB includes USB 1.1, USB 2.0, USB 3.0, USB 3.1, USB 4 and other versions, each with different transmission rates.

[0024] In this embodiment, the type of the target service platform is not limited. The target service platform may be, for example, a housekeeping service platform, a rental service platform, or a food delivery service platform, etc. If the target service platform is a housekeeping service platform, the target service personnel may be housekeeping service personnel; if the target service platform is a rental service platform, the target service personnel may be a house viewing personnel; if the target service platform is a food delivery service platform, the target service personnel may be a food delivery person, etc.

[0025] In addition, this embodiment does not limit the implementation form of the target service platform. The target service platform 101 can be a single server device, a cloud server array, or a virtual machine (VM) running in a cloud server array. In addition, the target service platform can also be other computing devices with corresponding service capabilities, such as computers and other terminal devices (running service programs), etc.

[0026] This embodiment does not limit the implementation form of the terminal device. The terminal device can be a smart handheld device, such as a smart phone, tablet computer, laptop computer or desktop computer, etc.; or, the terminal device can also be a smart wearable device, such as a smart watch, smart bracelet, etc.; or, the terminal device can also be various smart home appliances with display screens, such as smart TVs, smart large screens or smart robots, etc.

[0027] In this embodiment, when the timing for generating the audio introduction information arrives, the target service platform can trigger the generation mechanism of the audio introduction information. Since the target service platform may include many types of domestic staff, some service staff may meet the conditions for generating the audio introduction information, and some service staff may not meet the conditions for generating the audio introduction information. In the generation mechanism of the audio introduction information, first, the service staff who currently meets the conditions for generating the audio introduction information can be determined from the service staff who have signed a contract with the target service platform. This embodiment does not limit the generation conditions of the audio introduction information, for example, newly signed employees, directly hired employees, or employees with call recordings, etc. In addition, in order to make the audio introduction information closer to the voice characteristics of the target service staff himself, the target audio of the target service staff can also be generated based on the sound characteristic information contained in the audio sample of the target service staff. The audio sample can be, for example, call record information with the agent of the target service platform. For the convenience of description and distinction, the service staff who meet the conditions for generating the audio introduction information are referred to as target service staff.

[0028] Among them, this embodiment does not limit the generation time of audio introduction information and does not limit the triggering method of the generation mechanism of audio introduction information. For example, the triggering method can be to manually trigger the generation mechanism of audio introduction information in real time, such as when the target service platform has a demand for audio introduction information generation (i.e., the generation time of audio introduction information), the audio introduction information generation mechanism is triggered by the manual operation of the agent or other staff on the target service platform. Alternatively, the generation time of audio introduction information can be pre-set on the target service platform, and when the time is reached, the generation mechanism of audio introduction information is actively triggered. Among them, the pre-set time can be, for example, a time point in a day, or a trigger cycle or trigger event, etc.; a time point in a day can be, for example, one or more time points in 8 am, 12 noon or 8 pm, etc.; the trigger cycle can be, for example, triggered once a day, once a week, or once a month, etc.; the trigger event can be, for example, one or a combination of the number of service personnel without audio introduction information exceeding a threshold, the signing of a new service personnel, or the recruitment of service personnel through direct recruitment.

[0029] In an optional embodiment, target service personnel who meet the conditions for generating audio introduction information are determined from the service personnel who have signed contracts with the target service platform, including: when the target service platform has newly signed service personnel, determine whether the recruitment method of the newly signed service personnel is direct recruitment, and determine whether the broker has had a conversation with the newly signed service personnel; if both are yes, the newly signed service personnel is used as the target service personnel; and / or, from the service personnel who have signed contracts with the target service platform, select candidate service personnel for whom there is no audio introduction information page, determine whether the recruitment method of the candidate service personnel is direct recruitment, and determine whether the broker has had a conversation with the candidate service personnel; if both are yes, the candidate service personnel is used as the target service personnel. Among them, direct recruitment refers to the method of directly recruiting service personnel by the target service platform.

[0030] The above two implementation methods for determining the target service personnel who meet the conditions for generating audio introduction information can be selected from one of them, or the two methods can be combined. That is, from the service personnel who have signed contracts with the target service platform, the target service personnel who meet the conditions for generating audio introduction information are determined. One implementation method is: when the target service platform has a newly signed service personnel, it is determined whether the recruitment method of the newly signed service personnel is a direct recruitment method, and whether the broker has had a conversation with the newly signed service personnel; if both are yes, the newly signed service personnel is used as the first target service personnel. Alternatively, from the service personnel who have signed contracts with the target service platform, the target service personnel who meet the conditions for generating audio introduction information are determined. Another implementation method is: from the service personnel who have signed contracts with the target service platform, select candidate service personnel who do not have an audio introduction information page, determine whether the recruitment method of the candidate service personnel is a direct recruitment method, and determine whether the broker has had a conversation with the candidate service personnel; if both are yes, the candidate service personnel is used as the second target service personnel; the service personnel who overlap between the first target service personnel and the second target personnel are used as the target personnel. Alternatively, target service personnel who meet the conditions for generating audio introduction information are determined from among the service personnel who have signed contracts with the target service platform. Another implementation method is: when the target service platform has newly signed service personnel, determine whether the recruitment method of the newly signed service personnel is direct recruitment, and determine whether the agent has had a conversation with the newly signed service personnel; if both are yes, the newly signed service personnel is taken as the first target service personnel; and, from the service personnel who have signed contracts with the target service platform, select candidate service personnel for whom no audio introduction information page exists, determine whether the recruitment method of the candidate service personnel is direct recruitment, and determine whether the agent has had a conversation with the candidate service personnel; if both are yes, the candidate service personnel is taken as the second target service personnel; and the personnel included in the first target personnel and the second target personnel are taken as the target personnel.

[0031] Furthermore, the call record information between the agent of the target service platform and the target service personnel can be obtained from the call record database according to the identification information of the target service personnel. The call record database can store the call records of various call devices, and the identification ID of the service personnel is used as the index of the call record to facilitate the search for the corresponding call record.

[0032] In this embodiment, in order to further improve the authenticity of the audio introduction information and be closer to the voice characteristics of the target service personnel, the voice characteristics of the target service personnel can be learned through voice cloning technology, and based on the voice characteristics, audio introduction information that is infinitely close to the voice characteristics of the target service personnel can be generated. Based on this, the call record information of the target service personnel can be obtained, and based on the call record information and text introduction information of the target service personnel, the voice cloning model of deep learning can be used to perform voice cloning. To obtain the audio introduction information corresponding to the text information. The following embodiments are described using the call record information of the target service platform agent and the target service personnel as an example, but this does not constitute a limitation on this application.

[0033] In this embodiment, the call device between the agent of the target service platform and the target service personnel may include two channels, the two channels correspond to the agent and the target service personnel respectively, one channel is used to transmit the audio information of the agent, and the other channel is used to transmit the audio information of the target service personnel; accordingly, the call record information between the agent and the target service personnel is mixed with audio information of two channels. Based on this, in this embodiment, it can be determined whether the number of channels included in the call record information exceeds two, and when the call record information includes two channels, the call record information is parsed into audio information on the two channels, and the audio information on the channel corresponding to the target service personnel is obtained as the target audio information. For the convenience of description and distinction, the two channels can be referred to as the left channel and the right channel respectively. Among them, the agent can correspond to the left channel, and the target service personnel can correspond to the right channel; or, the agent can correspond to the right channel, and the target service personnel can correspond to the left channel. Take the example that the broker corresponds to the left channel and the target service personnel corresponds to the right channel, that is, the left channel is used to transmit the audio information of the broker, and the right channel is used to transmit the audio information of the target service personnel; accordingly, the call record between the broker and the target service personnel is mixed with the audio information of the left channel and the audio information of the right channel, and the audio information on the corresponding channel of the target service personnel is the audio information of the right channel.

[0034] Furthermore, there may be noise in the call record information, and the noise may be, for example, audio information of a third party, environmental noise or signal noise, etc., which is not limited in this embodiment. In order to avoid the interference of noise, the call record information may be subjected to a first noise reduction process to obtain the noise-reduced call record information. The first noise reduction process is a technical process aimed at extracting useful voice signals from noisy audio signals and suppressing or reducing background noise interference, which is crucial to improving voice quality, enhancing the listener's comfort, and improving the intelligibility of voice information transmission.

[0035] Among them, the first noise reduction processing can be divided into traditional signal processing methods and modern methods based on deep learning. For traditional signal processing methods, they mainly include spectral subtraction, Wiener filtering, subspace method or speech amplitude spectrum estimation method based on minimum mean square error, etc. Among them, spectral subtraction: it is to estimate the noise spectrum in the target audio information and subtract the noise spectrum from the noisy spectrum to obtain noise-free target audio information. Wiener filtering: it is a linear filter designed using the minimum mean square error criterion, which is suitable for steady-state noise. The noise wave is filtered out by this new filter to obtain noise-free target audio information. Subspace algorithm: it is to decompose the signal into signal subspace and noise subspace through singular value decomposition, remove the noise subspace, and obtain noise-free target audio information. Real-time noise reduction algorithm based on statistical model: such as Wiener filtering, which uses statistical methods to estimate the noise and language energy of each frequency point of the audio, removes the noise of each frequency point, and obtains noise-free target audio information. For deep learning methods, it mainly includes methods based on spectrum mapping, methods based on time-frequency masking or end-to-end methods, etc. Among them, the method based on spectrum mapping is to establish a mapping relationship table between noisy spectrum parameters and pure audio spectrum parameters through a deep neural network, and then remove the audio corresponding to the noise spectrum parameters in the mapping relationship table according to the mapping relationship table to obtain noise-free target audio information. The method based on time-frequency mask is to reflect the degree of noise suppression on each time-frequency unit by predicting the time-frequency mask, and then reconstruct the pure speech signal to obtain noise-free target audio information. The end-to-end method is to operate directly at the time domain waveform level, and directly input the noisy speech waveform through the model to obtain the enhanced waveform, so as to obtain noise-free target audio information.

[0036] Taking the signal processing method as an example, the first noise reduction process is described in detail, which specifically includes the following steps: Noise detection: First, the system needs to identify the noise in the target audio. Specifically, this can be achieved by analyzing the spectral characteristics of the audio signal. The noise usually has spectral characteristics different from the useful signal. Noise modeling: Once the noise is detected, the system will try to establish a mathematical model of the noise. This model can be simple, such as assuming that the noise is white noise (random noise with uniform power spectral density), or complex, such as a model based on a machine learning algorithm. Signal and noise separation: Using the noise model, the noise is separated from the original audio signal. This can be achieved by subtracting the output of the noise model, or through more complex algorithms such as spectral subtraction, Wiener filtering or Kalman filtering. Signal reconstruction: After removing the noise, the audio signal is reconstructed. This step includes filling the signal gaps that may be caused by noise reduction and enhancing the characteristics of the useful signal. Feedback and adjustment: After the noise reduction process, the processed audio information is analyzed to ensure that the noise reduction effect meets the expectations. If the effect is not good, the noise reduction parameters can be adjusted and reprocessed. Machine Learning: Machine learning algorithms can be used to improve noise reduction effects. These algorithms can be trained to learn the characteristics of noise and useful signals, so that they can more accurately identify and eliminate noise, thereby obtaining noise-free target audio information.

[0037] Furthermore, the target audio information may contain multiple audio clips, and there is a time interval between two adjacent audio clips. In order to ensure the accuracy of the target service personnel's voice characteristics and the cloning quality, an audio clip that meets certain requirements can be selected from the multiple sound clips as the basis for cloning. Specifically, multiple audio clips can be extracted from the target audio information, and an audio clip that meets the audio quality and duration requirements can be selected from the multiple audio clips as the target audio clip.

[0038] In an optional embodiment, multiple audio segments are extracted from the target audio information, including: performing a second noise reduction process on the target audio information, and identifying and removing silent segments from the target audio information after noise reduction to obtain the audio information to be segmented; dividing the audio information to be segmented into multiple audio segments according to the continuity and semantic integrity of the audio information to be segmented, and the pause intervals in the same audio segment are all less than the set interval threshold and express complete semantic information. Among them, the silent segment refers to the time period corresponding to the agent's speech contained in the target audio information, during which the target service personnel is in a silent state. Among them, the specific implementation method of the second noise reduction process can refer to the relevant description of the first noise reduction process, which will not be repeated here.

[0039] Among them, in order to ensure that each extracted audio segment has continuous and complete semantics, the audio information to be segmented is divided into multiple audio segments according to the continuity and semantic integrity of the audio information to be segmented. An optional implementation method includes: identifying pause points in the audio information to be segmented and the pause duration of each pause point; and identifying whether the semantic information of the audio segment between two adjacent pause points is complete; taking the pause point corresponding to the pause duration exceeding the preset duration and making the audio segment after segmentation have complete semantics as the segmentation point; and dividing the audio information to be segmented into multiple audio segments according to the segmentation point.

[0040] Among them, an optional implementation method for identifying whether the semantic information of the audio segment between two adjacent pause points is complete includes: extracting text information of the audio segment between two adjacent pause points; performing semantic analysis on the text information to obtain corresponding semantic information; extracting multiple key semantic words in the semantic information; calculating the closeness of the causal relationship between each semantic word; and determining that the semantic information of the audio segment between the two adjacent pause points is complete when the closeness of the causal relationship meets a threshold.

[0041] In an optional embodiment, a target audio segment that meets the audio quality and duration requirements is selected from multiple audio segments, including: performing duration statistics on multiple audio segments to obtain the durations of multiple audio segments, and selecting candidate audio segments that meet the duration requirements; performing subjective quality analysis and / or objective quality analysis on the candidate audio segments to obtain the subjective quality score and / or objective quality score of the candidate audio segments; and selecting a candidate audio segment whose subjective quality score and / or objective quality score is greater than the corresponding score threshold from the candidate audio segments as the target audio segment. For subjective quality analysis, a sufficient number of testers can be organized to listen to multiple audio segments and give each audio segment a score. For objective quality analysis, some indicators required for evaluating the quality of audio information, such as signal-to-noise ratio (SNR), spectrum difference, etc., can be calculated mathematically, and then these indicators are combined to fit an objective score.

[0042] More specifically, selecting a target audio segment that meets the audio quality and duration requirements from a plurality of audio segments includes: performing duration statistics on a plurality of audio segments to obtain the durations of a plurality of audio segments, and selecting a candidate audio segment that meets the duration requirements; performing subjective quality analysis on the candidate audio segments to obtain the subjective quality scores of the candidate audio segments; and selecting a candidate audio segment whose subjective quality scores are greater than a corresponding score threshold from the candidate audio segments as the target audio segment. Alternatively, selecting a target audio segment that meets the audio quality and duration requirements from a plurality of audio segments includes: performing duration statistics on a plurality of audio segments to obtain the durations of a plurality of audio segments, and selecting a candidate audio segment that meets the duration requirements; performing objective quality analysis on the candidate audio segments to obtain the objective quality scores of the candidate audio segments; and selecting a candidate audio segment whose objective quality scores are greater than a corresponding score threshold from the candidate audio segments as the target audio segment. Alternatively, a target audio segment that meets the audio quality and duration requirements is selected from multiple audio segments, including: performing duration statistics on the multiple audio segments to obtain the durations of the multiple audio segments, and selecting candidate audio segments that meet the duration requirements; performing subjective quality analysis and objective quality analysis on the candidate audio segments to obtain subjective quality scores and objective quality scores of the candidate audio segments; calculating a weighted average of the subjective quality scores and the objective quality scores; and selecting, from the candidate audio segments, a candidate audio segment whose weighted average corresponds to a score threshold as the target audio segment.

[0043] Furthermore, after obtaining the target audio segment that meets the audio quality and duration requirements, the target audio segment and the text introduction information of the target service personnel can be input into a deep learning-based voice cloning model for voice cloning to obtain the audio introduction information corresponding to the text introduction information. Among them, the voice cloning model can be an artificial intelligence AI model that is pre-trained based on text information samples and audio information samples corresponding to the text information samples, has learned voice features, and generates audio information corresponding to the text information based on the voice features and text information. The voice cloning model can be, for example, a CosyVoice model, which has the following core features: Multi-language support: It can support speech synthesis in multiple languages ​​including Chinese, English, Japanese, Cantonese, Korean, etc.; Zero-sample voice cloning: It can imitate the voice of the target speaker with only a few seconds of voice samples to achieve zero-sample voice cloning; Cross-language speech synthesis: It can use a voice sample of one language to synthesize the voice of another language, and has the ability of cross-language speech synthesis; Command control: Users can control the details of the generated voice, such as emotion and tone, through specific commands, to achieve fine control of the speech synthesis process; Full-stack capabilities: CosyVoice provides a complete solution from model training, reasoning to deployment. The working principle of CosyVoice can be summarized into the following key steps: Preprocessing: The input text is first preprocessed, including word segmentation, language detection, etc., to prepare text data; Text-to-speech synthesis: Use an end-to-end neural network model to convert text into speech. In this process, the model learns the semantics and context of the text to generate speech with the correct intonation, rhythm, and emotion; Voice cloning: Through zero-sample learning technology, CosyVoice can learn and imitate the voice characteristics of the target speaker in a short time (3 to 10 seconds), including rhythm and emotion; Cross-language synthesis: CosyVoice can use speech samples of one language to synthesize speech in another language, realizing cross-language speech synthesis; Command control: Users can control the emotion, tone, speed, etc. of the generated speech through specific commands, realizing fine-tuning of the speech synthesis process; Streaming reasoning: CosyVoice supports streaming reasoning mode, which can realize real-time speech synthesis, suitable for application scenarios that require instant response. The technical implementation of CosyVoice involves multiple aspects: large-scale pre-training: based on more than 170,000 hours of multilingual audio data for training, a powerful speech generation capability has been established; innovative model architecture: the use of advanced neural network architectures, such as Transformer and flow matching (Flow Matching) and other technologies, improves the performance and efficiency of the model; multi-task learning: through the multi-task learning framework, CosyVoice can simultaneously handle multi-language, multi-speaker, and multi-style speech synthesis tasks; efficient reasoning optimization: use KV cache, SDPA and other technologies to optimize the reasoning process and improve real-time performance.

[0044] In this embodiment, the sound cloning model includes at least: a preprocessing network, a feature extraction network and a sound synthesis network, and the target audio clip and the text introduction information of the target service personnel are input into the sound cloning model based on deep learning for sound cloning to obtain the audio introduction information corresponding to the text introduction information, including: inputting the target audio clip into the preprocessing network in the sound cloning model based on deep learning, performing the third noise reduction process on the target audio clip to obtain the intermediate audio clip; inputting the intermediate audio clip into the feature extraction network for sound feature extraction to obtain spectrum feature information, the spectrum feature information is a spectrum representation of the intermediate audio clip, including at least one of the frequency, amplitude and phase information of the intermediate audio clip; inputting the spectrum feature information and the text introduction information into the sound synthesis network, and performing sound synthesis on the text introduction information according to the spectrum feature information to obtain the audio introduction information. Among them, the specific implementation method of the third noise reduction process can refer to the relevant description of the first noise reduction process mentioned above, which will not be repeated here.

[0045] The intermediate audio clip is input into the feature extraction network for sound feature extraction to obtain spectral feature information, including the following steps: windowing: in order to reduce the discontinuity of the frame boundary, each frame of data is multiplied by a window function (such as a Hamming window) to avoid spectrum leakage; Fourier transform: fast Fourier transform (FFT) or short-time Fourier transform (STFT) is performed on each frame of windowed data to convert the time domain signal into a frequency domain signal to obtain a spectrum graph; feature extraction: useful features are extracted from the spectrum graph, such as Mel Frequency Cepstra l Coefficients (MFCC), spectrum centroid, spectrum bandwidth, Mel spectrum, etc.; feature regularization: the extracted features are regularized, such as mean normalization or standardization, to eliminate the differences between different recordings; classification: the extracted features are further learned and classified; feature learning: more abstract sound features are extracted; output: based on more abstract features, spectral feature information is obtained. It should be noted that the output can also be a classification label, a feature vector or other forms of representation, and the form of the output information depends on the specific application requirements.

[0046] The frequency spectrum feature information and the text introduction information are input into the sound synthesis network, and the text introduction information is synthesized according to the frequency spectrum feature information to obtain the audio introduction information, including the following steps: preprocessing the text introduction information: preprocessing the text introduction information, including word segmentation, removing stop words, converting to a standard form, etc., so as to extract useful semantic information; feature extraction: extracting key features from the preprocessed text using natural language processing technology, such as TF-IDF, Word2Vec, Bidirectional Encoder Representation fromTransformers, BERT), etc. to obtain the semantic representation of the text; spectral feature processing: analyze the spectral feature information, which may include Mel-frequency cepstral coefficients, spectrograms, etc. These features can describe the physical characteristics of the sound; mapping processing: learn the mapping relationship between text features and spectral features; sound synthesis: combine text features and spectral features to generate a sound waveform; generate audio introduction information: generate the corresponding audio introduction information according to the text introduction information and the sound waveform; post-processing: post-process the generated audio introduction information to obtain the processed audio introduction information, such as volume normalization, denoising, etc., to improve the sound quality.

[0047] It should be noted that the sound cloning model may also include a text information conversion network, through which the key information in the target audio information can be extracted, and whether the text introduction information contains all the key information in the target audio information can be determined. If the determination result is no, the key information missing from the text introduction information is updated to the text introduction information, so that the cloning model can generate the corresponding audio introduction information based on the updated text introduction information. In addition, the sound cloning model may also include a text information simplification network, which is used to extract key words and phrases in the text information, so that the sound synthesis network can generate the target audio resume information based on the extracted key words and phrases, thereby simplifying the audio resume content, so that users can more quickly understand the relevant services of the target service personnel.

[0048] Furthermore, in order to facilitate users to find and listen to the audio introduction information of the target service personnel, after obtaining the audio introduction information of the target service personnel, the audio introduction information can be added to the details page of the target service personnel to obtain the audio introduction information page of the target service personnel. This includes: adding the storage address of the audio introduction information in the audio database to the details page of the target service personnel; when the user initiates an access request to the details page, according to the storage address, the audio introduction information is added to the details page to obtain the audio introduction information page, and sent to the user's terminal device, so that the terminal device can display the audio introduction information page to the user.

[0049] Optionally, in order to make it easier for users to find and listen to the audio introduction information of the target service personnel, the storage address of the audio introduction information in the audio database is added to the details page of the target service personnel, including: determining the location information of the target card area from the details page of the target service personnel, the target card area is the top area of ​​the details page or the area for introducing the identity of the target service personnel; according to the location information of the target card area, binding the storage address of the audio introduction information in the associated area of ​​the specific information in the target card area, so as to display the audio introduction information in the associated area. Among them, the specific information can be the avatar, name or brief text introduction of the target service personnel, etc.; accordingly, the avatar, name or brief text introduction is divided into corresponding display areas; the associated area of ​​the specific information can be the adjacent area of ​​the avatar, name or brief text introduction divided into corresponding display areas.

[0050] For the terminal device, the terminal device can run the target application and display the service personnel list page provided by the target application.

[0051] Specifically, the service personnel list page includes the introduction information of multiple service personnel; in response to the trigger operation of the introduction information of any service personnel, an access request for the details page of any service personnel is sent to the target service platform; in the case of receiving the audio introduction information page of any service personnel returned by the target service platform, the service personnel list page is jumped to the audio introduction information page, and the audio introduction information is displayed on the audio introduction information page; in the case of receiving the details page of any service personnel returned by the target service platform, the service personnel list page is jumped to the details page. Wherein, if the target service platform includes the storage address of the audio introduction information in the details page of any service personnel, the audio introduction information is added to the details page to obtain the audio introduction information page and returned to the terminal device; or, if the target service platform does not include the storage address of the audio introduction information in the details page of any service personnel, the details page is returned to the terminal device.

[0052] Furthermore, the audio introduction information is displayed on the audio introduction information page. This embodiment does not limit the implementation form of the audio introduction information. For example, it can be a rectangular audio bar with a play button on the audio bar and the length of the voice will be displayed, usually in seconds. Further, in response to the user's play trigger operation on the play button, the audio introduction information is played, and a simple waveform chart will be displayed on the audio bar during the playback process to indicate the ups and downs of the voice; correspondingly, in response to the user's stop trigger operation on the play button, the audio introduction information is stopped. In addition, the audio bar also has the function of voice-to-text conversion. For example, long pressing the audio bar will cause a pop-up window to appear in the associated area of ​​the audio bar. The pop-up window includes at least a translation option and various translation language options under the translation option. In response to the selection operation of the translation option and the target translation language option, the audio information will be converted into the language and text corresponding to the target translation voice option for easy reading.

[0053] In the above-mentioned embodiments of the present application, according to the identification information of the target service personnel who meets the conditions for generating audio introduction information, the call records between the agent of the target service platform and the target service personnel are obtained, and the audio information containing only the call content of the target service personnel is extracted through the sound channel of the calling device, and clones are generated based on the audio information, which can avoid the sound interference of the target service personnel; further, a target audio segment that meets the audio quality and duration requirements can be selected from the audio information, and the target audio segment and the text introduction information of the target service personnel are input into a sound cloning model based on deep learning, and the sound characteristics of the target service personnel and other information are extracted based on the target sound segment, and audio introduction information with the sound characteristics of the target service personnel is generated based on the sound characteristics and text introduction of the target service personnel. As a result, the richness of the resume can be improved, and the user can quickly understand the useful information related to the target service personnel by listening to the audio resume information, thereby improving the conversion rate of the resume landing page.

[0054] Figure 2 The following is a flow chart of a method for generating an audio resume provided by an exemplary embodiment of the present application. Figure 2 As shown, the method is applied to a target service platform, and the method includes:

[0055] 201. Determine target service personnel who meet the conditions for generating audio resumes from among the service personnel who have signed contracts with the target service platform;

[0056] 202. Acquire call record information between a broker of the target service platform and the target service personnel from a call record database according to the identification information of the target service personnel;

[0057] 203. When the call record information includes two channels, parse the call record information into audio information on the two channels, and obtain the audio information on the channel corresponding to the target service personnel as the target audio information;

[0058] 204. Extract multiple audio segments from the target audio information, and select a target audio segment that meets audio quality and duration requirements from the multiple audio segments;

[0059] 205. Input the target audio clip and the text introduction information of the target service personnel into a deep learning-based voice cloning model to perform voice cloning to obtain audio introduction information corresponding to the text introduction information;

[0060] 206. Add the audio introduction information to the details page of the target service personnel to obtain the audio resume page of the target service personnel.

[0061] In an optional embodiment, target service personnel who meet the conditions for generating an audio resume are determined from among the service personnel who have signed contracts with the target service platform, including: when the target service platform has a newly signed service personnel, determining whether the recruitment method of the newly signed service personnel is a direct recruitment method, and determining whether the agent has had a conversation with the newly signed service personnel; if both are yes, the newly signed service personnel is used as the target service personnel; and / or selecting candidate service personnel for whom an audio resume page does not exist from among the service personnel who have signed contracts with the target service platform, determining whether the recruitment method of the candidate service personnel is a direct recruitment method, and determining whether the agent has had a conversation with the candidate service personnel; if both are yes, the candidate service personnel is used as the target service personnel.

[0062] In an optional embodiment, multiple audio segments are extracted from the target audio information, including: performing noise reduction processing on the target audio information, and identifying and removing silent segments from the target audio information after noise reduction to obtain audio information to be segmented; according to the continuity and semantic integrity of the audio information to be segmented, the audio information to be segmented is segmented into multiple audio segments, and the pause intervals in the same audio segment are all less than the set interval threshold and express complete semantic information.

[0063] In an optional embodiment, a target audio segment that meets audio quality and duration requirements is selected from multiple audio segments, including: performing duration statistics on the multiple audio segments to obtain the durations of the multiple audio segments, and selecting candidate audio segments that meet the duration requirements; performing subjective quality analysis and / or objective quality analysis on the candidate audio segments to obtain subjective quality scores and / or objective quality scores of the candidate audio segments; and selecting, from the candidate audio segments, a candidate audio segment whose subjective quality score and / or objective quality score is greater than a corresponding score threshold as a target audio segment.

[0064] In an optional embodiment, the sound cloning model includes: a preprocessing network, a feature extraction network and a sound synthesis network; the target audio clip and the text introduction information of the target service personnel are input into the sound cloning model based on deep learning for sound cloning to obtain the audio introduction information corresponding to the text introduction information, including: inputting the target audio clip into the preprocessing network in the sound cloning model based on deep learning, and performing noise reduction processing on the target audio clip to obtain an intermediate audio clip; inputting the intermediate audio clip into the feature extraction network for sound feature extraction to obtain spectrum feature information, the spectrum feature information is a spectrum representation of the intermediate audio clip, including at least one of the frequency, amplitude and phase information of the intermediate audio clip; inputting the spectrum feature information and the text introduction information into the sound synthesis network, and performing sound synthesis on the text introduction information according to the spectrum feature information to obtain the audio introduction information.

[0065] In an optional embodiment, audio introduction information is added to the details page of the target service personnel to obtain the audio resume page of the target service personnel, including: adding the storage address of the audio introduction information in the audio database to the details page of the target service personnel; when the user initiates an access request to the details page, according to the storage address, the audio introduction information is added to the details page to obtain the audio resume page, and sent to the user's terminal device, so that the terminal device can display the audio resume page to the user.

[0066] In an optional embodiment, the storage address of the audio introduction information in the audio database is added to the details page of the target service personnel, including: determining the location information of the target card area from the details page of the target service personnel, the target card area is the top area of ​​the details page or the area for introducing the identity of the target service personnel; according to the location information of the target card area, binding the storage address of the audio introduction information to the associated area of ​​specific information in the target card area, so as to display the audio introduction information in the associated area.

[0067] Figure 3 The following is a flow chart of a page access method provided by an exemplary embodiment of the present application. Figure 3 As shown, the method is applied to a terminal device, the terminal device runs a target application, and includes:

[0068] 301. Displaying a service personnel list page provided by the target application, where the service personnel list page includes introduction information of multiple service personnel;

[0069] 302. In response to a triggering operation on the introduction information of any service personnel, sending an access request for a detail page of any service personnel to a target service platform;

[0070] 303. When receiving the audio resume page of any service personnel returned by the target service platform, jump from the service personnel list page to the audio resume page, and display the audio introduction information on the audio resume page;

[0071] 304. When receiving the detail page of any service personnel returned by the target service platform, jump from the service personnel list page to the detail page;

[0072] Among them, if the target service platform includes the storage address of the audio introduction information in the details page of any service personnel, the target service platform adds the audio introduction information to the details page to obtain an audio resume page and returns it to the terminal device; or, if the target service platform does not include the storage address of the audio introduction information in the details page of any service personnel, the target service platform returns the details page to the terminal device.

[0073] The detailed implementation methods and beneficial effects of each step provided in the embodiments of the present application have been described in detail in the aforementioned embodiments and will not be elaborated here.

[0074] Figure 4 The following is a schematic diagram of a structure of an audio resume generating device provided by an exemplary embodiment of the present application. Figure 4 As shown, the device comprises:

[0075] A determination module 41 is used to determine target service personnel who meet the conditions for generating audio resumes from the service personnel who have signed contracts with the target service platform;

[0076] The first acquisition module 42 is used to acquire the call record information between the agent of the target service platform and the target service personnel from the call record database according to the identification information of the target service personnel, wherein the two channels of the target call device correspond to the agent and the target service personnel respectively;

[0077] The second acquisition module 43 is used to parse the call record information into audio information on the two channels when the call record information includes two channels, and acquire the audio information on the channel corresponding to the target service personnel as the target audio information;

[0078] A selection module 44 is used to extract multiple audio segments from the target audio information, and select a target audio segment that meets the audio quality and duration requirements from the multiple audio segments;

[0079] A cloning module 45 is used to input the target audio clip and the text introduction information of the target service personnel into a deep learning-based voice cloning model to perform voice cloning to obtain audio introduction information corresponding to the text introduction information;

[0080] The adding module 46 is used to add the audio introduction information to the details page of the target service personnel to obtain the audio resume page of the target service personnel.

[0081] In an optional embodiment, target service personnel who meet the conditions for generating an audio resume are determined from among the service personnel who have signed contracts with the target service platform, including: when the target service platform has a newly signed service personnel, determining whether the recruitment method of the newly signed service personnel is a direct recruitment method, and determining whether the agent has had a conversation with the newly signed service personnel; if both are yes, the newly signed service personnel is used as the target service personnel; and / or selecting candidate service personnel for whom an audio resume page does not exist from among the service personnel who have signed contracts with the target service platform, determining whether the recruitment method of the candidate service personnel is a direct recruitment method, and determining whether the agent has had a conversation with the candidate service personnel; if both are yes, the candidate service personnel is used as the target service personnel.

[0082] In an optional embodiment, when the selection module extracts multiple audio segments from the target audio information, it is specifically used to: perform noise reduction processing on the target audio information, and identify and remove silent segments from the target audio information after noise reduction to obtain the audio information to be segmented; according to the continuity and semantic integrity of the audio information to be segmented, the audio information to be segmented is segmented into multiple audio segments, and the pause intervals in the same audio segment are all less than the set interval threshold and express complete semantic information.

[0083] In an optional embodiment, when the selection module selects a target audio segment that meets the audio quality and duration requirements from multiple audio segments, the selection module is specifically used to: perform duration statistics on the multiple audio segments to obtain the durations of the multiple audio segments, and select candidate audio segments that meet the duration requirements; perform subjective quality analysis and / or objective quality analysis on the candidate audio segments to obtain the subjective quality score and / or objective quality score of the candidate audio segments; and select, from the candidate audio segments, a candidate audio segment whose subjective quality score and / or objective quality score is greater than the corresponding score threshold as the target audio segment.

[0084] In an optional embodiment, the sound cloning model includes: a preprocessing network, a feature extraction network and a sound synthesis network; when the cloning module inputs the target audio clip and the text introduction information of the target service personnel into the sound cloning model based on deep learning for sound cloning to obtain the audio introduction information corresponding to the text introduction information, it is specifically used to: input the target audio clip into the preprocessing network in the sound cloning model based on deep learning, and perform noise reduction on the target audio clip to obtain an intermediate audio clip; input the intermediate audio clip into the feature extraction network to perform sound feature extraction to obtain spectrum feature information, the spectrum feature information is a spectrum representation of the intermediate audio clip, including at least one of the frequency, amplitude and phase information of the intermediate audio clip; input the spectrum feature information and the text introduction information into the sound synthesis network, and perform sound synthesis on the text introduction information according to the spectrum feature information to obtain the audio introduction information.

[0085] In an optional embodiment, when the adding module adds the audio introduction information to the details page of the target service personnel to obtain the audio resume page of the target service personnel, it is specifically used to: add the storage address of the audio introduction information in the audio database to the details page of the target service personnel; when the user initiates an access request to the details page, according to the storage address, add the audio introduction information to the details page to obtain the audio resume page, and send it to the user's terminal device, so that the terminal device can display the audio resume page to the user.

[0086] Optionally, when the adding module adds the storage address of the audio introduction information in the audio database to the details page of the target service personnel, it is specifically used to: determine the location information of the target card area from the details page of the target service personnel, the target card area is the top area of ​​the details page or the area for introducing the identity of the target service personnel; according to the location information of the target card area, bind the storage address of the audio introduction information to the associated area of ​​specific information in the target card area, so as to display the audio introduction information in the associated area.

[0087] The detailed implementation methods and beneficial effects of each step provided in the embodiments of the present application have been described in detail in the aforementioned embodiments and will not be elaborated here.

[0088] Figure 5 The following is a schematic diagram of a page access device provided by an exemplary embodiment of the present application. Figure 5 As shown, the device comprises:

[0089] A display module 51 is used to display a service personnel list page provided by the target application, where the service personnel list page includes introduction information of multiple service personnel;

[0090] The sending module 52 is used to send an access request for a detail page of any service personnel to the target service platform in response to a triggering operation on the introduction information of any service personnel;

[0091] The display module 53 is used to jump from the service personnel list page to the audio resume page when receiving the audio resume page of any service personnel returned by the target service platform, and display the audio introduction information on the audio resume page;

[0092] The jump module 54 is used to jump from the service personnel list page to the detail page when receiving the detail page of any service personnel returned by the target service platform;

[0093] Among them, if the target service platform includes the storage address of the audio introduction information in the details page of any service personnel, the target service platform adds the audio introduction information to the details page to obtain an audio resume page and returns it to the terminal device; or, if the target service platform does not include the storage address of the audio introduction information in the details page of any service personnel, the target service platform returns the details page to the terminal device.

[0094] The detailed implementation methods and beneficial effects of each step provided in the embodiments of the present application have been described in detail in the aforementioned embodiments and will not be elaborated here.

[0095] Figure 6 The following is a schematic diagram of the structure of an electronic device provided by an exemplary embodiment of the present application. Figure 6 As shown, it includes: a memory 60a and a processor 60b; the memory 60a is used to store computer programs; the processor 60b is coupled to the memory 60a and is used to execute the computer programs to implement the steps in the above-mentioned method embodiments.

[0096] Furthermore, if Figure 6 As shown, the electronic device also includes: a communication component 60c, a display 60d, a power component 60e, an audio component 60f and other components. Figure 6 Only some components are shown schematically, which does not mean that the electronic device only includes Figure 6 Components shown.

[0097] The detailed implementation and beneficial effects of the electronic device provided in the embodiments of the present application have been described in detail in the aforementioned embodiments and will not be elaborated here.

[0098] An exemplary embodiment of the present application further provides a computer-readable storage medium storing a computer program, wherein when the computer program is executed by a processor, the processor is caused to implement the steps in the above-mentioned method embodiments.

[0099] An exemplary embodiment of the present application further provides a computer program product, which includes a computer program / instructions. When the computer program / instructions are executed by a processor, the processor is enabled to implement the steps in the above-mentioned method embodiments.

[0100] The above-mentioned memory can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, magnetic disk or optical disk.

[0101] The above-mentioned communication component is configured to facilitate wired or wireless communication between the device where the communication component is located and other devices. The device where the communication component is located can access a wireless network based on a communication standard, such as WiFi, 2G, 3G, 4G / LTE, 5G and other mobile communication networks, or a combination thereof. In an exemplary embodiment, the communication component receives a broadcast signal or broadcast-related information from an external broadcast management system via a broadcast channel. In an exemplary embodiment, the communication component also includes a near field communication (NFC) module to facilitate short-range communication. For example, the NFC module can be based on radio frequency identification (RFID) technology, infrared data association (IrDA) technology, ultra-wideband (Ultra Wide Band, UWB) technology, Bluetooth (BlueTooth, BT) technology and other technologies.

[0102] The above-mentioned display includes a screen, and the screen may include a liquid crystal display (LCD) and a touch panel (TP). If the screen includes a touch panel, the screen may be implemented as a touch screen to receive input signals from a user. The touch panel includes one or more touch sensors to sense touch, slide, and gestures on the touch panel. The touch sensor may not only sense the boundary of a touch or slide action, but also detect the duration and pressure associated with the touch or slide operation.

[0103] The power supply assembly provides power to various components of the device where the power supply assembly is located. The power supply assembly may include a power management system, one or more power supplies, and other components associated with generating, managing, and distributing power to the device where the power supply assembly is located.

[0104] The above-mentioned audio component can be configured to output and / or input audio signals. For example, the audio component includes a microphone (Microphone, MIC), and when the device where the audio component is located is in an operating mode, such as a call mode, a recording mode, and a speech recognition mode, the microphone is configured to receive an external audio signal. The received audio signal can be further stored in a memory or sent via a communication component. In some embodiments, the audio component also includes a speaker for outputting an audio signal.

[0105] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented on one or more computer-readable storage media (including but not limited to disk storage, compact disk read-only memory (Compact Disk Read-Only Memory, CD-ROM), optical storage, etc.) that contain computer-usable program code.

[0106] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0107] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0108] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0109] In a typical configuration, a computing device includes one or more processors (Central Processing Unit, CPU), input / output interface, network interface and memory.

[0110] The memory may include non-permanent memory in a computer-readable medium, random access memory (RAM) and / or non-volatile memory in the form of read-only memory (ROM) or flash memory (flashRAM). The memory is an example of a computer-readable medium.

[0111] Computer-readable media include permanent and non-permanent, removable and non-removable media that can be used to store information by any method or technology. Information can be computer-readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase-change random access memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, read-only compact disc read-only memory (CD-ROM), digital versatile disc (DVD) or other optical storage, magnetic cassettes, tape disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer-readable media does not include transitory media such as modulated data signals and carrier waves.

[0112] It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, commodity or device. In the absence of more restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the process, method, commodity or device including the elements.

[0113] The above are only embodiments of the present application and are not intended to limit the present application. For those skilled in the art, the present application may have various changes and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included within the scope of the claims of the present application.

Claims

1. A method for generating an audio resume, characterized in that: Applied to the target service platform, the method comprises: Determine target service personnel who meet the conditions for generating audio resumes from among the service personnel who have signed contracts with the target service platform; Acquire call record information between a broker of a target service platform and the target service personnel from a call record database according to the identification information of the target service personnel; In the case where the call record information includes two channels, parsing the call record information into audio information on the two channels, and acquiring the audio information on the channel corresponding to the target service personnel as the target audio information; Extracting multiple audio segments from the target audio information, and selecting a target audio segment that meets audio quality and duration requirements from the multiple audio segments; Inputting the target audio clip and the text introduction information of the target service personnel into a deep learning-based voice cloning model to perform voice cloning, so as to obtain audio introduction information corresponding to the text introduction information; The audio introduction information is added to the details page of the target service personnel to obtain the audio resume page of the target service personnel.

2. The method according to claim 1, characterized in that From the service personnel who have signed contracts with the target service platform, determine the target service personnel who meet the conditions for generating audio resumes, including: When the target service platform has a newly signed service personnel, determine whether the newly signed service personnel is recruited by direct recruitment, and determine whether the broker has had a call with the newly signed service personnel; if both are yes, the newly signed service personnel is used as the target service personnel; and / or From the service personnel who have signed contracts with the target service platform, select candidate service personnel who do not have an audio resume page, determine whether the candidate service personnel is recruited by direct recruitment, and determine whether the agent has had a conversation with the candidate service personnel; if both are true, the candidate service personnel will be used as the target service personnel.

3. The method according to claim 1, characterized in that Extracting multiple audio segments from the target audio information includes: Performing noise reduction processing on the target audio information, and identifying and removing silent segments from the target audio information after noise reduction to obtain audio information to be segmented; According to the continuity and semantic integrity of the audio information to be segmented, the audio information to be segmented is segmented into multiple audio segments, and the pause intervals in the same audio segment are all smaller than a set interval threshold and express complete semantic information.

4. The method according to claim 1, characterized in that: Selecting a target audio segment that meets audio quality and duration requirements from the multiple audio segments includes: Performing duration statistics on the multiple audio segments to obtain the durations of the multiple audio segments, and selecting candidate audio segments that meet the duration requirement; Performing subjective quality analysis and / or objective quality analysis on the candidate audio segment to obtain a subjective quality score and / or objective quality score of the candidate audio segment; From the candidate audio segments, a candidate audio segment having a subjective quality score and / or an objective quality score greater than a corresponding score threshold is selected as the target audio segment.

5. The method according to any one of claims 1 to 4, characterized in that: The sound cloning model includes: a preprocessing network, a feature extraction network and a sound synthesis network; the target audio clip and the text introduction information of the target service personnel are input into the sound cloning model based on deep learning for sound cloning, so as to obtain the audio introduction information corresponding to the text introduction information, including: Inputting the target audio segment into a preprocessing network in a deep learning-based sound cloning model, and performing noise reduction processing on the target audio segment to obtain an intermediate audio segment; Inputting the intermediate audio segment into the feature extraction network to extract sound features to obtain spectrum feature information, where the spectrum feature information is a spectrum representation of the intermediate audio segment, including at least one of frequency, amplitude and phase information of the intermediate audio segment; The frequency spectrum feature information and the text introduction information are input into the sound synthesis network, and the text introduction information is sound synthesized according to the frequency spectrum feature information to obtain the audio introduction information.

6. The method according to any one of claims 1 to 4, characterized in that: Adding the audio introduction information to the details page of the target service personnel to obtain the audio resume page of the target service personnel includes: Add the storage address of the audio introduction information in the audio database to the details page of the target service personnel; When the user initiates an access request to the details page, the audio introduction information is added to the details page according to the storage address to obtain the audio resume page, and sent to the user's terminal device so that the terminal device can display the audio resume page to the user.

7. The method according to claim 6, characterized in that Adding the storage address of the audio introduction information in the audio database to the details page of the target service personnel includes: Determine the location information of a target card area from the target service personnel's details page, where the target card area is the top area of ​​the details page or the area for introducing the identity of the target service personnel; According to the position information of the target card area, the storage address of the audio introduction information is bound to the associated area of ​​the specific information in the target card area, so as to display the audio introduction information in the associated area.

8. A page access method, characterized in that: Applied to a terminal device, the terminal device runs a target application, and the method includes: Displaying a service personnel list page provided by the target application, wherein the service personnel list page includes introduction information of multiple service personnel; In response to a triggering operation on the introduction information of any service personnel, sending a request for accessing the detail page of any service personnel to the target service platform; When receiving the audio resume page of any service personnel returned by the target service platform, jump from the service personnel list page to the audio resume page, and display the audio introduction information on the audio resume page; When receiving the detail page of any service personnel returned by the target service platform, jump from the service personnel list page to the detail page; In which, when the storage address of the audio introduction information is included in the details page of any service personnel, the target service platform adds the audio introduction information to the details page to obtain the audio resume page and returns it to the terminal device; or, when the storage address of the audio introduction information is not included in the details page of any service personnel, the target service platform returns the details page to the terminal device.

9. An electronic device, characterized in that: include: Memory and processor; The memory is used to store a computer program; the processor is coupled to the memory and is used to execute the computer program to implement the steps in the method according to any one of claims 1 to 7 or 8.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the processor is caused to implement the steps in the method according to any one of claims 1 to 7 or 8.

11. A computer program product, characterized in that The computer program product comprises a computer program / instruction, which, when executed by a processor, enables the processor to implement the steps of any one of the methods of claims 1 to 7 or 8.

Images