KR102583799B1 - Method for detect voice activity in audio data based on anomaly detection - Google Patents

Abstract

In accordance with one embodiment of the present disclosure, a method for detecting voice activity in audio data, performed by a computing device, is disclosed. The method includes inputting audio data into a voice activity detection (VAD) model; Predicting whether the audio data is classified as normal data or abnormal data using the VAD model; And when the audio data is classified as normal data, using the VAD model, determining that the audio data corresponds to voice data.

Description

{METHOD FOR DETECT VOICE ACTIVITY IN AUDIO DATA BASED ON ANOMALY DETECTION}

The present invention relates to a method for detecting voice activity in audio data, and more specifically, to a technology for detecting voice activity in audio data based on anomaly detection.

Voice Activity Detection (VAD) is a technology that determines whether voice activity has been detected in audio data (audio signal, audio stream, etc.). In other words, VAD technology is a technology that detects whether audio data (or part of the audio data) corresponds to voice data or non-voice data, and is a technology for voice recognition (STT). ; Speech-To-Text) This is a technology widely used in the process of selecting data to be input into the model.

However, conventional VAD technology is implemented based on supervised learning or by utilizing statistical evidence such as probability distribution. Therefore, the conventional VAD technology had a problem in that it was difficult to prepare voice data or non-voice data, or the performance of the VAD was difficult to guarantee when the imbalance between voice data and non-voice data was severe.

On the other hand, Anomaly Detection (AD) is a technology that detects and detects abnormal data that is outside of normal data. Although it can be viewed as classification in the sense of classifying data, classification There is a difference in the learning method. This anomaly detection (AD) is an algorithm that can be used when it is difficult to obtain normal or abnormal data or when the imbalance between the two data is severe. ) can be learned using. However, such anomaly detection is only used in the image or vision field, and has been difficult to utilize in the voice recognition field due to the nature of the task.

Republic of Korea Patent No. 10-1831078 (2018.02.13) discloses a voice activation detection method and device.

The purpose of the present disclosure is to provide a method for detecting voice activity in input audio data based on anomaly detection.

Meanwhile, the technical problem to be achieved by the present disclosure is not limited to the technical problems mentioned above, and may include various technical problems within the scope of what is apparent to those skilled in the art from the contents described below.

A method performed by a computing device according to an embodiment of the present disclosure for realizing the above-described problem is disclosed. The method includes inputting audio data into a voice activity detection (VAD) model; Predicting whether the audio data is classified as normal data or abnormal data using the VAD model; And when the audio data is classified as normal data, using the VAD model, determining that the audio data corresponds to voice data.

In one embodiment, the method may further include performing speech recognition (STT; Speech-to-Text) on the audio data when determining that the audio data corresponds to voice data.

In one embodiment, the VAD model is a model learned using supervised learning or unsupervised learning based on audio data for training, and the audio data for training may include only voice data. .

In one embodiment, the VAD model includes: a first type of VAD model that detects voice activation using a feature space; Alternatively, it may include at least one VAD model among a second type of VAD model that detects voice activity using reconstruction error.

In one embodiment, the step of predicting whether the audio data is classified as normal data or abnormal data includes extracting a feature vector of the audio data using the first type of VAD model, and extracting a feature vector of the audio data. Projecting the feature vector into a feature space; and determining that the audio data is classified as normal data when the feature vector of the audio data is projected to a predetermined region on the feature space, wherein the predetermined region is a predetermined region based only on voice data. can correspond to .

In one embodiment, predicting whether the audio data is classified as normal data or abnormal data includes encoding the audio data using the second type of VAD model to generate a latent vector of the audio data. After generating, decoding a latent vector of the audio data to generate an output corresponding to the audio data; calculating a restoration error between the audio data and the output; and determining that the audio data is classified as normal data when the calculated reconstruction error is smaller than a predetermined threshold.

According to an embodiment of the present disclosure for realizing the above-described task, a method of learning a voice activity detection (VAD) model performed by a computing device is disclosed. The method includes obtaining audio data for training that includes only voice data; Using the VAD model, generating a prediction result regarding whether the voice data is classified as normal data or abnormal data; And learning the VAD model using a loss function that reduces the loss value when the voice data is predicted to be normal data and increases the loss value when the voice data is predicted to be abnormal data. It can be included.

In one embodiment, generating a prediction result regarding whether the speech data is classified as normal data or abnormal data utilizes a first type of VAD model to transform the feature vector of the speech data into a feature space. generating a prediction result regarding whether the voice data is classified as normal data or abnormal data, based on the projection on the image; Or at least one step of generating a prediction result regarding whether the voice data is classified as normal data or abnormal data, based on a restoration error of the voice data, using a second type of VAD model. may include.

In one embodiment, the loss function for the first type of VAD model has a loss value that decreases when a plurality of feature vectors of a plurality of voice data form a cluster in the feature space, and the loss function for the first type of VAD model is When feature vectors do not form clusters in the feature space, the loss value may be designed to increase.

In one embodiment, the loss function for the first type of VAD model is such that the loss value is reduced when the feature vector of the speech data is projected onto a predetermined region in the feature space, and the additional feature vector is calculated from the predetermined region. It can be additionally designed so that the loss value increases if the projection is outside of the determined area.

In one embodiment, the loss function for the second type of VAD model may be designed so that the loss value decreases as the restoration error of the voice data decreases, and the loss value increases as the restoration error of the voice data increases. there is.

According to an embodiment of the present disclosure for realizing the above-described object, a computer program stored in a computer-readable storage medium is disclosed. The computer program, when executed by one or more processors, causes the one or more processors to perform the following operations for detecting voice activity in audio data, the operations comprising: voice activity detection (VAD) in audio data; Voice activity detection) Actions input to the model; Predicting whether the audio data is classified as normal data or abnormal data using the VAD model; and, when the audio data is classified as normal data, determining that the audio data corresponds to voice data using the VAD model.

A computing device according to an embodiment of the present disclosure for realizing the above-described problem is disclosed. The device includes at least one processor; and a memory, wherein the at least one processor is configured to: input audio data into a voice activity detection (VAD) model; Using the VAD model, predict whether the audio data is classified as normal data or abnormal data; And, by using the VAD model, if the audio data is classified as normal data, it may be configured to determine that the audio data corresponds to voice data.

The present disclosure may provide a method for detecting voice activity in input audio data based on anomaly detection. For example, the present disclosure provides a technology for detecting voice activity based on anomaly detection using a Voice Activity Detection (VAD) model learned based on training audio data containing only voice data. You can. Therefore, the present disclosure provides speech data that is relatively easy to obtain, even if non-speech data (e.g., all data other than human speech, various noises) is lacking or there are not many types of non-speech data. By learning with normal data, the performance of the VAD model can be guaranteed.

Additionally, the present disclosure can provide a VAD model capable of responding to various types of non-speech data. For example, the VAD model according to an embodiment of the present disclosure can classify all types of audio signals other than human speech as non-speech data, so all non-speech data on the learning data Even if types are not included, it can respond to various types of non-voice data.

Meanwhile, the effects of the present disclosure are not limited to the effects mentioned above, and various effects may be included within the range apparent to those skilled in the art from the contents described below.

1 is a block diagram of a computing device that detects voice activity in audio data according to an embodiment of the present disclosure.
Figure 2 is a schematic diagram showing a network function according to an embodiment of the present disclosure.
3 is a flowchart of a method for detecting voice activity in audio data according to an embodiment of the present disclosure.
FIG. 4 is a diagram schematically showing the learning operation of a first type of VAD model that detects voice activity using a feature space, according to an embodiment of the present disclosure.
Figure 5 schematically shows the operation of predicting whether audio data is classified as normal data or abnormal data using a first type of VAD model, according to an embodiment of the present disclosure. It is a drawing.
FIG. 6 is a diagram schematically showing the learning operation of a second type of VAD model that detects voice activity by utilizing reconstruction error according to an embodiment of the present disclosure.
Figure 7 schematically shows the operation of predicting whether audio data is classified as normal data or abnormal data using a second type of VAD model, according to an embodiment of the present disclosure. It is a drawing.
Figure 8 is a flowchart of a method of learning a voice activity detection (VAD) model according to an embodiment of the present disclosure.
9 is a brief, general schematic diagram of an example computing environment in which embodiments of the present disclosure may be implemented.

Various embodiments are now described with reference to the drawings. In this specification, various descriptions are presented to provide an understanding of the disclosure. However, it is clear that these embodiments may be practiced without these specific descriptions.

As used herein, the terms “component,” “module,” “system,” and the like refer to a computer-related entity, hardware, firmware, software, a combination of software and hardware, or an implementation of software. For example, a component may be, but is not limited to, a process running on a processor, a processor, an object, a thread of execution, a program, and/or a computer. For example, both an application running on a computing device and the computing device can be a component. One or more components may reside within a processor and/or thread of execution. A component may be localized within one computer. A component may be distributed between two or more computers. Additionally, these components can execute from various computer-readable media having various data structures stored thereon. Components can transmit signals, for example, with one or more data packets (e.g., data and/or signals from one component interacting with other components in a local system, a distributed system, to other systems and over a network such as the Internet). Depending on the data being transmitted, they may communicate through local and/or remote processes.

Additionally, the term “or” is intended to mean an inclusive “or” and not an exclusive “or.” That is, unless otherwise specified or clear from context, “X utilizes A or B” is intended to mean one of the natural implicit substitutions. That is, either X uses A; X uses B; Or, if X uses both A and B, “X uses A or B” can apply to either of these cases. Additionally, the term “and/or” as used herein should be understood to refer to and include all possible combinations of one or more of the related listed items.

Additionally, the terms “comprise” and/or “comprising” should be understood to mean that the corresponding feature and/or element is present. However, the terms “comprise” and/or “comprising” should be understood as not excluding the presence or addition of one or more other features, elements and/or groups thereof. Additionally, unless otherwise specified or the context is clear to indicate a singular form, the singular terms herein and in the claims should generally be construed to mean “one or more.”

And, the term “at least one of A or B” should be interpreted to mean “a case containing only A,” “a case containing only B,” and “a case of combining A and B.”

Those skilled in the art will additionally recognize that the various illustrative logical blocks, components, modules, circuits, means, logic, and algorithm steps described in connection with the embodiments disclosed herein may be implemented using electronic hardware, computer software, or a combination of both. It must be recognized that it can be implemented with To clearly illustrate the interchangeability of hardware and software, various illustrative components, blocks, configurations, means, logics, modules, circuits, and steps have been described above generally in terms of their functionality. Whether such functionality is implemented in hardware or software will depend on the specific application and design constraints imposed on the overall system. A skilled technician can implement the described functionality in a variety of ways for each specific application. However, such implementation decisions should not be construed as causing a departure from the scope of the present disclosure.

The description of the presented embodiments is provided to enable anyone skilled in the art to use or practice the present invention. Various modifications to these embodiments will be apparent to those skilled in the art. The general principles defined herein may be applied to other embodiments without departing from the scope of the disclosure. Therefore, the present invention is not limited to the embodiments presented herein. The present invention is to be interpreted in the broadest scope consistent with the principles and novel features presented herein.

In this disclosure, network function, artificial neural network, and neural network may be used interchangeably.

1 is a block diagram of a computing device for detecting voice activity in audio data according to an embodiment of the present disclosure.

The configuration of the computing device 100 shown in FIG. 1 is only a simplified example. In one embodiment of the present disclosure, the computing device 100 may include different configurations for performing the computing environment of the computing device 100, and only some of the disclosed configurations may configure the computing device 100.

The computing device 100 may include a processor 110, a memory 130, and a network unit 150.

The processor 110 may be composed of one or more cores, and may include a central processing unit (CPU), a general purpose graphics processing unit (GPGPU), and a tensor processing unit (TPU) of a computing device. unit) may include a processor for data analysis and deep learning. The processor 110 may read a computer program stored in the memory 130 and perform data processing for machine learning according to an embodiment of the present disclosure. According to an embodiment of the present disclosure, the processor 110 may perform an operation for learning a neural network. The processor 110 is used for learning neural networks, such as processing input data for learning in deep learning (DL), extracting features from input data, calculating errors, and updating the weights of the neural network using backpropagation. Calculations can be performed. At least one of the CPU, GPGPU, and TPU of the processor 110 may process learning of the network function. For example, CPU and GPGPU can work together to process learning of network functions and data classification using network functions. Additionally, in one embodiment of the present disclosure, the processors of a plurality of computing devices can be used together to process learning of network functions and data classification using network functions. Additionally, a computer program executed in a computing device according to an embodiment of the present disclosure may be a CPU, GPGPU, or TPU executable program.

According to an embodiment of the present disclosure, the memory 130 may store any type of information generated or determined by the processor 110 and any type of information received by the network unit 150.

According to an embodiment of the present disclosure, the memory 130 is a flash memory type, hard disk type, multimedia card micro type, or card type memory (e.g. (e.g. SD or -Only Memory), and may include at least one type of storage medium among magnetic memory, magnetic disk, and optical disk. The computing device 100 may operate in connection with web storage that performs a storage function of the memory 130 on the Internet. The description of the memory described above is merely an example, and the present disclosure is not limited thereto.

The network unit 150 according to an embodiment of the present disclosure includes Public Switched Telephone Network (PSTN), x Digital Subscriber Line (xDSL), Rate Adaptive DSL (RADSL), Multi Rate DSL (MDSL), and VDSL ( A variety of wired communication systems can be used, such as Very High Speed DSL), Universal Asymmetric DSL (UADSL), High Bit Rate DSL (HDSL), and Local Area Network (LAN).

In addition, the network unit 150 presented in this specification includes Code Division Multi Access (CDMA), Time Division Multi Access (TDMA), Frequency Division Multi Access (FDMA), Orthogonal Frequency Division Multi Access (OFDMA), and SC-FDMA ( A variety of wireless communication systems can be used, such as Single Carrier-FDMA) and other systems.

In the present disclosure, the network unit 150 may be configured regardless of communication mode, such as wired or wireless, and may include a local area network (LAN), a personal area network (PAN), or a wide area network (WAN). It can be composed of various communication networks such as Wide Area Network. Additionally, the network may be the well-known World Wide Web (WWW), or may use wireless transmission technology used for short-distance communication, such as Infrared Data Association (IrDA) or Bluetooth.

The techniques described herein can be used in the networks mentioned above, as well as other networks.

Figure 2 is a schematic diagram showing a network function according to an embodiment of the present disclosure.

Throughout this specification, computational model, neural network, network function, and neural network may be used interchangeably. A neural network can generally consist of a set of interconnected computational units, which can be referred to as nodes. These nodes may also be referred to as neurons. A neural network consists of at least one node. Nodes (or neurons) that make up neural networks may be interconnected by one or more links.

Within a neural network, one or more nodes connected through a link may form a relative input node and output node relationship. The concepts of input node and output node are relative, and any node in an output node relationship with one node may be in an input node relationship with another node, and vice versa. As described above, input node to output node relationships can be created around links. One or more output nodes can be connected to one input node through a link, and vice versa.

In a relationship between an input node and an output node connected through one link, the value of the data of the output node may be determined based on the data input to the input node. Here, the link connecting the input node and the output node may have a weight. Weights may be variable and may be varied by the user or algorithm in order for the neural network to perform the desired function. For example, when one or more input nodes are connected to one output node by respective links, the output node is set to the values input to the input nodes connected to the output node and the links corresponding to each input node. The output node value can be determined based on the weight.

As described above, in a neural network, one or more nodes are interconnected through one or more links to form an input node and output node relationship within the neural network. The characteristics of the neural network can be determined according to the number of nodes and links within the neural network, the correlation between the nodes and links, and the value of the weight assigned to each link. For example, if the same number of nodes and links exist and two neural networks with different weight values of the links exist, the two neural networks may be recognized as different from each other.

A neural network may consist of a set of one or more nodes. A subset of nodes that make up a neural network can form a layer. Some of the nodes constituting the neural network may form one layer based on the distances from the first input node. For example, a set of nodes with a distance n from the initial input node may constitute n layers. The distance from the initial input node can be defined by the minimum number of links that must be passed to reach the node from the initial input node. However, this definition of a layer is arbitrary for explanation purposes, and the order of a layer within a neural network may be defined in a different way than described above. For example, a layer of nodes may be defined by distance from the final output node.

The initial input node may refer to one or more nodes in the neural network through which data is directly input without going through links in relationships with other nodes. Alternatively, in a neural network network, in the relationship between nodes based on links, it may mean nodes that do not have other input nodes connected by links. Similarly, the final output node may refer to one or more nodes that do not have an output node in their relationship with other nodes among the nodes in the neural network. Additionally, hidden nodes may refer to nodes constituting a neural network other than the first input node and the last output node.

The neural network according to an embodiment of the present disclosure is a neural network in which the number of nodes in the input layer may be the same as the number of nodes in the output layer, and the number of nodes decreases and then increases again as it progresses from the input layer to the hidden layer. You can. In addition, the neural network according to another embodiment of the present disclosure may be a neural network in which the number of nodes in the input layer may be less than the number of nodes in the output layer, and the number of nodes decreases as it progresses from the input layer to the hidden layer. there is. In addition, the neural network according to another embodiment of the present disclosure may be a neural network in which the number of nodes in the input layer may be greater than the number of nodes in the output layer, and the number of nodes increases as it progresses from the input layer to the hidden layer. You can. A neural network according to another embodiment of the present disclosure may be a neural network that is a combination of the above-described neural networks.

A deep neural network (DNN) may refer to a neural network that includes multiple hidden layers in addition to the input layer and output layer. Deep neural networks allow you to identify latent structures in data. In other words, it is possible to identify the potential structure of a photo, text, video, voice, or music (e.g., what object is in the photo, what the content and emotion of the text are, what the content and emotion of the voice are, etc.) . Deep neural networks include convolutional neural networks (CNN), recurrent neural networks (RNN), auto encoders, generative adversarial networks (GAN), and restricted Boltzmann machines (RBM). machine), deep belief network (DBN), Q network, U network, Siamese network, Generative Adversarial Network (GAN), etc. The description of the deep neural network described above is only an example and the present disclosure is not limited thereto.

In one embodiment of the present disclosure, the network function may include an autoencoder. An autoencoder may be a type of artificial neural network to output output data similar to input data. The autoencoder may include at least one hidden layer, and an odd number of hidden layers may be placed between input and output layers. The number of nodes in each layer may be reduced from the number of nodes in the input layer to an intermediate layer called the bottleneck layer (encoding), and then expanded symmetrically and reduced from the bottleneck layer to the output layer (symmetrical to the input layer). Autoencoders can perform nonlinear dimensionality reduction. The number of input layers and output layers can be corresponded to the dimension after preprocessing of the input data. In an auto-encoder structure, the number of nodes in the hidden layer included in the encoder may have a structure that decreases as the distance from the input layer increases. If the number of nodes in the bottleneck layer (the layer with the fewest nodes located between the encoder and decoder) is too small, not enough information may be conveyed, so if it is higher than a certain number (e.g., more than half of the input layers, etc.) ) may be maintained.

Neural networks learn by at least one of supervised learning, self-supervised learning, unsupervised learning, semi-supervised learning, or reinforcement learning. It can be. Learning of a neural network may be a process of applying knowledge for the neural network to perform a specific operation to the neural network.

Neural networks can be trained to minimize output errors. In neural network learning, learning data is repeatedly input into the neural network, the output of the neural network and the error of the target for the learning data are calculated, and the error of the neural network is transferred from the output layer of the neural network to the input layer in the direction of reducing the error. This is the process of updating the weight of each node in the neural network through backpropagation. In the case of teacher learning, learning data in which the correct answer is labeled in each learning data is used (i.e., labeled learning data), and in the case of non-teacher learning, the correct answer may not be labeled in each learning data. That is, for example, in the case of teacher learning regarding data classification, the learning data may be data in which each learning data is labeled with a category. Labeled training data is input to the neural network, and the error can be calculated by comparing the output (category) of the neural network with the label of the training data. As another example, in the case of non-teachable learning for data classification, the error can be calculated by comparing the input training data with the neural network output. The calculated error is backpropagated in the reverse direction (i.e., from the output layer to the input layer) in the neural network, and the connection weight of each node in each layer of the neural network can be updated according to backpropagation. The amount of change in the connection weight of each updated node may be determined according to the learning rate. The neural network's calculation of input data and backpropagation of errors can constitute a learning cycle (epoch). The learning rate may be applied differently depending on the number of repetitions of the learning cycle of the neural network. For example, in the early stages of neural network training, a high learning rate can be used to increase efficiency by allowing the neural network to quickly achieve a certain level of performance, and in the later stages of training, a low learning rate can be used to increase accuracy.

In the learning of neural networks, the training data can generally be a subset of real data (i.e., the data to be processed using the learned neural network), and thus the error for the training data is reduced, but the error for the real data is reduced. There may be an incremental learning cycle. Overfitting is a phenomenon in which errors in actual data increase due to excessive learning on training data. For example, a phenomenon in which a neural network that learned a cat by showing a yellow cat fails to recognize that it is a cat when it sees a non-yellow cat may be a type of overfitting. Overfitting can cause errors in machine learning algorithms to increase. To prevent such overfitting, various optimization methods can be used. To prevent overfitting, methods such as increasing the learning data, regularization, dropout to disable some of the network nodes during the learning process, and use of a batch normalization layer can be applied. You can.

Speech recognition (STT or ASR; Speech To Text, or Automatic Speech Recognition) according to an embodiment of the present disclosure is a dictation technology that converts voice into text. In other words, speech recognition (STT) is a technology that generates text (correct spelling) corresponding to speech. The input of this voice recognition (STT) may include at least one of a voice signal, a spectrogram converted from a voice signal, or a voice feature. Additionally, the output of speech recognition (STT) is text in the form of a string. Meanwhile, the speech recognition (STT) model can be implemented in various types of models, including a neural network model. Additionally, speech recognition (STT) models can be divided into modular and non-modular end-to-end (e2e) methods depending on how they are implemented. Here, the modular method is divided into an acoustic model (a model that indicates what form a voice signal can be expressed in), a language model (a model that assigns the probability of occurrence to words based on given sentences and words), and a pronunciation dictionary. It may include, but is not limited to, traditional models that perform recognition (e.g., some models of Kaldi toolkit-based ASR models, Hybrid-ASR models, etc.). On the other hand, non-modularized methods mainly refer to e2e models (e.g., transformer-based encoder decoder models, etc.), and models can be created by learning a lot of data without having sub-modules. Meanwhile, the Beam Search technique is a representative decoding technique. The Beam Search technique does not predict only one word that is closest to the correct answer depending on the situation, but opens up various possibilities and considers the entire sentence to find the most optimal one. This is a way to find the correct answer.

Voice Activity Detection (VAD) according to an embodiment of the present disclosure is a technology that determines whether voice activity is detected in audio data (audio signal, audio stream, etc.). Voice activation detection (VAD) may be subject to preprocessing techniques such as noise removal and speech enhancement before applying speech recognition (STT). In relation to this VAD implementation, in the past, a supervised learning-based binary classification algorithm or a probability distribution-based classification algorithm was mainly used.

Anomaly detection (AD) according to an embodiment of the present disclosure is a technology for detecting and detecting abnormal data that is outside of normal data, and can be considered classification in terms of classifying data. However, the learning method is different from classification. This anomaly detection (AD) is an algorithm that can be used when it is difficult to obtain normal or abnormal data or when the imbalance between the two data is severe. ) can be learned using. However, such anomaly detection is only used in the image or vision field, and has been difficult to utilize in the voice recognition field due to the nature of the task. The present disclosure proposes a technique for applying such anomaly detection particularly to VAD. Specifically, the present disclosure classifies human voices, which are relatively easy to obtain, as normal, and other sounds (i.e., sounds other than human voices) as abnormal, so as to detect voice activity in audio data. I would like to explain how to detect it.

On the other hand, since binary classification projects two data sets into the feature space and then classifies the data based on the determined plane, non-speech data not used for learning may not be properly projected into the feature space, and the resulting The classification result may be incorrect. In addition, distribution based classification algorithms may also have difficulty obtaining a reliable probability distribution in situations where sufficient non-speech data cannot be secured, which may reduce the accuracy of classification. Alternatively, when anomaly detection is applied to VAD, learning can be performed using only speech data, so the performance of the VAD model can be guaranteed even in situations where it is difficult to secure sufficient non-speech data. In addition, the anomaly detection-based VAD model can respond relatively robustly to various types of non-voice data that are not defined or identified in advance, and can maintain high voice activity detection performance.

Embodiments of the present disclosure, which will be discussed below, relate to a method of performing voice activity detection (VAD) on input audio data using anomaly detection (AD) technology. As seen above, in these embodiments, even if non-speech data (e.g., all non-speech data, various types of noise, etc.) is lacking or there are not many types, speech data that is relatively easy to obtain is used as normal data. By learning with data, VAD technology with excellent performance can be implemented.

3 is a flowchart of a method for detecting voice activity in audio data according to an embodiment of the present disclosure.

According to an embodiment of the present disclosure, the computing device 100 may input audio data into a voice activity detection (VAD) model (S110). For example, audio data may include a voice signal or a spectrogram converted from a voice signal. Additionally, audio data may be a token generated by dividing an audio signal. Additionally, audio data may include voice data and data other than voice (eg, noise data).

According to an embodiment of the present disclosure, the computing device 100 may utilize a VAD model to predict whether the audio data is classified as normal data or abnormal data (S120). . Here, the VAD model is a model learned using supervised learning or unsupervised learning based on audio data for learning, and the audio data for learning may include “voice data only.” For reference, the computing device 100 classifies voice data as normal data and other sounds (i.e., sounds other than human voices) as abnormal data to detect voice activity in the audio data. can do. In addition, the VAD model is either a first type VAD model that detects voice activity using a feature space or a second type VAD model that detects voice activity using a reconstruction error. It may contain at least one VAD model.

According to one embodiment of the present disclosure, the computing device 100 utilizes a VAD model (e.g., a first type VAD model or a second type VAD model) to determine if the audio data is classified as normal data. It may be determined that it corresponds to data (S130).

For example, the computing device 100 considers voice data uttered by a person as normal data, and when a random audio token with a short length is given to the learned VAD model, it determines whether the audio token is normal data or abnormal data. Based on the classification method, only the part (voice data) actually spoken by a person can be detected from the input audio. In addition, the computing device 100 classifies all sounds (i.e., noise data) except human voice data (i.e., normal data) as abnormal data, so it is necessary to separately learn noise data or collect various noise data. It is possible to create a VAD model that is robust to various types of noise.

Hereinafter, the operation of predicting whether the audio data is classified as normal data or abnormal data using the first type of VAD model through FIGS. 4 and 5 and FIGS. 6 and Through FIG. 7, the operation of predicting whether the audio data is classified as normal data or abnormal data using the second type of VAD model will be described in more detail. However, the VAD model of the present disclosure is not limited to the first type of VAD model or the second type of VAD model, and VAD models utilizing various anomaly detection techniques may be used. For example, the VAD model of the present disclosure includes not only a first type of VAD model and a second type of VAD model, but also VAD is performed based on methods used for anomaly detection such as Normalizing, GAN (Generative Adversarial Networks), or Representation. A model that does this may be included.

FIG. 4 is a diagram schematically showing the learning operation of a first type of VAD model that detects voice activity by utilizing feature space according to an embodiment of the present disclosure, and FIG. 5 is an embodiment of the present disclosure. According to an example, this is a diagram schematically showing the operation of predicting whether audio data is classified as normal data or abnormal data using a first type of VAD model.

Referring to FIG. 4, the first type of VAD model is a model that detects voice activity by utilizing a feature space. The first type of VAD model may be a model learned based on training audio data 10 containing “voice data only”. In other words, the first type of VAD model can perform learning by using “voice data only” as normal data during learning. The first type of VAD model extracts feature vectors from voice data, which is normal data, input through a feature extraction network, and stores the extracted feature vectors in feature space. It can be projected.

At this time, the first type of VAD model utilizes a loss function that reduces the loss value when the learning voice data is predicted to be the normal data and increases the loss value when the learning voice data is predicted to be the abnormal data. , learning can be performed. In addition, the loss function for the first type of VAD model has a loss value that decreases when a plurality of feature vectors of a plurality of learning voice data form a cluster on the feature space, and the plurality of feature vectors If they do not form a cluster in the feature space, the loss value may be designed to increase. For example, a cluster area may be formed when the mutual distance between a plurality of feature vectors of a plurality of learning voice data is within a preset distance. Additionally, the cluster area may be formed when a plurality of feature vectors of a plurality of learning voice data form a certain cluster within a feature space. Meanwhile, after the cluster area is formed, the loss function for the first type of VAD model may be additionally designed in such a way that feature vectors of additional speech data for training are projected onto the cluster area. For example, the loss function for the first type of VAD model has a loss value that decreases when the feature vector of the additional voice data for training is projected to the cluster area on the feature space, and the features of the additional voice data for training It may be additionally designed so that the loss value increases when the vector is projected outside the cluster area.

According to an embodiment of the present disclosure, the computing device 100 utilizes the first VAD model (i.e., a model based on feature extraction learned through FIG. 4) to determine if the input audio data is normal. ) It is possible to predict whether it will be classified as data or abnormal data (S120). More specifically, referring to FIG. 5 , the computing device 100 may input audio data 20 into a first type of VAD model. Here, the audio data 20 may include voice data and data other than voice data (eg, noise data). Additionally, the computing device 100 may extract a feature vector of the audio data 20 using the first type of VAD model. Additionally, the computing device 200 may project the extracted feature vector into a feature space. Additionally, the computing device 100 may determine that the audio data is classified as normal data when the feature vector of the audio data is projected to a predetermined area on the feature space. Here, the predetermined area may correspond to a cluster area formed by a plurality of feature vectors of a plurality of learning voice data during the learning process. In addition, the predetermined area is an area designed to project feature vectors of speech data (for example, a loss function is used to determine a predetermined area on a feature space and to project feature vectors of speech data for learning into the predetermined area. It can correspond to the designed area).

According to an embodiment of the present disclosure, the computing device 100 may utilize the first VAD model to determine that the audio data corresponds to voice data when the audio data is classified as normal data (S130). In other words, the computing device 100 may generate a prediction result regarding whether the voice data is classified as normal data or abnormal data, based on projecting the feature vector of the voice data onto the feature space. You can. Alternatively, the computing device 100 may classify the voice data as normal data when the probability that the feature vector of the voice data included in the audio data is included in the feature space is greater than or equal to a threshold. Additionally, the computing device 100 may classify the voice data as abnormal data when the probability that the feature vector of the voice data included in the audio data is included in the feature space is less than a threshold. That is, the computing device 100 utilizes a first type of VAD model, and when voice data included in the audio data is included in a cluster area or a predetermined area in the feature space, the audio data is It may be determined that it corresponds to voice data.

FIG. 6 is a diagram schematically showing the learning operation of a second type of VAD model that detects voice activity by utilizing reconstruction error according to an embodiment of the present disclosure, and FIG. 7 is an embodiment of the present disclosure. According to an example, this is a diagram schematically showing the operation of predicting whether audio data is classified as normal data or abnormal data using a second type of VAD model.

Referring to FIG. 6, the second type of VAD model is a model that detects voice activity by utilizing reconstruction error. A second type of VAD model may be a model learned based on training audio data 10 containing “voice data only”. In other words, the second type of VAD model can perform learning by using “only voice data” as normal data during learning. The second type of VAD model can perform learning using an autoencoder that includes an encoder and a decoder. When learning the second type of VAD model, the encoder of the autoencoder receives training audio data (10) containing “voice data only”, extracts a latent vector, and then decoders ( You can restore data through decoder). For reference, the autoencoder cannot properly encode and decode sounds other than the learned voice data.

At this time, the second type of VAD model utilizes a loss function that reduces the loss value when the voice data for training is predicted to be normal data and increases the loss value when the voice data for training is predicted to be abnormal data, Learning can be done. More specifically, the loss function for the second type of VAD model can be designed so that the loss value decreases as the restoration error for the training voice data decreases, and the loss value increases as the restoration error of the learning voice data increases. there is. As an example, the loss function for the second type of VAD model can be learned such that ∥input - output∥ ^-2 - converges to 0.

According to an embodiment of the present disclosure, the computing device 100 uses a second VAD model (i.e., an autoencoder learned through FIG. 6) to classify input audio data as normal data. It is possible to predict whether or not the data is classified as abnormal data (S120). In other words, the computing device 100 may generate a prediction result regarding whether the voice data is classified as normal data or abnormal data based on the restoration error of the voice data. More specifically, referring to FIG. 7 , the computing device 100 may input audio data 20 into a second type of VAD model. Here, the audio data 20 may include voice data and data other than voice data (eg, noise data). Additionally, the computing device 100 may encode the audio data using the second type of VAD model to generate a latent vector of the audio data. Additionally, the computing device 100 may decode a latent vector of the audio data and generate output corresponding to the audio data. Additionally, the computing device 100 may calculate a restoration error between the audio data and the output. For example, the restoration error can be expressed as ∥input - output ∥ ^-2 -.

According to an embodiment of the present disclosure, the computing device 100 may utilize the second VAD model to determine that the audio data corresponds to voice data when the audio data is classified as normal data (S130). That is, the computing device 100 may determine that the audio data is classified as normal data when the calculated reconstruction error is smaller than a predetermined threshold (eg, ∥input - output ∥ ^{- 2} < threshold). Additionally, the computing device 100 may determine that the audio data is classified as abnormal data when the calculated reconstruction error is greater than a predetermined threshold (eg, ∥input - output ∥ - ² > threshold).

According to an embodiment of the present disclosure, when the computing device 100 determines that the audio data corresponds to voice data, the computing device 100 may perform speech recognition (STT; Speech-to-Text) on the audio data. In other words, the computing device 100 can perform speech recognition (STT; Speech-to-Text) by taking audio data determined as voice data as input and outputting it as text in the form of a string. The computing device 100 uses audio data (i.e., voice data) previously classified as normal data through FIGS. 4 to 7 to more simply detect the voice part, thereby enabling more accurate speech recognition (STT). Text) can be performed.

Figure 8 is a flowchart of a method of learning a voice activity detection (VAD) model according to an embodiment of the present disclosure.

Referring to FIG. 8, a method of learning a voice activity detection (VAD) model according to an embodiment of the present disclosure includes acquiring audio data for training including only voice data (S210), the VAD model Using , generating a prediction result regarding whether the voice data is classified as normal data or abnormal data (S220), and if the voice data is predicted to be normal data, reducing the loss value, and If the voice data is predicted to be abnormal data, a step (S230) of learning the VAD model may be included by using a loss function that increases the loss value. Additionally, the method of learning a voice activity detection model according to an embodiment of the present disclosure may be performed by the computing device 100.

The step S210 is a step of acquiring audio data for learning that includes only voice data.

The step S220 is a step of generating a prediction result regarding whether the voice data is classified as normal data or abnormal data using the VAD model. This step S220 determines whether the speech data is classified as normal data or abnormal data based on projecting the feature vector of the speech data onto the feature space using a first type of VAD model. Generating a prediction result or utilizing a second type of VAD model to generate a prediction result regarding whether the speech data is classified as normal data or abnormal data based on a restoration error of the speech data. It may include at least one of the steps. Here, the loss function for the first type of VAD model has a loss value that decreases when a plurality of feature vectors of a plurality of voice data form a cluster in the feature space, and the plurality of feature vectors If a cluster is not formed in the feature space, the loss value may be designed to increase. In addition, the loss function for the first type of VAD model has a loss value that decreases when the feature vector of the speech data is projected onto a predetermined region in the feature space, and the additional feature vector is distributed outside the region of the cluster. It can be additionally designed so that the loss value increases when projected. Meanwhile, the loss function for the second type of VAD model may be designed so that the loss value decreases as the restoration error of the voice data decreases, and the loss value increases as the restoration error of the voice data increases.

The step S230 uses a loss function that reduces the loss value when the voice data is predicted to be normal data and increases the loss value when the voice data is predicted to be abnormal data to construct the VAD model. This is the learning stage.

The steps mentioned in the above description may be further divided into additional steps or combined into fewer steps, depending on the implementation of the present disclosure. Additionally, some steps may be omitted or the order between steps may be changed as needed.

According to an embodiment of the present disclosure, a computer-readable medium storing a data structure is disclosed.

Data structure can refer to the organization, management, and storage of data to enable efficient access and modification of data. Data structure can refer to the organization of data to solve a specific problem (e.g., retrieving data, storing data, or modifying data in the shortest possible time). A data structure may be defined as a physical or logical relationship between data elements designed to support a specific data processing function. Logical relationships between data elements may include connection relationships between user-defined data elements. Physical relationships between data elements may include actual relationships between data elements that are physically stored in a computer-readable storage medium (e.g., a persistent storage device). A data structure may specifically include a set of data, relationships between data, and functions or instructions applicable to the data. Effectively designed data structures allow computing devices to perform computations while minimizing the use of the computing device's resources. Specifically, computing devices can increase the efficiency of operations, reading, insertion, deletion, comparison, exchange, and search through effectively designed data structures.

Data structures can be divided into linear data structures and non-linear data structures depending on the type of data structure. A linear data structure may be a structure in which only one piece of data is connected to another piece of data. Linear data structures may include List, Stack, Queue, and Deque. A list can refer to a set of data that has an internal order. The list may include a linked list. A linked list may be a data structure in which data is connected in such a way that each data is connected in a single line with a pointer. In a linked list, a pointer may contain connection information to the next or previous data. Depending on its form, a linked list can be expressed as a singly linked list, a doubly linked list, or a circularly linked list. A stack may be a data listing structure that allows limited access to data. A stack can be a linear data structure in which data can be processed (for example, inserted or deleted) at only one end of the data structure. Data stored in the stack may have a data structure (LIFO-Last in First Out) where the later it enters, the sooner it comes out. A queue is a data listing structure that allows limited access to data. Unlike the stack, it can be a data structure (FIFO-First in First Out) where data stored later is released later. A deck can be a data structure that can process data at both ends of the data structure.

A non-linear data structure may be a structure in which multiple pieces of data are connected behind one piece of data. Nonlinear data structures may include graph data structures. A graph data structure can be defined by vertices and edges, and an edge can include a line connecting two different vertices. Graph data structure may include a tree data structure. A tree data structure may be a data structure in which there is only one path connecting two different vertices among a plurality of vertices included in the tree. In other words, it may be a data structure that does not form a loop in the graph data structure.

Throughout this specification, computational model, neural network, network function, and neural network may be used interchangeably. Below, it is described in a unified manner as a neural network. Data structures may include neural networks. And the data structure including the neural network may be stored in a computer-readable medium. Data structures including neural networks also include data preprocessed for processing by a neural network, data input to the neural network, weights of the neural network, hyperparameters of the neural network, data acquired from the neural network, activation functions associated with each node or layer of the neural network, neural network It may include a loss function for learning. A data structure containing a neural network may include any of the components disclosed above. In other words, the data structure including the neural network includes preprocessed data for processing by the neural network, data input to the neural network, weights of the neural network, hyperparameters of the neural network, data acquired from the neural network, activation functions associated with each node or layer of the neural network, neural network It may be configured to include all or any combination of the loss function for learning. In addition to the configurations described above, a data structure containing a neural network may include any other information that determines the characteristics of the neural network. Additionally, the data structure may include all types of data used or generated in the computational process of a neural network and is not limited to the above. Computer-readable media may include computer-readable recording media and/or computer-readable transmission media. A neural network can generally consist of a set of interconnected computational units, which can be referred to as nodes. These nodes may also be referred to as neurons. A neural network consists of at least one node.

The data structure may include data input to the neural network. A data structure containing data input to a neural network may be stored in a computer-readable medium. Data input to the neural network may include learning data input during the neural network learning process and/or input data input to the neural network on which training has been completed. Data input to the neural network may include data that has undergone pre-processing and/or data subject to pre-processing. Preprocessing may include a data processing process to input data into a neural network. Therefore, the data structure may include data subject to preprocessing and data generated by preprocessing. The above-described data structure is only an example and the present disclosure is not limited thereto.

The data structure may include the weights of the neural network. (In this specification, weights and parameters may be used with the same meaning.) And the data structure including the weights of the neural network may be stored in a computer-readable medium. A neural network may include multiple weights. Weights may be variable and may be varied by the user or algorithm in order for the neural network to perform the desired function. For example, when one or more input nodes are connected to one output node by respective links, the output node is set to the values input to the input nodes connected to the output node and the links corresponding to each input node. Based on the weight, the data value output from the output node can be determined. The above-described data structure is only an example and the present disclosure is not limited thereto.

As an example and not a limitation, the weights may include weights that are changed during the neural network learning process and/or weights for which neural network learning has been completed. Weights that change during the neural network learning process may include weights that change at the start of the learning cycle and/or weights that change during the learning cycle. Weights for which neural network training has been completed may include weights for which a learning cycle has been completed. Therefore, the data structure including the weights of the neural network may include weights that are changed during the neural network learning process and/or the data structure including the weights for which neural network learning has been completed. Therefore, the above-mentioned weights and/or combinations of each weight are included in the data structure including the weights of the neural network. The above-described data structure is only an example and the present disclosure is not limited thereto.

The data structure including the weights of the neural network may be stored in a computer-readable storage medium (e.g., memory, hard disk) after going through a serialization process. Serialization can be the process of converting a data structure into a form that can be stored on the same or a different computing device and later reorganized and used. Computing devices can transmit and receive data over a network by serializing data structures. Data structures containing the weights of a serialized neural network can be reconstructed on the same computing device or on a different computing device through deserialization. The data structure including the weights of the neural network is not limited to serialization. Furthermore, the data structure including the weights of the neural network is a data structure to increase computational efficiency while minimizing the use of computing device resources (e.g., in non-linear data structures, B-Tree, Trie, m-way search tree, AVL tree, Red-Black Tree) may be included. The foregoing is merely an example and the present disclosure is not limited thereto.

The data structure may include hyper-parameters of a neural network. And the data structure including the hyperparameters of the neural network can be stored in a computer-readable medium. A hyperparameter may be a variable that can be changed by the user. Hyperparameters include, for example, learning rate, cost function, number of learning cycle repetitions, weight initialization (e.g., setting the range of weight values subject to weight initialization), Hidden Unit. It may include a number (e.g., number of hidden layers, number of nodes in hidden layers). The above-described data structure is only an example and the present disclosure is not limited thereto.

9 is a brief, general schematic diagram of an example computing environment in which embodiments of the present disclosure may be implemented.

Although the present disclosure has generally been described above as being capable of being implemented by a computing device, those skilled in the art will understand that the present disclosure can be implemented in combination with computer-executable instructions and/or other program modules that can be executed on one or more computers and/or in hardware and software. It will be well known that it can be implemented as a combination.

Typically, program modules include routines, programs, components, data structures, etc. that perform specific tasks or implement specific abstract data types. Additionally, those skilled in the art will understand that the methods of the present disclosure are applicable to single-processor or multiprocessor computer systems, minicomputers, mainframe computers, as well as personal computers, handheld computing devices, microprocessor-based or programmable consumer electronics, etc. It will be appreciated that each of these may be implemented in other computer system configurations, including those capable of operating in conjunction with one or more associated devices.

The described embodiments of the disclosure can also be practiced in distributed computing environments where certain tasks are performed by remote processing devices that are linked through a communications network. In a distributed computing environment, program modules may be located in both local and remote memory storage devices.

Computers typically include a variety of computer-readable media. Computer-readable media can be any medium that can be accessed by a computer, and such computer-readable media includes volatile and non-volatile media, transitory and non-transitory media, removable and non-transitory media. Includes removable media. By way of example, and not limitation, computer-readable media may include computer-readable storage media and computer-readable transmission media. Computer-readable storage media refers to volatile and non-volatile media, transient and non-transitory media, removable and non-removable, implemented in any method or technology for storage of information such as computer readable instructions, data structures, program modules or other data. Includes media. Computer readable storage media may include RAM, ROM, EEPROM, flash memory or other memory technology, CD-ROM, digital video disk (DVD) or other optical disk storage, magnetic cassette, magnetic tape, magnetic disk storage or other magnetic storage. This includes, but is not limited to, a device, or any other medium that can be accessed by a computer and used to store desired information.

A computer-readable transmission medium typically implements computer-readable instructions, data structures, program modules, or other data on a modulated data signal, such as a carrier wave or other transport mechanism. Includes all information delivery media. The term modulated data signal refers to a signal in which one or more of the characteristics of the signal have been set or changed to encode information within the signal. By way of example, and not limitation, computer-readable transmission media includes wired media such as a wired network or direct-wired connection, and wireless media such as acoustic, RF, infrared, and other wireless media. Combinations of any of the above are also intended to be included within the scope of computer-readable transmission media.

An example environment 1100 is shown that implements various aspects of the present disclosure, including a computer 1102, which includes a processing unit 1104, a system memory 1106, and a system bus 1108. do. System bus 1108 couples system components, including but not limited to system memory 1106, to processing unit 1104. Processing unit 1104 may be any of a variety of commercially available processors. Dual processors and other multiprocessor architectures may also be used as processing unit 1104.

System bus 1108 may be any of several types of bus structures that may further be interconnected to a memory bus, peripheral bus, and local bus using any of a variety of commercial bus architectures. System memory 1106 includes read only memory (ROM) 1110 and random access memory (RAM) 1112. The basic input/output system (BIOS) is stored in non-volatile memory 1110, such as ROM, EPROM, and EEPROM, and is a basic input/output system that helps transfer information between components within the computer 1102, such as during startup. Contains routines. RAM 1112 may also include high-speed RAM, such as static RAM, for caching data.

Computer 1102 may also include an internal hard disk drive (HDD) 1114 (e.g., EIDE, SATA)—the internal hard disk drive 1114 may also be configured for external use within a suitable chassis (not shown). Yes - a magnetic floppy disk drive (FDD) 1116 (e.g., for reading from or writing to a removable diskette 1118), and an optical disk drive 1120 (e.g., a CD-ROM for reading the disk 1122 or reading from or writing to other high-capacity optical media such as DVDs). Hard disk drive 1114, magnetic disk drive 1116, and optical disk drive 1120 are connected to system bus 1108 by hard disk drive interface 1124, magnetic disk drive interface 1126, and optical drive interface 1128, respectively. ) can be connected to. The interface 1124 for implementing an external drive includes at least one or both of Universal Serial Bus (USB) and IEEE 1394 interface technologies.

These drives and their associated computer-readable media provide non-volatile storage of data, data structures, computer-executable instructions, and the like. For computer 1102, drive and media correspond to storing any data in a suitable digital format. Although the description of computer-readable media above refers to removable optical media such as HDDs, removable magnetic disks, and CDs or DVDs, those skilled in the art will also recognize removable optical media such as zip drives, magnetic cassettes, flash memory cards, cartridges, etc. It will be appreciated that other types of computer-readable media, such as the like, may also be used in the example operating environment and that any such media may contain computer-executable instructions for performing the methods of the present disclosure.

A number of program modules may be stored in drives and RAM 1112, including an operating system 1130, one or more application programs 1132, other program modules 1134, and program data 1136. All or portions of the operating system, applications, modules and/or data may also be cached in RAM 1112. It will be appreciated that the present disclosure may be implemented on various commercially available operating systems or combinations of operating systems.

A user may enter commands and information into computer 1102 through one or more wired/wireless input devices, such as a keyboard 1138 and a pointing device such as mouse 1140. Other input devices (not shown) may include microphones, IR remote controls, joysticks, game pads, stylus pens, touch screens, etc. These and other input devices are connected to the processing unit 1104 through an input device interface 1142, which is often connected to the system bus 1108, but may also include a parallel port, an IEEE 1394 serial port, a game port, a USB port, an IR interface, It can be connected by other interfaces, etc.

A monitor 1144 or other type of display device is also connected to system bus 1108 through an interface, such as a video adapter 1146. In addition to monitor 1144, computers typically include other peripheral output devices (not shown) such as speakers, printers, etc.

Computer 1102 may operate in a networked environment using logical connections to one or more remote computers, such as remote computer(s) 1148, via wired and/or wireless communications. Remote computer(s) 1148 may be a workstation, computing device computer, router, personal computer, portable computer, microprocessor-based entertainment device, peer device, or other conventional network node, and is generally connected to computer 1102. For simplicity, only memory storage device 1150 is shown, although it includes many or all of the components described. The logical connections depicted include wired/wireless connections to a local area network (LAN) 1152 and/or a larger network, such as a wide area network (WAN) 1154. These LAN and WAN networking environments are common in offices and companies and facilitate enterprise-wide computer networks, such as intranets, all of which can be connected to a worldwide computer network, such as the Internet.

When used in a LAN networking environment, computer 1102 is connected to local network 1152 through wired and/or wireless communication network interfaces or adapters 1156. Adapter 1156 may facilitate wired or wireless communication to LAN 1152, which also includes a wireless access point installed thereon for communicating with wireless adapter 1156. When used in a WAN networking environment, the computer 1102 may include a modem 1158 or be connected to a communicating computing device on the WAN 1154 or to establish communications over the WAN 1154, such as via the Internet. Have other means. Modem 1158, which may be internal or external and a wired or wireless device, is coupled to system bus 1108 via serial port interface 1142. In a networked environment, program modules described for computer 1102, or portions thereof, may be stored in remote memory/storage device 1150. It will be appreciated that the network connections shown are exemplary and that other means of establishing a communications link between computers may be used.

Computer 1102 may be associated with any wireless device or object deployed and operating in wireless communications, such as a printer, scanner, desktop and/or portable computer, portable data assistant (PDA), communications satellite, wirelessly detectable tag. Performs actions to communicate with any device or location and telephone. This includes at least Wi-Fi and Bluetooth wireless technologies. Accordingly, communication may be a predefined structure as in a conventional network or may simply be ad hoc communication between at least two devices.

Wi-Fi (Wireless Fidelity) allows connection to the Internet, etc. without wires. Wi-Fi is a wireless technology, like cell phones, that allows these devices, such as computers, to send and receive data indoors and outdoors, anywhere within the coverage area of a base station. Wi-Fi networks use wireless technology called IEEE 802.11 (a, b, g, etc.) to provide secure, reliable, and high-speed wireless connections. Wi-Fi can be used to connect computers to each other, the Internet, and wired networks (using IEEE 802.3 or Ethernet). Wi-Fi networks can operate in the unlicensed 2.4 and 5 GHz wireless bands, for example, at data rates of 11 Mbps (802.11a) or 54 Mbps (802.11b), or in products that include both bands (dual band). .

Those skilled in the art will understand that information and signals may be represented using any of a variety of different technologies and techniques. For example, data, instructions, commands, information, signals, bits, symbols and chips that may be referenced in the above description include voltages, currents, electromagnetic waves, magnetic fields or particles, optical fields. It can be expressed by particles or particles, or any combination thereof.

Those skilled in the art will understand that the various illustrative logical blocks, modules, processors, means, circuits and algorithm steps described in connection with the embodiments disclosed herein may be used in electronic hardware, (for convenience) It will be understood that it may be implemented by various forms of program or design code (referred to herein as software) or a combination of both. To clearly illustrate this interoperability of hardware and software, various illustrative components, blocks, modules, circuits and steps have been described above generally with respect to their functionality. Whether this functionality is implemented as hardware or software depends on the specific application and design constraints imposed on the overall system. A person skilled in the art of this disclosure may implement the described functionality in various ways for each specific application, but such implementation decisions should not be construed as departing from the scope of this disclosure.

The various embodiments presented herein may be implemented as a method, apparatus, or article of manufacture using standard programming and/or engineering techniques. The term article of manufacture includes a computer program, carrier, or media accessible from any computer-readable storage device. For example, computer-readable storage media include magnetic storage devices (e.g., hard disks, floppy disks, magnetic strips, etc.), optical disks (e.g., CDs, DVDs, etc.), smart cards, and flash. Includes, but is not limited to, memory devices (e.g., EEPROM, cards, sticks, key drives, etc.). Additionally, various storage media presented herein include one or more devices and/or other machine-readable media for storing information.

It is to be understood that the specific order or hierarchy of steps in the processes presented is an example of illustrative approaches. It is to be understood that the specific order or hierarchy of steps in processes may be rearranged within the scope of the present disclosure, based on design priorities. The appended method claims present elements of the various steps in a sample order but are not meant to be limited to the particular order or hierarchy presented.

The description of the presented embodiments is provided to enable any person skilled in the art to make or use the present disclosure. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be applied to other embodiments without departing from the scope of the disclosure. Thus, the present disclosure is not limited to the embodiments presented herein but is to be interpreted in the broadest scope consistent with the principles and novel features presented herein.

Claims (13)

1. A method of detecting voice activity in audio data, performed by a computing device, comprising:
generating a plurality of audio data based on dividing the audio signal into a plurality of tokens;
Inputting the plurality of audio data into a voice activity detection (VAD) model using a feature space;
Extracting a feature vector from each of the plurality of audio data using the VAD model and projecting the extracted feature vector into a feature space;
Among the plurality of audio data, filtering audio data in which the feature vector is projected onto a predetermined area in the feature space, where the predetermined area corresponds to a predetermined area based only on voice data; and
performing speech recognition (STT; Speech-to-Text) based on the filtered audio data;
Including,
The filtered audio data includes voice data,
The VAD model is,
Obtaining audio data for learning including only voice data;
Generating a prediction result for the learning audio data using the VAD model; and
Learning the VAD model using a loss function that reduces a loss value when the audio data for training is predicted to be speech data and increases a loss value when the speech data is predicted to be non-speech data;
It is a learned artificial neural network model including,
The loss function is,
When a plurality of feature vectors of a plurality of voice data form a cluster on the feature space, the loss value decreases, and when the plurality of feature vectors do not form a cluster on the feature space, the loss value increases. designed to,
method.
delete According to claim 1,
The VAD model is a model learned using supervised learning or unsupervised learning based on audio data for learning,
The learning audio data includes only voice data,
method.
delete delete delete delete According to claim 1,
The step of generating a prediction result for the learning audio data using the VAD model includes:
Using the VAD model, a prediction result is generated as to whether the training audio data is classified as speech data or non-speech data, based on projecting the feature vector of the training audio data onto the feature space. steps;
Including,
method.
delete According to claim 8,
The loss function for the VAD model is:
The loss value is reduced when the feature vector of the additional speech data for training is projected onto a predetermined area in the feature space, and the loss value is increased when the feature vector of the additional speech data for training is projected outside the predetermined area. designed,
method.
delete A computer program stored in a computer-readable storage medium, wherein the computer program, when executed by one or more processors, causes the one or more processors to perform the following operations for detecting voice activity in audio data, said operations: heard:
generating a plurality of audio data based on dividing the audio signal into a plurality of tokens;
Inputting the plurality of audio data into a voice activity detection (VAD) model using a feature space;
Extracting a feature vector from each of the plurality of audio data using the VAD model and projecting the extracted feature vector into a feature space;
Among the plurality of audio data, filtering audio data in which the feature vector is projected onto a predetermined area in the feature space, where the predetermined area corresponds to a predetermined area based only on voice data; and
An operation of performing speech recognition (STT; Speech-to-Text) based on the filtered audio data;
Including,
The filtered audio data includes voice data,
The VAD model is,
Obtaining audio data for learning that includes only voice data;
An operation of generating a prediction result for the learning audio data using the VAD model; and
An operation of training the VAD model using a loss function that reduces a loss value when the audio data for training is predicted to be speech data and increases a loss value when the speech data is predicted to be non-speech data;
It is a learned artificial neural network model including,
The loss function is,
Designed to reduce the loss value when a plurality of feature vectors of a plurality of voice data form a cluster on the feature space and increase the loss value when the plurality of feature vectors do not form a cluster on the feature space. ,
A computer program stored on a computer-readable storage medium.
A computing device for detecting voice activity in audio data, comprising:
at least one processor; and
Memory
Including,
The at least one processor,
Generating a plurality of audio data based on dividing the audio signal into a plurality of tokens,
Input the plurality of audio data into a voice activity detection (VAD) model using a feature space based on anomaly detection,
Using the VAD model, extract a feature vector of each of the plurality of audio data, project the extracted feature vector onto a feature space,
Among the plurality of audio data, filter audio data in which the feature vector is projected onto a predetermined region in the feature space, where the predetermined region corresponds to a predetermined region based only on voice data, and
Configured to perform speech recognition (STT; Speech-to-Text) based on the filtered audio data,
The filtered audio data includes voice data,
The VAD model is,
Obtaining audio data for training that includes only speech data;
Generating a prediction result for the learning audio data using the VAD model; and
training the VAD model using a loss function that reduces a loss value when the audio data for training is predicted to be speech data and increases a loss value when the speech data is predicted to be non-speech data;
It is a learned artificial neural network model including,
The loss function is,
When a plurality of feature vectors of a plurality of voice data form a cluster on the feature space, the loss value decreases, and when the plurality of feature vectors do not form a cluster on the feature space, the loss value increases. designed to,
Device.

Images