KR20240040577A - Method of adjusting sensitivity for spearker verification and electronic device therefor - Google Patents

Abstract

An electronic device including a display, memory, and processor is disclosed. The processor may obtain a voice signal including a wakeup word and attempt speaker verification based on the voice signal. If user authentication is received within a specified time after speaker verification fails, the processor can adjust the threshold for speaker verification using the voice signal that failed speaker verification.

Description

Sensitivity adjustment method for speaker verification and electronic device therefor {METHOD OF ADJUSTING SENSITIVITY FOR SPEARKER VERIFICATION AND ELECTRONIC DEVICE THEREFOR}

Embodiments disclosed in this document relate to a method for adjusting sensitivity for speaker verification and an electronic device for the same.

Technologies that control electronic devices based on user voice commands are widely used. For example, the electronic device may include a voice assistant application configured to identify the user's intent from the user's utterance and perform an action corresponding to the identified intent. Users can easily control electronic devices through voice commands. For example, a user can call a voice assistant application by uttering a wake-up word to call the voice assistant application.

In one example, the electronic device can verify the speaker based on voice commands. The electronic device may store a voice model corresponding to at least one user's voice. The voice model may include a user-specific model learned using the user's voice. The electronic device may verify the speaker of the voice command based on similarity between the received voice command and the stored voice model. For example, the electronic device can verify the speaker of the voice command to unlock the lock mode. The electronic device can verify the speaker to perform an operation corresponding to the voice command. The electronic device may verify the speaker of the voice command to identify the corresponding speaker among the plurality of stored voice models.

The electronic device may store a user's voice model. The voice model may be a vector corresponding to the speaker's voice. In the enrollment process, the electronic device can create a voice model using the user's voice samples. Once the voice model is created, the electronic device can calculate a score based on similarity using the received voice and the voice model. If the score exceeds the set threshold, the electronic device can verify that the received voice corresponds to the voice of the speaker in the voice model.

A user's voice may change temporarily. For example, if a user catches a cold, the user's voice may temporarily change. In this case, the electronic device may fail speaker verification based on the user's voice. The stored voice model and the user's voice may have a similarity that is less than a set threshold. Typically, the threshold is set by the manufacturer of the electronic device, so the user experience may be degraded due to repeated speaker authentication failures.

An electronic device according to an embodiment disclosed in this document may include a display, memory, and a processor. When executed by a processor, the memory can enable the processor to perform the following operations. The processor may acquire a plurality of voice signals including a wakeup word. The processor may determine whether to verify the speaker for each of the plurality of voice signals by comparing the similarity between each of the plurality of voice signals and the speaker verification model stored in the memory with a first threshold. The processor may receive user authentication for the electronic device based on an authentication method other than speaker verification within a specified time from the time of reception of the last voice signal among the plurality of voice signals, after speaker verification of the plurality of voice signals fails. . The processor may adjust the first threshold based on the similarity between at least some of the plurality of voice signals and the speaker verification model.

A method for adjusting speaker verification sensitivity of an electronic device according to an embodiment disclosed in this document may include an operation of acquiring voice data including a wakeup word. The method may include performing speaker verification on the speech data by comparing a similarity between the speech data and a speaker verification model stored in a memory of the electronic device to a first threshold. The method may include, if speaker verification fails, determining whether successful user authentication is received within a specified time from the time of acquisition of the voice data. The method may include adjusting the first threshold using a plurality of voice data stored in memory when successful user authentication is received. Each of the plurality of voice data includes the wakeup word and may correspond to voice data that failed the speaker verification.

An electronic device according to an embodiment of the present disclosure can adaptively adjust the threshold for speaker authentication.

An electronic device according to an embodiment of the present disclosure can increase user convenience.

The electronic device according to an example of the present disclosure can increase the frequency of use of the electronic device by improving user convenience.

1 is a block diagram of an electronic device in a network environment according to various embodiments.
Figure 2 is a block diagram showing an integrated intelligence system according to an embodiment.
Figure 3 is a diagram showing how relationship information between concepts and actions is stored in a database, according to an embodiment.
FIG. 4 is a diagram illustrating a user terminal displaying a screen for processing voice input received through an intelligent app, according to one embodiment.
Figure 5 shows a voice recognition environment according to one embodiment.
Figure 6 shows a block diagram of an electronic device according to an embodiment.
Figure 7 shows modules of an electronic device according to an embodiment.
Figure 8 shows a flowchart of a sensitivity adjustment method according to one embodiment.
Figure 9 shows a flowchart of a sensitivity adjustment method according to one embodiment.
10 illustrates an example UI of an electronic device according to an embodiment.
11 illustrates an example UI of an electronic device according to an embodiment.
12 illustrates an example UI of an electronic device according to an embodiment.

Hereinafter, various embodiments of the present invention are described with reference to the accompanying drawings. However, this is not intended to limit the present invention to specific embodiments, and should be understood to include various modifications, equivalents, and/or alternatives to the embodiments of the present invention.

1 is a block diagram of an electronic device 101 in a network environment 100, according to various embodiments. Referring to FIG. 1, in the network environment 100, the electronic device 101 communicates with the electronic device 102 through a first network 198 (e.g., a short-range wireless communication network) or a second network 199. It is possible to communicate with at least one of the electronic device 104 or the server 108 through (e.g., a long-distance wireless communication network). According to one embodiment, the electronic device 101 may communicate with the electronic device 104 through the server 108. According to one embodiment, the electronic device 101 includes a processor 120, a memory 130, an input module 150, an audio output module 155, a display module 160, an audio module 170, and a sensor module ( 176), interface 177, connection terminal 178, haptic module 179, camera module 180, power management module 188, battery 189, communication module 190, subscriber identification module 196 , or may include an antenna module 197. In some embodiments, at least one of these components (eg, the connection terminal 178) may be omitted or one or more other components may be added to the electronic device 101. In some embodiments, some of these components (e.g., sensor module 176, camera module 180, or antenna module 197) are integrated into one component (e.g., display module 160). It can be.

The processor 120, for example, executes software (e.g., program 140) to operate at least one other component (e.g., hardware or software component) of the electronic device 101 connected to the processor 120. It can be controlled and various data processing or calculations can be performed. According to one embodiment, as at least part of data processing or computation, the processor 120 stores commands or data received from another component (e.g., sensor module 176 or communication module 190) in volatile memory 132. The commands or data stored in the volatile memory 132 can be processed, and the resulting data can be stored in the non-volatile memory 134. According to one embodiment, the processor 120 includes a main processor 121 (e.g., a central processing unit or an application processor) or an auxiliary processor 123 that can operate independently or together (e.g., a graphics processing unit, a neural network processing unit ( It may include a neural processing unit (NPU), an image signal processor, a sensor hub processor, or a communication processor). For example, if the electronic device 101 includes a main processor 121 and a secondary processor 123, the secondary processor 123 may be set to use lower power than the main processor 121 or be specialized for a designated function. You can. The auxiliary processor 123 may be implemented separately from the main processor 121 or as part of it.

The auxiliary processor 123 may, for example, act on behalf of the main processor 121 while the main processor 121 is in an inactive (e.g., sleep) state, or while the main processor 121 is in an active (e.g., application execution) state. ), together with the main processor 121, at least one of the components of the electronic device 101 (e.g., the display module 160, the sensor module 176, or the communication module 190) At least some of the functions or states related to can be controlled. According to one embodiment, co-processor 123 (e.g., image signal processor or communication processor) may be implemented as part of another functionally related component (e.g., camera module 180 or communication module 190). there is. According to one embodiment, the auxiliary processor 123 (eg, neural network processing unit) may include a hardware structure specialized for processing artificial intelligence models. Artificial intelligence models can be created through machine learning. For example, such learning may be performed in the electronic device 101 itself on which the artificial intelligence model is performed, or may be performed through a separate server (e.g., server 108). Learning algorithms may include, for example, supervised learning, unsupervised learning, semi-supervised learning, or reinforcement learning, but It is not limited. An artificial intelligence model may include multiple artificial neural network layers. Artificial neural networks include deep neural network (DNN), convolutional neural network (CNN), recurrent neural network (RNN), restricted boltzmann machine (RBM), belief deep network (DBN), bidirectional recurrent deep neural network (BRDNN), It may be one of deep Q-networks or a combination of two or more of the above, but is not limited to the examples described above. In addition to hardware structures, artificial intelligence models may additionally or alternatively include software structures.

The memory 130 may store various data used by at least one component (eg, the processor 120 or the sensor module 176) of the electronic device 101. Data may include, for example, input data or output data for software (e.g., program 140) and instructions related thereto. Memory 130 may include volatile memory 132 or non-volatile memory 134.

The program 140 may be stored as software in the memory 130 and may include, for example, an operating system 142, middleware 144, or application 146.

The input module 150 may receive commands or data to be used in a component of the electronic device 101 (e.g., the processor 120) from outside the electronic device 101 (e.g., a user). The input module 150 may include, for example, a microphone, mouse, keyboard, keys (eg, buttons), or digital pen (eg, stylus pen).

The sound output module 155 may output sound signals to the outside of the electronic device 101. The sound output module 155 may include, for example, a speaker or a receiver. Speakers can be used for general purposes such as multimedia playback or recording playback. The receiver can be used to receive incoming calls. According to one embodiment, the receiver may be implemented separately from the speaker or as part of it.

The display module 160 can visually provide information to the outside of the electronic device 101 (eg, a user). The display module 160 may include, for example, a display, a hologram device, or a projector, and a control circuit for controlling the device. According to one embodiment, the display module 160 may include a touch sensor configured to detect a touch, or a pressure sensor configured to measure the intensity of force generated by the touch.

The audio module 170 can convert sound into an electrical signal or, conversely, convert an electrical signal into sound. According to one embodiment, the audio module 170 acquires sound through the input module 150, the sound output module 155, or an external electronic device (e.g., directly or wirelessly connected to the electronic device 101). Sound may be output through the electronic device 102 (e.g., speaker or headphone).

The sensor module 176 detects the operating state (e.g., power or temperature) of the electronic device 101 or the external environmental state (e.g., user state) and generates an electrical signal or data value corresponding to the detected state. can do. According to one embodiment, the sensor module 176 includes, for example, a gesture sensor, a gyro sensor, an air pressure sensor, a magnetic sensor, an acceleration sensor, a grip sensor, a proximity sensor, a color sensor, an IR (infrared) sensor, a biometric sensor, It may include a temperature sensor, humidity sensor, or light sensor.

The interface 177 may support one or more designated protocols that can be used to connect the electronic device 101 directly or wirelessly with an external electronic device (eg, the electronic device 102). According to one embodiment, the interface 177 may include, for example, a high definition multimedia interface (HDMI), a universal serial bus (USB) interface, an SD card interface, or an audio interface.

The connection terminal 178 may include a connector through which the electronic device 101 can be physically connected to an external electronic device (eg, the electronic device 102). According to one embodiment, the connection terminal 178 may include, for example, an HDMI connector, a USB connector, an SD card connector, or an audio connector (eg, a headphone connector).

The haptic module 179 can convert electrical signals into mechanical stimulation (e.g., vibration or movement) or electrical stimulation that the user can perceive through tactile or kinesthetic senses. According to one embodiment, the haptic module 179 may include, for example, a motor, a piezoelectric element, or an electrical stimulation device.

The camera module 180 can capture still images and moving images. According to one embodiment, the camera module 180 may include one or more lenses, image sensors, image signal processors, or flashes.

The power management module 188 can manage power supplied to the electronic device 101. According to one embodiment, the power management module 188 may be implemented as at least a part of, for example, a power management integrated circuit (PMIC).

The battery 189 may supply power to at least one component of the electronic device 101. According to one embodiment, the battery 189 may include, for example, a non-rechargeable primary battery, a rechargeable secondary battery, or a fuel cell.

Communication module 190 is configured to provide a direct (e.g., wired) communication channel or wireless communication channel between electronic device 101 and an external electronic device (e.g., electronic device 102, electronic device 104, or server 108). It can support establishment and communication through established communication channels. Communication module 190 operates independently of processor 120 (e.g., an application processor) and may include one or more communication processors that support direct (e.g., wired) communication or wireless communication. According to one embodiment, the communication module 190 is a wireless communication module 192 (e.g., a cellular communication module, a short-range wireless communication module, or a global navigation satellite system (GNSS) communication module) or a wired communication module 194 (e.g., : LAN (local area network) communication module, or power line communication module) may be included. Among these communication modules, the corresponding communication module is a first network 198 (e.g., a short-range communication network such as Bluetooth, wireless fidelity (WiFi) direct, or infrared data association (IrDA)) or a second network 199 (e.g., legacy It may communicate with an external electronic device 104 through a telecommunication network such as a cellular network, a 5G network, a next-generation communication network, the Internet, or a computer network (e.g., LAN or WAN). These various types of communication modules may be integrated into one component (e.g., a single chip) or may be implemented as a plurality of separate components (e.g., multiple chips). The wireless communication module 192 uses subscriber information (e.g., International Mobile Subscriber Identifier (IMSI)) stored in the subscriber identification module 196 within a communication network such as the first network 198 or the second network 199. The electronic device 101 can be confirmed or authenticated.

The wireless communication module 192 may support 5G networks after 4G networks and next-generation communication technologies, for example, NR access technology (new radio access technology). NR access technology provides high-speed transmission of high-capacity data (eMBB (enhanced mobile broadband)), minimization of terminal power and access to multiple terminals (mMTC (massive machine type communications)), or high reliability and low latency (URLLC (ultra-reliable and low latency). -latency communications)) can be supported. The wireless communication module 192 may support high frequency bands (eg, mmWave bands), for example, to achieve high data rates. The wireless communication module 192 uses various technologies to secure performance in high frequency bands, for example, beamforming, massive array multiple-input and multiple-output (MIMO), and full-dimensional multiplexing. It can support technologies such as input/output (FD-MIMO: full dimensional MIMO), array antenna, analog beam-forming, or large scale antenna. The wireless communication module 192 may support various requirements specified in the electronic device 101, an external electronic device (e.g., electronic device 104), or a network system (e.g., second network 199). According to one embodiment, the wireless communication module 192 supports Peak data rate (e.g., 20 Gbps or more) for realizing eMBB, loss coverage (e.g., 164 dB or less) for realizing mmTC, or U-plane latency (e.g., 164 dB or less) for realizing URLLC. Example: Downlink (DL) and uplink (UL) each of 0.5 ms or less, or round trip 1 ms or less) can be supported.

The antenna module 197 may transmit or receive signals or power to or from the outside (eg, an external electronic device). According to one embodiment, the antenna module 197 may include an antenna including a radiator made of a conductor or a conductive pattern formed on a substrate (eg, PCB). According to one embodiment, the antenna module 197 may include a plurality of antennas (eg, an array antenna). In this case, at least one antenna suitable for a communication method used in a communication network such as the first network 198 or the second network 199 is connected to the plurality of antennas by, for example, the communication module 190. can be selected. Signals or power may be transmitted or received between the communication module 190 and an external electronic device through the at least one selected antenna. According to some embodiments, in addition to the radiator, other components (eg, radio frequency integrated circuit (RFIC)) may be additionally formed as part of the antenna module 197.

According to various embodiments, the antenna module 197 may form a mmWave antenna module. According to one embodiment, a mmWave antenna module includes: a printed circuit board, an RFIC disposed on or adjacent to a first side (e.g., bottom side) of the printed circuit board and capable of supporting a designated high frequency band (e.g., mmWave band); And a plurality of antennas (e.g., array antennas) disposed on or adjacent to the second side (e.g., top or side) of the printed circuit board and capable of transmitting or receiving signals in the designated high frequency band. can do.

At least some of the components are connected to each other through a communication method between peripheral devices (e.g., bus, general purpose input and output (GPIO), serial peripheral interface (SPI), or mobile industry processor interface (MIPI)) and signal ( (e.g. commands or data) can be exchanged with each other.

According to one embodiment, commands or data may be transmitted or received between the electronic device 101 and the external electronic device 104 through the server 108 connected to the second network 199. Each of the external electronic devices 102 or 104 may be of the same or different type as the electronic device 101. According to one embodiment, all or part of the operations performed in the electronic device 101 may be executed in one or more of the external electronic devices 102, 104, or 108. For example, when the electronic device 101 needs to perform a certain function or service automatically or in response to a request from a user or another device, the electronic device 101 may perform the function or service instead of executing the function or service on its own. Alternatively, or additionally, one or more external electronic devices may be requested to perform at least part of the function or service. One or more external electronic devices that have received the request may execute at least part of the requested function or service, or an additional function or service related to the request, and transmit the result of the execution to the electronic device 101. The electronic device 101 may process the result as is or additionally and provide it as at least part of a response to the request. For this purpose, for example, cloud computing, distributed computing, mobile edge computing (MEC), or client-server computing technology can be used. The electronic device 101 may provide an ultra-low latency service using, for example, distributed computing or mobile edge computing. In another embodiment, the external electronic device 104 may include an Internet of Things (IoT) device. Server 108 may be an intelligent server using machine learning and/or neural networks. According to one embodiment, the external electronic device 104 or server 108 may be included in the second network 199. The electronic device 101 may be applied to intelligent services (e.g., smart home, smart city, smart car, or healthcare) based on 5G communication technology and IoT-related technology.

Electronic devices according to various embodiments disclosed in this document may be of various types. Electronic devices may include, for example, portable communication devices (e.g., smartphones), computer devices, portable multimedia devices, portable medical devices, cameras, wearable devices, or home appliances. Electronic devices according to embodiments of this document are not limited to the above-described devices.

The various embodiments of this document and the terms used herein are not intended to limit the technical features described in this document to specific embodiments, and should be understood to include various changes, equivalents, or replacements of the embodiments. In connection with the description of the drawings, similar reference numbers may be used for similar or related components. The singular form of a noun corresponding to an item may include one or more of the above items, unless the relevant context clearly indicates otherwise. As used herein, “A or B”, “at least one of A and B”, “at least one of A or B”, “A, B or C”, “at least one of A, B and C”, and “A Each of phrases such as “at least one of , B, or C” may include any one of the items listed together in the corresponding phrase, or any possible combination thereof. Terms such as "first", "second", or "first" or "second" may be used simply to distinguish one component from another, and to refer to that component in other respects (e.g., importance or order) is not limited. One (e.g., first) component is said to be “coupled” or “connected” to another (e.g., second) component, with or without the terms “functionally” or “communicatively.” When mentioned, it means that any of the components can be connected to the other components directly (e.g. wired), wirelessly, or through a third component.

The term “module” used in various embodiments of this document may include a unit implemented in hardware, software, or firmware, and is interchangeable with terms such as logic, logic block, component, or circuit, for example. It can be used as A module may be an integrated part or a minimum unit of the parts or a part thereof that performs one or more functions. For example, according to one embodiment, the module may be implemented in the form of an application-specific integrated circuit (ASIC).

Various embodiments of the present document are one or more instructions stored in a storage medium (e.g., built-in memory 136 or external memory 138) that can be read by a machine (e.g., electronic device 101). It may be implemented as software (e.g., program 140) including these. For example, a processor (e.g., processor 120) of a device (e.g., electronic device 101) may call at least one command among one or more commands stored from a storage medium and execute it. This allows the device to be operated to perform at least one function according to the at least one instruction called. The one or more instructions may include code generated by a compiler or code that can be executed by an interpreter. A storage medium that can be read by a device may be provided in the form of a non-transitory storage medium. Here, 'non-transitory' only means that the storage medium is a tangible device and does not contain signals (e.g. electromagnetic waves), and this term refers to cases where data is semi-permanently stored in the storage medium. There is no distinction between temporary storage cases.

According to one embodiment, methods according to various embodiments disclosed in this document may be included and provided in a computer program product. Computer program products are commodities and can be traded between sellers and buyers. The computer program product may be distributed in the form of a machine-readable storage medium (e.g. compact disc read only memory (CD-ROM)) or through an application store (e.g. Play Store™) or on two user devices (e.g. It can be distributed (e.g. downloaded or uploaded) directly between smart phones) or online. In the case of online distribution, at least a portion of the computer program product may be at least temporarily stored or temporarily created in a machine-readable storage medium, such as the memory of a manufacturer's server, an application store's server, or a relay server.

According to various embodiments, each component (e.g., module or program) of the above-described components may include a single or plural entity, and some of the plurality of entities may be separately placed in other components. there is. According to various embodiments, one or more of the components or operations described above may be omitted, or one or more other components or operations may be added. Alternatively or additionally, multiple components (eg, modules or programs) may be integrated into a single component. In this case, the integrated component may perform one or more functions of each component of the plurality of components in the same or similar manner as those performed by the corresponding component of the plurality of components prior to the integration. . According to various embodiments, operations performed by a module, program, or other component may be executed sequentially, in parallel, iteratively, or heuristically, or one or more of the operations may be executed in a different order, or omitted. Alternatively, one or more other operations may be added.

Figure 2 is a block diagram showing an integrated intelligence system according to an embodiment.

Referring to FIG. 2, the integrated intelligence system of one embodiment may include a user terminal 201, an intelligent server 300, and a service server 400.

The user terminal 201 (e.g., the electronic device 101 in FIG. 1) of one embodiment may be a terminal device (or electronic device) capable of connecting to the Internet, for example, a mobile phone, a smartphone, or a personal digital assistant (PDA). It may be a digital assistant, a laptop computer, a television (TV), a white appliance, a wearable device, a head mounted device (HMD), or a smart speaker.

According to the illustrated embodiment, the user terminal 201 may include a communication interface 290, a microphone 270, a speaker 255, a display 260, a memory 230, and/or a processor 220. there is. The components listed above may be operatively or electrically connected to each other.

The communication interface 290 (e.g., the communication module 190 of FIG. 1) may be connected to an external device and configured to transmit and receive data. The microphone 270 (e.g., the audio module 170 of FIG. 1) may receive sound (e.g., a user's speech) and convert it into an electrical signal. The speaker 255 (e.g., the sound output module 155 of FIG. 1) may output an electrical signal as sound (e.g., voice). Display 260 (e.g., display module 160 of FIG. 1) may be configured to display images or video. The display 260 of one embodiment may also display a graphic user interface (GUI) of an app (or application program) being executed.

The memory 230 (eg, memory 130 in FIG. 1) in one embodiment may store a client module 231, a software development kit (SDK) 233, and a plurality of applications. The client module 231 and SDK 233 may form a framework (or solution program) for performing general functions. Additionally, the client module 231 or SDK 233 may configure a framework for processing voice input.

The plurality of applications (eg, 235a, 235b) may be programs for performing designated functions. According to one embodiment, the plurality of applications may include a first app 235a and/or a second app 235b. According to one embodiment, each of the plurality of applications may include a plurality of operations to perform a designated function. For example, the applications may include an alarm app, a messaging app, and/or a schedule app. According to one embodiment, a plurality of applications are executed by the processor 220 to sequentially execute at least some of the plurality of operations.

The processor 220 in one embodiment may control the overall operation of the user terminal 201. For example, the processor 220 may be electrically connected to the communication interface 290, microphone 270, speaker 255, and display 260 to perform designated operations. For example, processor 220 may include at least one processor.

The processor 220 of one embodiment may also execute a program stored in the memory 230 to perform a designated function. For example, the processor 220 may execute at least one of the client module 231 or the SDK 233 and perform the following operations for processing voice input. The processor 220 may control the operation of a plurality of applications through, for example, the SDK 233. The following operations described as operations of the client module 231 or SDK 233 may be operations performed by execution of the processor 220.

The client module 231 in one embodiment may receive voice input. For example, the client module 231 may receive a voice signal corresponding to a user utterance detected through the microphone 270. The client module 231 may transmit the received voice input (eg, voice signal) to the intelligent server 300. The client module 231 may transmit status information of the user terminal 201 to the intelligent server 300 along with the received voice input. The status information may be, for example, execution status information of an app.

The client module 231 of one embodiment may receive a result corresponding to the received voice input from the intelligent server 300. For example, if the intelligent server 300 can calculate a result corresponding to the received voice input, the client module 231 may receive a result corresponding to the received voice input. The client module 231 may display the received result on the display 260.

The client module 231 of one embodiment may receive a plan corresponding to the received voice input. The client module 231 can display the results of executing multiple operations of the app according to the plan on the display 260. For example, the client module 231 may sequentially display execution results of a plurality of operations on a display. For another example, the user terminal 201 may display only some results of executing a plurality of operations (eg, the result of the last operation) on the display.

According to one embodiment, the client module 231 may receive a request from the intelligent server 300 to obtain information necessary to calculate a result corresponding to the voice input. According to one embodiment, the client module 231 may transmit the necessary information to the intelligent server 300 in response to the request.

The client module 231 in one embodiment may transmit information as a result of executing a plurality of operations according to the plan to the intelligent server 300. The intelligent server 300 can use the result information to confirm that the received voice input has been processed correctly.

The client module 231 in one embodiment may include a voice recognition module. According to one embodiment, the client module 231 may recognize voice input performing a limited function through the voice recognition module. For example, the client module 231 may execute an intelligent app for processing voice input by performing an organic action in response to a designated voice input (e.g., wake up!).

In one embodiment, the intelligent server 300 receives information related to the user voice input from the user terminal 201 through the network 299 (e.g., the first network 198 and/or the second network 199 in FIG. 1). You can receive it. According to one embodiment, the intelligent server 300 may change data related to the received voice input into text data. According to one embodiment, the intelligent server 300 may generate at least one plan for performing a task corresponding to the user's voice input based on the text data.

According to one embodiment, the plan may be generated by an artificial intelligence (AI) system. An artificial intelligence system may be a rule-based system, a neural network-based system (e.g., a feedforward neural network (FNN)), and/or a recurrent neural network. network(RNN))). Alternatively, it may be a combination of the above or a different artificial intelligence system. According to one embodiment, a plan may be selected from a set of predefined plans or may be generated in real time in response to a user request. For example, an artificial intelligence system can select at least one plan from a plurality of predefined plans.

The intelligent server 300 of one embodiment may transmit a result according to the generated plan to the user terminal 201 or transmit the generated plan to the user terminal 201. According to one embodiment, the user terminal 201 may display results according to the plan on the display 260. According to one embodiment, the user terminal 201 may display the results of executing an operation according to the plan on the display 260.

The intelligent server 300 of one embodiment includes a front end 310, a natural language platform 320, a capsule database 330, an execution engine 340, It may include an end user interface (350), a management platform (360), a big data platform (370), or an analytic platform (380).

The front end 310 of one embodiment may receive a voice input received by the user terminal 201 from the user terminal 201 . The front end 310 may transmit a response corresponding to the voice input to the user terminal 201.

According to one embodiment, the natural language platform 320 includes an automatic speech recognition module (ASR module) 321, a natural language understanding module (NLU module) 323, and a planner module ( It may include a planner module (325), a natural language generator module (NLG module) (327), and/or a text to speech module (TTS module) (329).

The automatic voice recognition module 321 of one embodiment may convert voice input received from the user terminal 201 into text data. The natural language understanding module 323 in one embodiment may determine the user's intention using text data of voice input. For example, the natural language understanding module 323 may perform syntactic analysis and/or semantic analysis to determine the user's intention. The natural language understanding module 323 in one embodiment determines the meaning of a word extracted from a voice input using linguistic features (e.g., grammatical elements) of a morpheme or phrase, and matches the meaning of the identified word to the user's intention. You can determine your intention.

The planner module 325 in one embodiment may generate a plan using the intent and parameters determined by the natural language understanding module 323. According to one embodiment, the planner module 325 may determine a plurality of domains required to perform the task based on the determined intention. The planner module 325 may determine a plurality of operations included in each of the plurality of domains determined based on the intention. According to one embodiment, the planner module 325 may determine parameters required to execute the determined plurality of operations or result values output by executing the plurality of operations. The parameters and the result values may be defined as concepts of a specified type (or class). Accordingly, the plan may include a plurality of operations and/or a plurality of concepts determined by the user's intention. The planner module 325 may determine the relationship between the plurality of operations and the plurality of concepts in a stepwise (or hierarchical) manner. For example, the planner module 325 may determine the execution order of a plurality of operations determined based on the user's intention based on a plurality of concepts. In other words, the planner module 325 may determine the execution order of the plurality of operations based on the parameters required for execution of the plurality of operations and the results output by executing the plurality of operations. Accordingly, the planner module 325 may generate a plan that includes association information (eg, ontology) between a plurality of operations and a plurality of concepts. The planner module 325 can create a plan using information stored in the capsule database 330, which stores a set of relationships between concepts and operations.

The natural language generation module 327 of one embodiment may change specified information into text form. The information changed to the text form may be in the form of natural language speech. The text-to-speech conversion module 329 in one embodiment can change information in text form into information in voice form.

According to one embodiment, some or all of the functions of the natural language platform 320 may be implemented in the user terminal 201. For example, the user terminal 201 may include an automatic speech recognition module and/or a natural language understanding module. After the user terminal 201 recognizes the user's voice command, it can transmit text information corresponding to the recognized voice command to the intelligent server 300. For example, the user terminal 201 may include a text-to-speech module. The user terminal 201 may receive text information from the intelligent server 300 and output the received text information as voice.

The capsule database 330 may store information about the relationship between a plurality of concepts and operations corresponding to a plurality of domains. The capsule according to one embodiment may include a plurality of action objects (or action information) and/or concept objects (or concept information) included in the plan. According to one embodiment, the capsule database 330 may store a plurality of capsules in the form of CAN (concept action network). According to one embodiment, a plurality of capsules may be stored in a function registry included in the capsule database 330.

The capsule database 330 may include a strategy registry in which strategy information necessary for determining a plan corresponding to a voice input is stored. The strategy information may include standard information for determining one plan when there are multiple plans corresponding to voice input. According to one embodiment, the capsule database 330 may include a follow up registry in which information on follow-up actions is stored to suggest follow-up actions to the user in a specified situation. The follow-up action may include, for example, follow-up speech. According to one embodiment, the capsule database 330 may include a layout registry that stores layout information of information output through the user terminal 201. According to one embodiment, the capsule database 330 may include a vocabulary registry where vocabulary information included in capsule information is stored. According to one embodiment, the capsule database 330 may include a dialogue registry in which information about dialogue (or interaction) with a user is stored. The capsule database 330 can update stored objects through a developer tool. The developer tool may include, for example, a function editor for updating operation objects or concept objects. The developer tool may include a vocabulary editor for updating the vocabulary. The developer tool may include a strategy editor that creates and registers a strategy for determining the plan. The developer tool may include a dialogue editor that creates a dialogue with the user. The developer tool may include a follow up editor that can edit follow-up utterances to activate follow-up goals and provide hints. The subsequent goal may be determined based on currently set goals, user preferences, or environmental conditions. In one embodiment, the capsule database 330 may also be implemented within the user terminal 201.

The execution engine 340 of one embodiment may calculate a result using the generated plan. The end user interface 350 may transmit the calculated result to the user terminal 201. Accordingly, the user terminal 201 can receive the result and provide the received result to the user. The management platform 360 of one embodiment can manage information used in the intelligent server 300. The big data platform 370 in one embodiment may collect user data. The analysis platform 380 of one embodiment may manage quality of service (QoS) of the intelligent server 300. For example, the analytics platform 380 can manage the components and processing speed (or efficiency) of the intelligent server 300.

The service server 400 in one embodiment may provide a designated service (eg, food ordering or hotel reservation) to the user terminal 201. According to one embodiment, the service server 400 may be a server operated by a third party. The service server 400 in one embodiment may provide the intelligent server 300 with information for creating a plan corresponding to the received voice input. The provided information may be stored in the capsule database 330. Additionally, the service server 400 may provide result information according to the plan to the intelligent server 300. The service server 400 may communicate with the intelligent server 300 and/or the user terminal 201 through the network 299. The service server 400 can communicate with the intelligent server 300 through a separate connection. Although the service server 400 is shown as one server in FIG. 2, the embodiments of this document are not limited thereto. At least one of the services 401, 402, and 403 of the service server 400 may be implemented as a separate server.

In the integrated intelligence system described above, the user terminal 201 can provide various intelligent services to the user in response to user input. The user input may include, for example, input through a physical button, touch input, or voice input.

In one embodiment, the user terminal 201 may provide a voice recognition service through an internally stored intelligent app (or voice recognition app). In this case, for example, the user terminal 201 recognizes a user utterance or voice input received through the microphone 270 and provides a service corresponding to the recognized voice input to the user. can do.

In one embodiment, the user terminal 201 may perform a designated operation alone or together with the intelligent server 300 and/or the service server 400 based on the received voice input. For example, the user terminal 201 may run an app corresponding to a received voice input and perform a designated operation through the executed app.

In one embodiment, when the user terminal 201 provides a service together with the intelligent server 300 and/or the service server 400, the user terminal 201 uses the microphone 270 to It is possible to detect an utterance and generate a signal (or voice data) corresponding to the detected user utterance. The user terminal 201 may transmit the voice data to the intelligent server 300 using the communication interface 290.

In response to a voice input received from the user terminal 201, the intelligent server 300 according to one embodiment provides a plan for performing a task corresponding to the voice input, or an operation according to the plan. can produce results. For example, the plan may include a plurality of operations for performing a task corresponding to a user's voice input and/or a plurality of concepts related to the plurality of operations. The concept may define parameters input to the execution of the plurality of operations or result values output by the execution of the plurality of operations. The plan may include association information between multiple operations and/or multiple concepts.

The user terminal 201 in one embodiment may receive the response using the communication interface 290. The user terminal 201 uses the speaker 255 to output a voice signal generated inside the user terminal 201 to the outside, or uses the display 260 to output an image generated inside the user terminal 201 to the outside. It can be output as .

Figure 3 is a diagram showing how relationship information between concepts and actions is stored in a database, according to an embodiment.

The capsule database (e.g., capsule database 330) of the intelligent server 300 may store capsules in CAN (concept action network) format. The capsule database may store operations for processing tasks corresponding to the user's voice input, and parameters necessary for the operations in CAN (concept action network) format.

The capsule database may store a plurality of capsules (capsule A 331, capsule B 334) corresponding to each of a plurality of domains (eg, applications). According to one embodiment, one capsule (e.g., capsule A 331) may correspond to one domain (e.g., location (geo), application). Additionally, one capsule contains at least one capsule of a service provider (e.g., CP 1 (332), CP 2 (333), CP3 (335), and/or CP4 (336) to perform functions for the domain associated with the capsule. )) can correspond. According to one embodiment, one capsule may include at least one operation 330a and at least one concept 330b for performing a designated function.

The natural language platform 320 may create a plan for performing a task corresponding to the received voice input using the capsule stored in the capsule database 330. For example, the planner module 325 of the natural language platform can create a plan using capsules stored in the capsule database. For example, create a plan 337 using the operations 331a, 332a and concepts 331b, 332b of capsule A 331 and the operations 334a and concept 334b of capsule B 334. can do.

Figure 4 is a diagram illustrating a screen in which a user terminal processes voice input received through an intelligent app according to an embodiment.

The user terminal 201 can run an intelligent app to process user input through the intelligent server 300.

According to one embodiment, on the first screen 210, when the user terminal 201 recognizes a designated voice input (eg, wake up!) or receives an input through a hardware key (eg, a dedicated hardware key), the user terminal 201 It can run intelligent apps to process input. For example, the user terminal 201 may run an intelligent app while executing a schedule app. According to one embodiment, the user terminal 201 may display an object (eg, an icon) 211 corresponding to an intelligent app on the display 260. According to one embodiment, the user terminal 201 may receive voice input through a user's utterance. For example, the user terminal 201 may receive a voice input saying “Tell me this week’s schedule!” According to one embodiment, the user terminal 201 may display a user interface (UI) 213 (e.g., input window) of an intelligent app displaying text data of a received voice input on the display.

According to one embodiment, on the second screen 215, the user terminal 201 may display a result corresponding to the received voice input on the display. For example, the user terminal 201 may receive a plan corresponding to the received user input and display ‘this week’s schedule’ on the display according to the plan.

Figure 5 shows a voice recognition environment according to one embodiment.

Referring to FIG. 5 , the voice recognition environment 500 according to an embodiment may include an electronic device 501, a server device 511, and/or a peripheral electronic device 521. For example, the electronic device 501 may correspond to the electronic device 101 of FIG. 1 or the user terminal 201 of FIG. 2. For example, server device 511 may correspond to server 108 in FIG. 1 or intelligent server 300 in FIG. 2. For example, the peripheral electronic device 521 may correspond to the electronic device 102 of FIG. 1 .

The electronic device 501 may be referred to as a listener device that receives the utterance 590 of the speaker 599. The electronic device 501 may include a voice assistant application (eg, the client module 231 of FIG. 2). The electronic device 501 may receive the utterance 590 of the speaker 599 using a voice reception circuit (eg, the audio module 170 of FIG. 1). The electronic device 501 may receive the speaker 599's speech 590 using a peripheral electronic device 521 (eg, an earbud).

For example, the utterance 590 may include a designated call word (eg, wakeup word) for calling a voice assistant application. Electronic device 501 may identify a wakeup word from utterance 590, for example. For example, the electronic device 501 may perform automatic voice recognition for the utterance 590. The electronic device 501 may identify the designated call word from the result of automatic voice recognition (e.g., text corresponding to the utterance 590).

In one example, the electronic device 501 may perform speaker verification using the utterance 590. In this disclosure, speaker verification may be referred to as a procedure for determining whether the voice corresponding to utterance 590 corresponds to the voice of a specific speaker. In one example, the electronic device 501 may perform authentication for the speaker 599 based on the speaker verification result. A detailed description of speaker verification may be described later in relation to FIG. 7 .

In one example, utterance 590 may include a control command. For example, a control command may mean a command containing the user's intention corresponding to control of the electronic device 501 or another electronic device. The speaker 599 may speak a control command along with a designated calling word, or may speak a control command while the voice assistant application is called. In one example, the electronic device 501 may be set to transmit speech data corresponding to the speech 590 to the server device 511. The electronic device 501 may be set to transmit speech data to the server device 511 through a network such as the Internet. The electronic device 501 may receive a plan corresponding to speech data from the server device 511. The electronic device 501 may perform a task corresponding to the intention of the utterance 590 by sequentially performing various operations corresponding to the plan. For example, the electronic device 501 may perform a task corresponding to the intention of the utterance 590 without the intervention of the server 511. The electronic device 501 may perform at least some of the operations of the natural language platform 320 described above with reference to FIG. 2 to generate a plan corresponding to speech data. In this case, the server device 511 may be omitted.

According to one embodiment, the electronic device 501 may perform speaker verification based on at least one condition. For example, the electronic device 501 may perform speaker verification to call a voice assistant application. When the utterance 590 includes a designated call word, the electronic device 501 may perform speaker verification to determine whether the utterance 590 is by a designated user. If the utterance 590 includes a designated call word and speaker verification based on the utterance 590 is successful, the electronic device 501 can execute the voice assistant application. For example, the electronic device 501 may perform speaker verification to unlock. When the electronic device 501 receives an utterance 590 while in a locked state, the electronic device 501 may perform speaker verification based on the utterance 590. If speaker verification based on the utterance 590 is successful, the electronic device 501 may unlock the electronic device 501. For example, the electronic device 501 may perform speaker verification to transition from the standby state to the wake-up state. In the standby state, the electronic device 501 may control a part of the processor 620 (e.g., the auxiliary processor 123 in FIG. 1) to be in an active state and control the remainder to be in a low-power state. When receiving an utterance 590 in a standby state, the electronic device 501 may perform speaker verification based on the utterance 590. If speaker verification based on the utterance 590 is successful, the electronic device 501 may switch from a standby state (e.g., low-power state) to a wake-up state (e.g., high-power state). The electronic device 501 may enter the wake-up state by activating the remainder of the processor 620 (e.g., the main processor 121 in Figure 1). For example, the electronic device 501 may perform unlocking and voice assistant functions. Speaker verification may be performed to execute the application. When the electronic device 501 receives an utterance 590 while in a locked state, the electronic device 501 may perform speaker verification based on the utterance 590. When speaker verification based on the utterance 590 is successful, the electronic device 501 can unlock the electronic device 501 and run a voice assistant application. For example, the electronic device 501 may perform speaker verification to perform the control command. When the utterance 590 includes a control command, user authentication may be required to perform the control command. The electronic device 501 may perform user authentication. For this purpose, speaker verification based on the utterance 590 can be performed.

Server device 511 may include at least one server device. For example, the server device 511 may be configured to receive speech data from the electronic device 501 and process the speech data. The server device 511 may determine the intent of the speaker 599 included in the received speech data by processing the speech data received from the electronic device 501. The server device 511 may create a plan corresponding to the intention and transmit the generated plan to the electronic device 501.

The voice recognition environment 500 described above with reference to FIG. 5 is illustrative, and embodiments of the present disclosure are not limited thereto. Those skilled in the art will understand that the server device 511 and/or the peripheral electronic device 521 may be omitted.

Figure 6 shows a block diagram of an electronic device according to an embodiment.

Referring to Figure 6, according to one embodiment, the electronic device 501 may include a processor 620, memory 630, microphone 650, display 660, and/or communication circuitry 690. there is. For example, the processor 620 may correspond to the processor 120 of FIG. 1 or the processor 220 of FIG. 2. For example, memory 630 may correspond to memory 130 of FIG. 1 or memory 230 of FIG. 2 . For example, the microphone 650 may correspond to the audio module 170 of FIG. 1 or the microphone 270 of FIG. 2. For example, display 660 may correspond to display module 160 of FIG. 1 or display 260 of FIG. 2 . For example, the communication circuit 690 may correspond to the communication module 190 of FIG. 1 or the communication interface 290 of FIG. 2. The configuration of the electronic device 601 in FIG. 6 is an example, and the electronic device 601 may further include a configuration not shown in FIG. 6 .

Processor 620 may be electrically, operatively, or functionally connected to memory 630, microphone 650, display 660, and/or communication circuitry 690. there is. In the present disclosure, when one component is “operably” connected to another component, it may mean that one component is connected to operate the other component. For example, one component may actuate another component by transmitting a control signal to the other component, either directly or via another component. In the present disclosure, when one component is “functionally” connected to another component, it may mean that one component is connected to execute the function of the other component. For example, one component may execute the function of another component by transmitting a control signal to the other component directly or via another component.

Memory 630 may store instructions. When executed by the processor 620, the instructions may cause the electronic device 501 to perform various operations. In this disclosure, the operation of the electronic device 501 may be referred to as an operation performed by the processor 620 by executing instructions stored in the memory 630. In one embodiment, the memory 630 may store at least one speaker verification model for speaker verification. The speaker verification model may be described later with respect to FIG. 7 .

The microphone 650 can convert sound signals into electrical signals. The electronic device 501 may convert the electrical signal received from the microphone 650 into a digital signal and perform data processing based on the digital signal.

The communication circuit 690 may provide communication between the electronic device 501 and other electronic devices (eg, the server device 611 and/or the peripheral electronic device 621 of FIG. 5 ). Communication circuitry 690 may support wired and/or wireless communication. Communication circuitry 690 may support short-range wireless communication and/or long-range wireless communication.

The electronic device 501 can perform speaker verification. The electronic device 501 may acquire speech data using the microphone 650 or the communication circuit 690. The electronic device 501 can extract features about speech data. For example, the electronic device 501 may extract a vector of features for speech included in speech data (eg, a speech vector). The electronic device 501 may compare the speech vector corresponding to the speech data with the speaker verification vector stored in the memory 630. If the similarity between the voice vector and the speaker verification vector is greater than or equal to the threshold, the electronic device 501 may verify that the speaker of the utterance corresponding to the voice vector is the same as the speaker in the stored speaker verification vector.

According to one embodiment, the electronic device 501 may adjust sensitivity for speaker verification. For example, the electronic device 501 may adjust sensitivity by adjusting a threshold for speaker verification (e.g., a threshold for similarity between a speech vector and a speaker verification model). The electronic device 501 may adjust the threshold for speaker verification using a plurality of voice samples corresponding to the speaker's utterance. For example, electronic device 501 may provide recommendations for adjustment of the threshold. Below, various examples for adjusting the threshold of the electronic device 501 may be described with reference to FIGS. 7 to 12 .

In one example, the electronic device 501 may include a display 660, a memory 630, and a processor 620. The memory 630 may store instructions that, when executed by the processor 620, allow the processor 620 to perform various operations. The processor 620 may acquire a plurality of voice signals including a wakeup word. The processor 620 may determine whether to verify the speaker for each of the plurality of voice signals by comparing the similarity between each of the plurality of voice signals and the speaker verification model stored in the memory 630 with a first threshold. For example, the processor 620 determines that speaker verification for each of the plurality of speech signals has failed if the similarity between each of the plurality of speech signals and the speaker verification model stored in the memory 630 is less than the first threshold. You can. The processor 620 may store a voice signal that fails speaker verification in the memory 630.

The processor 620 receives user authentication for the electronic device based on an authentication method other than speaker verification within a specified time from the reception of the last voice signal among the plurality of voice signals, after speaker verification of the plurality of voice signals fails. can do. For example, user authentication may include authentication based on at least one of a password, fingerprint, iris, or pattern. For example, the processor 620 may change the state of the electronic device 501 from the locked state to the unlocked state based on receipt of user authentication. For example, the processor 620 may display a user interface (UI) recommending adjustment of the first threshold on the display 630 based on user authentication. The processor 620 may display a UI when the voice agent application of the electronic device 501 is executed after changing to the unlocked state. For example, the UI may include information about a plurality of voice signals.

The processor 620 may adjust the first threshold based on the similarity between at least some of the plurality of voice signals and the speaker verification model based on the input to the UI. The processor 620 may adjust the first threshold using at least some of a plurality of voice signals selected based on a user input to the UI. For example, the processor 620 may adjust the first threshold to the second threshold based on the similarity between each of at least some of the plurality of voice signals and the stored speaker verification model. The processor 620 may identify the second threshold based on the highest first similarity among the similarities between each of at least some of the plurality of voice signals and the stored speaker verification model. The processor 620 may identify the second threshold based on the weighted sum of the first similarity and the first threshold.

When the instructions are executed by the processor, the processor displays the UI when the voice agent application of the electronic device is executed after changing to the unlocked state.

Figure 7 shows modules of an electronic device according to an embodiment.

Referring to FIGS. 6 and 7 , according to one embodiment, the electronic device 501 may include a plurality of modules. For example, the components of the electronic device 501 described with reference to FIG. 7 may be software modules implemented by executing instructions stored in the memory 630 by the processor 620. At least some of the modules of the electronic device 501 may be hardware modules. In one example, the memory 630 includes a wakeup word recognition model database 725, an evaluation data database 735, and a speaker verification model database 745. You can save it. In one example, speaker verification module 740, speaker verification manager 760, and voice assistant client 750 are part of processor 630 (e.g., application processor or main processor 121 of FIG. 1). ) can be implemented by. In one example, the preprocessing module 710, the wakeup word recognition module 720, and the data management module 730 may be connected to another part of the processor 620 (e.g., a digital signal processor (DSP) or a co-processor of FIG. 1). 122)). Operations of modules described in relation to FIG. 7 may be referred to as operations of the processor 620.

The preprocessing module 710 may perform preprocessing on acquired voice data. For example, the preprocessing module 710 may perform preprocessing on voice data acquired through the microphone 650 or voice data received from another electronic device through the communication circuit 690. Preprocessing may include various operations for the voice data to be used by other modules. For example, the preprocessing module 710 may perform analog to digital conversion (ADC) on the voice signal acquired by the microphone 650. The preprocessing module 710 performs echo cancellation, noise suppression, automatic gain control, acoustic echo cancellation (AEC), feature extraction, and/or windowing on audio data. You can.

The wake-up recognition module 720 can recognize a wake-up word using voice data (hereinafter referred to as voice data) processed by the pre-processing module 710. For example, the wakeup word recognition model DB 725 may include at least one wakeup word recognition model corresponding to a designated call word (eg, wakeup word). The wakeup word recognition model may be, for example, a vector generated based on a wakeup word uttered by the user of the electronic device 501 in an enrollment process. The wakeup word recognition module 720 may compare the wakeup word recognition model and the voice data to determine whether the voice data includes a wakeup word. For example, the wake-up recognition module 720 may determine whether the voice data includes an utterance corresponding to a wake-up word based on automatic voice recognition of the voice data.

In one example, when the voice data includes a wakeup word, the electronic device 501 may execute the voice assistant client 750. In one example, if the voice data includes a wakeup word and speaker verification based on the voice data is successful, the electronic device 501 may execute the voice assistant client 750. In one example, the electronic device 501 may transition from a standby state (e.g., a low power state) to an active state (e.g., a high power state) when the voice data includes a wakeup word. The voice assistant client 750 may correspond to the client module 231 of FIG. 2.

The data management module 730 can manage the evaluation data DB 735. The data management module 730 may store the voice data received from the pre-processing module 710 in the evaluation data DB 735. For example, the data management module 730 may store a specified number of voice samples (eg, voice data) in the evaluation data DB 735. For example, the data management module 730 may store the most recently received N (eg, N is an integer equal to or greater than 2) voice samples in the evaluation data DB 735. For example, the data management module 730 may store voice samples received within a specified period (eg, 3 days) in the evaluation data DB 735.

In one example, the data management module 730 may store voice data including a wake-up word in the evaluation data DB 735. When the wakeup word recognition module 720 identifies the wakeup word from the voice data, the data management module 730 may store the voice data in the evaluation data DB 735. In one example, the data management module 730 may store voice data that includes a wakeup word and fails speaker verification in the evaluation data DB 735. If the wakeup word recognition module 720 identifies the wakeup word from the voice data, and the speaker verification module 740 fails to verify the speaker from the voice data, the data management module 730 records the voice in the evaluation data DB 735. Data can be saved.

In one example, the data management module 730 may delete voice samples (eg, voice data) stored in the evaluation data DB 735. The data management module 730 may delete the oldest voice sample when the number of stored voice samples exceeds a specified number. The data management module 730 may delete the voice sample selected by the user based on the user input. The data management module 730 may delete the used voice samples once they have been used to adjust sensitivity (e.g., threshold) for speaker verification. The data management module 730 may delete the stored voice sample when a specified time elapses after the voice sample is stored.

In one example, the data management module 730 may map each voice sample to a user identifier and store it. For example, for a voice sample that fails speaker verification, the user can specify a user identifier corresponding to the voice sample. In this case, the data management module 730 may map the designated user identifier to the corresponding voice sample and store it.

In one example, data management module 730 can classify voice samples. The data management module 730 may classify voice samples based on their similarity. For example, similarity between voice samples of the data management module 730 may be calculated. Similarity may be calculated based on, for example, pitch, gender, age, and/or any voice characteristic. For example, the similarity may be calculated based on the similarity (e.g., distance) between speech vectors corresponding to speech samples. The similarity may be calculated, for example, based on the similarity (e.g., distance) between speech vectors corresponding to speech samples and speaker verification vectors. In one example, the similarity between the voice sample and the speaker verification vector may be calculated by the speaker verification module 740. The data management module 730 may map the calculated similarity to the voice sample and store it in the evaluation data DB 735.

In one example, the data management module 730 may group voice samples whose similarity is greater than or equal to a specified value and store them in the evaluation data DB 735. In one example, the data management module 730 may store in the evaluation data DB 735 only voice samples whose similarity is greater than or equal to a specified value.

For example, the speaker verification model DB 745 may include at least one speaker verification model (eg, speaker verification vector). For example, the electronic device 501 may receive a plurality of utterances corresponding to a designated word (e.g., a wakeup word) from the user and generate a speaker verification model corresponding to the user based on the plurality of utterances. . The electronic device 501 may store the speaker verification model in the speaker verification model DB 745. In one example, the electronic device 501 may store the speaker verification model in a secure area of the memory 630.

The speaker verification module 740 may perform speaker verification using voice data processed by the pre-processing module 710. For example, the speaker verification module 740 may perform speaker verification by comparing the speaker verification model stored in the speaker verification model DB 745 with the speech data received from the pre-processing module 710. The speaker verification module 740 may perform speaker verification based on the similarity between the speech vector corresponding to speech data and the speaker verification model. In one example, the speaker verification model may include eigenvectors (e.g., i-vector, x-vector, etc.) extracted from the speaker's voice (e.g., the voice received during the speaker's registration process). For example, the speaker verification model may be previously created during the speaker registration process and pre-stored in the speaker verification model 745.

For example, the speaker verification module 740 may identify the similarity between the speech vector and the speaker verification model based on likelihood ratio test (LRT), Euclidian distance, or Cosine similarity. For example, speaker verification module 740 may identify similarity based on the distance between the speech vector and the speaker verification model.

The speaker verification module 740 may compare the identified similarity with a threshold to determine whether speaker verification is successful. If the similarity is below the threshold, the speaker verification module 740 may determine that speaker verification based on voice data has failed. If the similarity exceeds the threshold, the speaker verification module 740 may determine that speaker verification based on speech data was successful.

The threshold for speaker verification may be, for example, a value set by the manufacturer of the electronic device 501. For example, the threshold may be a value set by the user among values (eg, low, medium, high) set by the manufacturer of the electronic device 501. The threshold may be a value adjusted by the speaker verification manager 760, as will be described later.

According to one embodiment, the speaker verification manager 760 may identify (eg, calculate) a threshold for speaker verification. The threshold identified by the speaker verification manager 760 may be referred to as a candidate threshold. The candidate threshold may be used to adjust the threshold used by speaker verification module 740. In one example, the speaker verification manager 760 may calculate a candidate threshold for speaker verification based on the setting value of the electronic device 501. If the user sets the speaker verification threshold to allow adjustment, the speaker verification manager 760 can calculate a candidate threshold.

The speaker verification manager 760 may identify (e.g., determine) a candidate threshold when user authentication is performed by a method other than speaker verification after speaker verification fails. For example, user authentication by a method other than speaker verification may include at least one of user authentication based on password input, user authentication based on fingerprint, user authentication based on pattern input, or iris-based user authentication. User authentication by methods other than speaker verification is not limited to the methods listed and may include any user authentication method. For example, speaker verification manager 760 may identify a candidate threshold when user authentication (e.g., successful user authentication) is performed after a specified number of speaker verification failures. If speaker verification based on repeated utterances more than a specified number of times fails, the electronic device 501 may adjust the threshold using the speaker verification manager 760. For example, speaker verification manager 760 may identify a candidate threshold when user authentication is performed within a specified time from the last speaker verification failure. As described above, speaker verification manager 760 may identify candidate thresholds when an adjustment of the thresholds is established. In this disclosure, identifying a candidate threshold may include determining a candidate threshold.

For example, speaker verification manager 760 may identify candidate thresholds when user authentication is performed after speaker verification failure. For example, speaker verification manager 760 may identify candidate thresholds when user authentication is performed and voice assistant client 750 is launched. For example, the speaker verification manager 760 may identify a candidate threshold when displaying a UI recommending update of the threshold (e.g., operation 820 of FIG. 8 or operation 925 of FIG. 9).

The speaker verification manager 760 may identify a candidate threshold using voice samples stored in the evaluation data DB 735. Speaker verification manager 760 may identify a candidate threshold using, for example, a specified number of speech samples that have a high degree of similarity to the most recently stored speech sample. Speaker verification manager 760 may identify a candidate threshold using, for example, a specified number of most recently stored voice samples. Speaker verification manager 760 may identify a candidate threshold using, for example, voice samples selected by the user.

The speaker verification manager 760 may identify a candidate threshold using the similarity between each of the voice samples and the speaker verification model. For example, the speaker verification manager 760 may identify the similarity (e.g., similarity score) of the voice sample with the highest similarity as the candidate threshold. For example, the speaker verification manager 760 adds a value obtained by applying a first weight to at least one of the similarities of voice samples (e.g., the similarity of the highest voice sample) and a value obtained by applying a second weight to the current threshold to create a candidate candidate. Thresholds can be identified. The sum of the first weight and the second weight may be, for example, 1.

For example, voice samples such as those shown in Table 1 below can be used to identify candidate thresholds.

voice samples Similarity score sample 1 63 sample 2 62 sample 3 65 sample 4 64 sample 5 61

The speaker verification manager 760 may identify the similarity of sample 4, which has the highest similarity, as a candidate threshold. For example, the currently set threshold may be 70. The speaker verification manager 760 may identify the value obtained by multiplying the threshold value by the second weight and the similarity of sample 4 by multiplying the first weight value as a candidate threshold value.

Speaker verification manager 760 may identify a plurality of candidate thresholds (eg, a low threshold, a middle threshold, and a high threshold). For example, speaker verification manager 760 may identify a plurality of candidate thresholds by identifying candidate thresholds based on speech samples and identifying thresholds that have a specified difference from the identified candidate threshold.

For example, the threshold value currently set in the electronic device 501 may be a low threshold value. Speaker verification manager 760 may identify a candidate threshold based on voice samples. Speaker verification manager 760 may identify the identified candidate threshold as a candidate for a lower threshold. Speaker verification manager 760 may identify an intermediate candidate threshold, for example, by adding the first value. Speaker verification manager 760 may identify a high candidate threshold by adding a second value to the middle candidate threshold.

The speaker verification manager 760 may adjust the threshold value currently set in the electronic device 501 using the identified candidate threshold value (or a plurality of candidate threshold values). The speaker verification manager 760 can adjust the set threshold by setting the candidate threshold as a threshold for speaker verification. For example, speaker verification manager 760 may adjust thresholds once candidate thresholds are identified. For example, speaker verification manager 760 can adjust the threshold based on user input. The speaker verification manager 760 may adjust the threshold based on a user input to a UI that recommends updating the threshold (e.g., the UI in operation 820 of FIG. 8 or operation 925 of FIG. 9).

When the threshold is adjusted, the speaker verification module 750 can perform speaker verification using the adjusted threshold. For example, the electronic device 501 may receive speech after adjusting the threshold. The speaker verification module 740 may receive voice data corresponding to the received utterance from the preprocessing module 710. The speaker verification module 740 may identify the similarity between the speech data and the speaker verification model, and determine that speaker verification is successful if the identified similarity exceeds an adjusted threshold.

According to one embodiment, the electronic device 501 may restore the adjusted threshold to the threshold before adjustment (eg, the original threshold). The electronic device 501 may store the original threshold value in the memory 630. In one example, the electronic device 501 may restore the threshold based on a specified condition (eg, at least one of a specified time, similarity between speech data and a speaker verification model, or user input).

For example, the electronic device 501 may restore the adjusted threshold value to the original threshold value when a specified time (eg, two weeks) elapses after adjusting the threshold value. When a specified time elapses, the electronic device 501 provides a user interface to inquire about restoration of the threshold value, and restores the threshold value when a user input allowing restoration is received.

For example, after adjusting the threshold, the electronic device 501 may restore the adjusted threshold to the original threshold based on the similarity between the received voice data and the speaker verification model. For example, the electronic device 501 restores the adjusted threshold to the original threshold level based on the distribution of speaker recognition scores (e.g., similarity) of user voice inputs entered during a recent specified period (e.g., a week). can do. For example, if the average value of audio samples that succeeded in speaker verification among voice inputs attempted within a specified period is greater than the adjusted threshold or the specified threshold, the electronic device 501 changes the adjusted threshold to the original threshold. It can be reset. The electronic device 501 provides a user interface to inquire about restoration of the threshold when the average value of audio samples that have succeeded in speaker verification is greater than an adjusted threshold or a specified threshold, and when a user input allowing restoration is received, the threshold can be restored.

For example, after adjusting the threshold, the electronic device 501 may restore the adjusted threshold to the original threshold based on the user input. For example, the electronic device 501 may provide a user interface that can reset the threshold based on user input. When a user input instructing to reset the threshold value is received through the user interface, the electronic device 501 may restore the threshold value to the original threshold value.

Figure 8 shows a flowchart of a sensitivity adjustment method according to one embodiment.

Referring to FIGS. 6 and 8 , according to one embodiment, the electronic device 501 may acquire a plurality of voice signals including a wakeup word in operation 805. For example, the electronic device 501 may acquire a plurality of voice signals using the microphone 650. In the example of FIG. 8, the plurality of voice signals can be assumed to be voice signals that have successfully recognized the wakeup word. For example, each of the plurality of voice signals may be assumed to be a voice signal whose wakeup word has been successfully detected by the wakeup word recognition module 720 of FIG. 7 .

In operation 810, the electronic device 501 may determine whether speaker verification for each of the plurality of voice signals is successful. The electronic device 501 may compare each of the plurality of voice signals with a speaker verification model to determine whether speaker verification is successful. The electronic device 501 may determine that speaker verification is successful when the similarity between each of the plurality of voice signals and the speaker verification model exceeds a specified threshold (eg, a first threshold). The electronic device 501 may determine that speaker verification has failed if the similarity between each of the plurality of voice signals and the speaker verification model is less than or equal to a specified threshold. For example, the electronic device 501 may perform speaker verification according to the operation of the speaker verification module 740 described above with reference to FIG. 7 .

In FIG. 8, operation 810 is shown as a subsequent operation to operation 805, but the electronic device 501 may perform speaker verification (eg, operation 810) each time each of a plurality of voice signals is performed. Operations 810 and 815 may correspond to operations 905, 910, and 915 described below with respect to FIG. 9. In the example of FIG. 8, speaker verification for each of the plurality of voice signals may be assumed to have failed.

In operation 815, the electronic device 501 may receive user authentication based on an authentication method other than speaker verification within a specified time after speaker verification fails. The designated time may mean a time interval of a certain length set from the reception time of the voice signal that was most recently received and failed speaker verification. User authentication by a method other than speaker verification may include at least one of user authentication based on password input, user authentication based on fingerprint, user authentication based on pattern input, or iris-based user authentication. The user authentication of operation 815 may be referred to as successful user authentication. That is, through an utterance including a wake-up word, the user may have attempted to perform a designated action (e.g., exiting standby mode, calling a voice assistant application, and/or executing an action through a voice command), but failed. After failing to perform the specified operation, the user may attempt to perform the specified operation through direct user authentication rather than voice. For example, user authentication in operation 815 may be performed after a preset number of speaker verification failures. The electronic device 501 may perform operation 820 when user authentication is received after a preset number of speaker verification failures.

In operation 820, the electronic device 501 may display a UI recommending updating the threshold for speaker verification based on user authentication. For example, the electronic device 501 may display a UI when user authentication is successful, when a voice agent is executed after user authentication, or when the lock is unlocked according to user authentication.

For example, the UI may include content suggesting adjustment of the threshold for speaker verification. When a user input allowing adjustment of the threshold value is received through the UI, the electronic device 501 can adjust the threshold value.

For example, the UI may include an interface for selection of speech samples for adjustment of thresholds for speaker verification. For example, the electronic device 501 may adjust the threshold using only a portion of the complement voice signals stored in the memory. In one example, the UI may include at least one of an icon, a button, or a hyperlink for switching to a screen or page for adjusting a threshold for speaker verification. In one example, the electronic device 501 may display on the UI only information on voice samples (eg, voice samples corresponding to at least some of a plurality of voice signals) whose similarity is greater than or equal to a specified value. The electronic device 501 may perform primary filtering based on similarity. Filtering of voice samples may be performed based on, for example, similarity in gender, age, and/or pitch. In one example, the electronic device 501 may receive a user input for selecting a voice sample to be used for threshold adjustment among voice samples displayed on the UI. For example, the electronic device 501 may adjust the threshold using voice samples selected according to user input. Regarding UI, various examples can be described with reference to FIGS. 10 to 12.

In operation 825, the electronic device 501 may update the threshold based on at least some of the plurality of voice signals based on the input to the UI. For example, the electronic device 501 may receive a user input allowing update (eg, adjustment) of the threshold value through a UI. Based on reception of the user input, the electronic device 501 may update the threshold using at least some of the plurality of voice signals. For example, the electronic device 501 may change the threshold from a first threshold to a second threshold. For example, the electronic device 501 may identify a threshold to be changed (eg, a second threshold) according to the candidate threshold identification method of the speaker verification manager 760 of FIG. 7 .

With reference to FIG. 8 , the electronic device 501 has been described as providing a UI, but the embodiments of this document are not limited thereto. For example, the electronic device 501 may provide a UI through a peripheral electronic device (eg, the peripheral electronic device 521 in FIG. 5). For example, the electronic device 501 may provide an audio UI instead of a visual UI. The electronic device 501 may output a voice inquiry asking the user whether to adjust the threshold. In this case, the electronic device 501 may not include a display.

Figure 9 shows a flowchart of a sensitivity adjustment method according to one embodiment.

Referring to FIGS. 6 and 9 , in operation 905, the electronic device 501 may acquire voice data. For example, the electronic device 501 may acquire voice data by directly receiving a voice signal using the microphone 650. For example, the electronic device 501 may receive voice data from a peripheral electronic device (eg, the peripheral electronic device 521 of FIG. 5) through the communication circuit 690. Voice data may be referred to, for example, as a voice signal processed by the preprocessing module 710 of FIG. 7 .

In operation 910, the electronic device 501 may determine whether recognition of the wakeup word is successful from voice data. For example, the electronic device 501 may attempt to recognize a wakeup word by performing the operations of the wakeup word recognition module 720 of FIG. 7 .

If recognition of the wakeup word fails (e.g., operation 910-NO), the electronic device 501 may maintain the previous state. When the electronic device 501 is in the locked state or standby state, in response to failure to recognize the wakeup word, the electronic device 501 may maintain the locked state or the standby state.

If recognition of the wakeup word is successful (e.g., operation 910-YES), the electronic device 501 may determine whether speaker verification is successful in operation 915. The electronic device 501 may perform speaker verification, for example, by performing the operation of the speaker verification module 740 of FIG. 7 . In Figure 9, operation 915 is described as following operation 910, but embodiments of the present document are not limited thereto. Operations 915 and 910 may be performed substantially simultaneously. Operation 915 may be performed prior to operation 910.

If speaker verification is successful (e.g., operation 915-YES), in operation 930, the electronic device 501 may perform an operation corresponding to the voice data. For example, if the intention of the voice data is to call a voice assistant application, the electronic device 501 may execute the voice assistant application. The electronic device 501 may switch from a standby state or a locked state to an active state or an unlocked state depending on the success of speaker verification.

If speaker verification fails (e.g., operation 915-NO), in operation 920, the electronic device 501 may determine whether user authentication is successful within a specified time from speaker verification failure. User authentication may be referred to as user authentication based on methods other than speaker verification. If recognition of the wakeup word is successful, but speaker verification is unsuccessful, the electronic device 501 may store voice data in the memory. For example, the electronic device 501 may store voice data in memory according to the operations of the data management module 730 of FIG. 7 . The designated time may be referred to as a preset time from the time of last speaker verification failure (e.g., the time of acquisition of voice data that failed speaker verification).

If user authentication is not successful within a specified time (eg, operation 920-NO), the electronic device 501 may maintain the previous state. In this case, the electronic device 501 may delete voice data stored in the memory.

If user authentication is successful within a specified time (e.g., operation 920-YES), in operation 925, the electronic device 501 may provide a UI for adjusting the threshold. Operation 925 may correspond to operation 820 of FIG. 8 . As described above with reference to FIG. 8 , the electronic device 501 may display a UI when user authentication is successful, when a voice agent is executed after user authentication, or when the device is unlocked according to user authentication. As described above with reference to FIG. 8 , the electronic device 501 may adjust the threshold for speaker verification based on input to the UI.

10 illustrates an example UI of an electronic device according to an embodiment.

Referring to FIGS. 6 and 10 , according to one embodiment, the electronic device 501 may provide a UI 1000 recommending adjustment of a threshold for speaker verification. For example, the electronic device 501 may display the UI 1000 on the display 660 according to operation 820 of FIG. 8 or operation 925 of FIG. 9 .

In one example, the electronic device 501 may display information on a plurality of voice samples stored in the memory 630 (e.g., the evaluation data DB 735 of FIG. 7) on the UI 1000. Voice samples displayed on the UI 1000 may be selected by the electronic device 501 (eg, the data management module 730 of FIG. 7) from among voice samples stored in the memory. For example, the electronic device 501 may select a specified number of voice samples with high similarity among a plurality of voice samples stored in the memory and display them on the UI 1000. The electronic device 501 may determine the degree of similarity between voice samples based on at least one of gender, age, or pitch corresponding to the voice sample.

In the example of FIG. 10, the electronic device 501 includes a first voice sample 1011, a second voice sample 1012, a third voice sample 1013, a fourth voice sample 1014, and a fifth voice sample ( 1015) can be displayed on the UI (1000). When an input to one of the play buttons 1020 is received, the electronic device 501 may play a voice sample corresponding to the play button 501 from which the input was received. The user can check the voice sample to be used for sensitivity adjustment (e.g., threshold adjustment) through playback of the voice sample.

When an input for one of the delete buttons 1030 is received, the electronic device 501 may exclude from sensitivity adjustment the voice sample corresponding to the delete button 501 for which the input was received. For example, a user can play a voice sample and then delete voice samples that are not his or her own voice. By deleting a voice sample, the user can prevent the voice sample from being used for sensitivity adjustment.

When an input to the sensitivity update button 1050 is received, the electronic device 501 can adjust the sensitivity using the selected voice samples. Sensitivity adjustment based on user input may correspond to operation 825 of FIG. 8. In the example of FIG. 10, voice samples that were not deleted by the user may be referred to as voice samples selected by the user. The method for adjusting sensitivity may correspond to the method for identifying a candidate threshold and the method for adjusting a threshold using a candidate threshold described above with reference to FIG. 7 .

11 illustrates an example UI of an electronic device according to an embodiment.

Referring to FIGS. 6 and 11 , according to one embodiment, the electronic device 501 may provide a UI 1100 that recommends adjustment of a threshold for speaker verification. For example, the electronic device 501 may display the UI 1100 on the display 660 according to operation 820 of FIG. 8 or operation 925 of FIG. 9 . A description of a configuration having the same identification number as the UI 1000 of FIG. 10 may be referred to by the description of FIG. 10 .

In the example of FIG. 11 , UI 1100 may include a selection button 1140. The user can select a voice sample to be used for sensitivity adjustment through input to the selection button 1140. For example, the electronic device 501 may or may not use the corresponding voice sample for sensitivity adjustment based on the input to the selection button 1140.

Figure 12 shows an example UI of an electronic device according to an embodiment.

Referring to FIGS. 6 and 12 , according to one embodiment, the electronic device 501 may provide a UI 1200 that recommends adjustment of a threshold for speaker verification. For example, the electronic device 501 may display the UI 1200 on the display 660 according to operation 820 of FIG. 8 or operation 925 of FIG. 9 . A description of a configuration having the same identification number as the UI 1000 of FIG. 10 may be referred to by the description of FIG. 10 .

In the example of FIG. 12 , UI 1200 may include a user ID input interface 1260. For example, the user can specify a user ID (eg, speaker ID) to each voice sample through input to the user ID input interface 1260. In one example, the electronic device 501 may store a plurality of speaker verification vectors. Each of the plurality of speaker verification vectors may have a different speaker ID. In the examples in this document, it can be assumed that the voice sample used for sensitivity adjustment contains a wake word but fails speaker verification. Accordingly, the electronic device 501 may not be able to identify the speaker ID from the voice sample. For example, the user can listen to a voice sample through input to the play button 1020. After listening to the voice samples, the user can determine which user ID (e.g., speaker ID) each voice sample corresponds to. The user can specify a user ID for each voice sample through input to the user ID input interface 1260.

For example, in the example of FIG. 12, a voice sample with a user ID assigned to Sunny may be used for sensitivity adjustment associated with the speaker verification model corresponding to Sunny. A voice sample with the user ID assigned to Jessica can be used for sensitivity adjustment associated with the speaker verification model corresponding to Jessica.

Claims (10)

In electronic devices,
display;
Memory; and
Includes a processor,
When the memory is executed by the processor, the processor:
Acquire a plurality of voice signals including a wake-up word,
Compare the similarity between each of the plurality of voice signals and the speaker verification model stored in the memory with a first threshold to determine whether to verify the speaker for each of the plurality of voice signals,
After failure of the speaker verification for the plurality of voice signals, receiving user authentication for the electronic device based on an authentication method other than the speaker verification within a specified time from the time of reception of the last voice signal among the plurality of voice signals, and ,
Based on the user authentication, the electronic device stores instructions for adjusting the first threshold based on similarity between at least some of the plurality of voice signals and the speaker verification model. According to claim 1,
When the instructions are executed by the processor, the processor determines whether the similarity between each of the plurality of voice signals and the speaker verification model stored in the memory is less than the first threshold value for each of the plurality of voice signals. An electronic device that determines that speaker verification has failed. According to claim 2,
When the instructions are executed by the processor, the electronic device causes the processor to store the voice signal that failed the speaker verification in the memory. According to claim 1,
When the instructions are executed by the processor, the processor:
Displaying a user interface (UI) recommending adjustment of the first threshold on the display,
An electronic device that adjusts the first threshold using at least some of the plurality of voice signals based on a user input to the UI. According to claim 4,
The instructions, when executed by the processor, cause the processor to adjust the first threshold to a second threshold based on similarity between each of at least some of the plurality of speech signals and a stored speaker verification model. , electronic devices. According to claim 5,
When the instructions are executed by the processor, the processor:
a first similarity level that is the highest among similarities between each of at least some of the plurality of speech signals and a stored speaker verification model; or
An electronic device that identifies the second threshold based on a weighted sum of the first similarity and the first threshold. According to claim 1,
The instructions, when executed by the processor, cause the processor to change the state of the electronic device from a locked state to an unlocked state in response to receiving the user authentication. According to claim 7,
The user authentication includes authentication based on at least one of a password, fingerprint, iris, or pattern. According to claim 7,
When the instructions are executed by the processor, the processor displays a UI (user interface) recommending adjustment of the first threshold when the voice agent application of the electronic device is executed after the change to the unlock state. An electronic device that causes a display. A method for adjusting the speaker verification sensitivity of an electronic device, comprising:
Obtaining voice data including a wake-up word;
performing speaker verification on the speech data by comparing a similarity between the speech data and a speaker verification model stored in a memory of the electronic device to a first threshold;
If the speaker verification fails, determining whether successful user authentication is received within a specified time from the point of acquisition of the voice data; and
Upon receiving the successful user authentication, adjusting the first threshold using a plurality of voice data stored in the memory,
Each of the plurality of voice data includes the wakeup word and is voice data that fails the speaker verification.

Images