Speech existence probability generation method, device and robot

Abstract

The present application discloses a method for generating a probability of speech existence, comprising: obtaining audio information collected by a robot; performing speech enhancement on the audio information to generate enhanced audio information; and generating a probability of speech existence including speech in the audio information based on the audio information and the enhanced audio information. According to an embodiment of the present disclosure, speech enhancement can be performed on the audio information collected by the robot to generate enhanced audio information, and a probability of speech existence including speech in the audio information can be generated based on the audio information and the enhanced audio information, thereby improving the accuracy of speech extraction. Through the method of the embodiment of the present disclosure, the user's speech can be effectively and accurately extracted from the audio information collected by the robot, thereby providing a basis for subsequent speech recognition, etc.

Description

Technical Field

[0001] The present disclosure relates to the field of robot technology, and in particular to a method and device for generating speech existence probability, a robot, and a storage medium. Background Art

[0002] With the continuous development of robots, robot pets are becoming more and more popular. However, robot pets, such as legged robots, are constantly moving during voice interaction with people. Unlike traditional fixed smart devices (such as smart speakers), robot pets generate a lot of noise due to their constant movement, such as the noise of the drive motor, the mechanical transmission noise of the joints during movement, etc. These noises will greatly interfere with voice recognition.

[0003] Therefore, it can be seen that the noise generated by the robot pet itself will seriously affect the normal human-computer interaction process and speech recognition. However, since the noise of the robot pet is usually sudden and only generates noise when the robot pet moves, the traditional speech detection method often cannot cope with such sudden interference. Therefore, there is an urgent need for a method for generating the probability of detecting the presence of robot speech, so that the user's speech can be accurately extracted from the ambient audio with many interfering noises. Summary of the invention

[0004] The present disclosure provides a method, device, robot, and storage medium for generating a speech existence probability, which are used to solve the problem that the user's speech cannot be accurately extracted.

[0005] According to an embodiment of one aspect of the present disclosure, a method for generating a probability of speech existence is provided, comprising: acquiring audio information collected by a robot; performing speech enhancement on the audio information to generate enhanced audio information; and generating a probability of speech existence including speech in the audio information based on the audio information and the enhanced audio information.

[0006] In one embodiment of the present disclosure, the enhancing the audio information to generate enhanced audio information includes: inputting the audio information into a neural network model to generate the enhanced audio information, wherein the neural network model is trained based on the noise generated by the robot itself.

[0007] In one embodiment of the present disclosure, generating the probability of speech existence including speech in the audio information based on the audio information and the enhanced audio information includes: generating a first eigenvalue based on the audio information, and generating a second eigenvalue based on the enhanced audio information; generating the speech existence probability based on the first eigenvalue and the second eigenvalue.

[0008] In one embodiment of the present disclosure, generating a first eigenvalue based on the audio information and generating a second eigenvalue based on the enhanced audio information includes: performing a root mean square operation on the audio information to generate the first eigenvalue; and performing a root mean square operation on the enhanced audio information to generate the second eigenvalue.

[0009] In one embodiment of the present disclosure, generating the speech existence probability based on the first eigenvalue and the second eigenvalue includes: determining whether the first eigenvalue and the second eigenvalue are less than a first preset threshold; if both the first eigenvalue and the second eigenvalue are less than the first preset threshold, determining that the audio information is noise.

[0010] In one embodiment of the present disclosure, it also includes: if the first eigenvalue or the second eigenvalue is greater than or equal to the first preset threshold, the first eigenvalue and the second eigenvalue are converted into a first non-negative eigenvalue and a second non-negative eigenvalue; and the speech existence probability is generated according to the first non-negative eigenvalue and the second non-negative eigenvalue.

[0011] In one embodiment of the present disclosure, generating the probability of speech existence based on the first non-negative eigenvalue and the second non-negative eigenvalue includes: obtaining the difference generated by subtracting the second non-negative eigenvalue from the first non-negative eigenvalue; if the difference is less than a second preset threshold, determining that the audio information includes speech; if the difference is greater than or equal to the second preset threshold, generating the probability of speech existence based on the difference and the second non-negative eigenvalue.

[0012] In one embodiment of the present disclosure, generating the probability of speech existence based on the difference and the second non-negative eigenvalue includes: dividing the difference by the second non-negative eigenvalue to generate the probability that the audio information is noise; generating the probability of speech existence based on the probability that the audio information is noise.

[0013] According to another embodiment of the present disclosure, a device for generating the probability of speech existence is also proposed, including: an audio acquisition module for acquiring audio information collected by a robot; an enhancement module for performing speech enhancement on the audio information to generate enhanced audio information; and a probability generation module for generating the probability of speech existence including speech in the audio information based on the audio information and the enhanced audio information.

[0014] In one embodiment of the present disclosure, the enhancement module includes a neural network model, wherein the neural network model converts the audio information into the enhanced audio information, wherein the neural network model is trained based on the noise generated by the robot itself.

[0015] In one embodiment of the present disclosure, the probability generation module includes: a feature value generation submodule, used to generate a first feature value based on the audio information, and to generate a second feature value based on the enhanced audio information; and a probability generation submodule, used to generate the probability of speech existence based on the first feature value and the second feature value.

[0016] In one embodiment of the present disclosure, the eigenvalue generation submodule performs a root mean square operation on the audio information to generate the first eigenvalue, and performs a root mean square operation on the enhanced audio information to generate the second eigenvalue.

[0017] In an embodiment of the present disclosure, the probability generation submodule determines that the audio information is noise when both the first eigenvalue and the second eigenvalue are smaller than the first preset threshold.

[0018] In one embodiment of the present disclosure, the probability generation submodule converts the first eigenvalue and the second eigenvalue into a first non-negative eigenvalue and a second non-negative eigenvalue when the first eigenvalue or the second eigenvalue is greater than or equal to the first preset threshold, and generates the probability of speech existence based on the first non-negative eigenvalue and the second non-negative eigenvalue.

[0019] In one embodiment of the present disclosure, the probability generation submodule obtains the difference generated by subtracting the second non-negative eigenvalue from the first non-negative eigenvalue, and when the difference is less than a second preset threshold, determines that the audio information includes speech, and when the difference is greater than or equal to the second preset threshold, generates the probability of the existence of the speech based on the difference and the second non-negative eigenvalue.

[0020] In one embodiment of the present disclosure, the probability generation submodule divides the difference by the second non-negative eigenvalue to generate the probability that the audio information is noise, and generates the speech existence probability according to the probability that the audio information is noise.

[0021] According to yet another embodiment of the present disclosure, a robot is provided, comprising the device as described above.

[0022] According to another embodiment of the present disclosure, an electronic device is provided, comprising:

[0023] at least one processor; and

[0024] a memory communicatively connected to the at least one processor; wherein,

[0025] The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the method described above.

[0026] According to yet another embodiment of the present disclosure, a non-transitory computer-readable storage medium storing computer instructions is provided, wherein the computer instructions are used to enable the computer to execute the method described above.

[0027] According to the embodiments of the present disclosure, the audio information collected by the robot can be voice enhanced to generate enhanced audio information, and the probability of the existence of the voice including the voice in the audio information can be generated based on the audio information and the enhanced audio information, thereby improving the accuracy of voice extraction. Through the method of the embodiments of the present disclosure, the user's voice can be effectively and accurately extracted from the audio information collected by the robot, thereby providing a basis for subsequent voice recognition, etc. In addition, in the embodiments of the present disclosure, since the voice in the robot's audio information can be accurately identified, the user's voice command can be timely and accurately identified, and a timely response can be made. Thereby improving the response speed and response accuracy of the robot, especially the legged robot.

[0028] It should be understood that the content described in this section is not intended to identify the key or important features of the embodiments of the present disclosure, nor is it intended to limit the scope of the present disclosure. Other features of the present disclosure will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] The accompanying drawings are used to better understand the present solution and do not constitute a limitation of the present application.

[0030] Figure 1 This is a flow chart of a method for generating speech existence probability according to an embodiment of the present disclosure;

[0031] Figure 2 A flowchart of a method for generating speech existence probability according to an embodiment of the present disclosure;

[0032] Figure 3 This is a structural diagram of a device for generating a speech existence probability according to an embodiment of the present disclosure;

[0033] Figure 4 The invention is a block diagram of an electronic device for a method for generating a speech existence probability according to an embodiment of the present invention. DETAILED DESCRIPTION

[0034] The following is a description of exemplary embodiments of the present application in conjunction with the accompanying drawings, including various details of the embodiments of the present application to facilitate understanding, which should be considered as merely exemplary. Therefore, it should be recognized by those of ordinary skill in the art that various changes and modifications can be made to the embodiments described herein without departing from the scope and spirit of the present application. Similarly, for the sake of clarity and conciseness, the description of well-known functions and structures is omitted in the following description.

[0035] like Figure 1 As shown, it is a flow chart of a method for generating the probability of speech existence according to an embodiment of the present disclosure. In one embodiment of the present disclosure, the method for generating the probability of speech existence can be run in a robot, such as a robot pet, or a legged robot. Of course, in other embodiments of the present disclosure, the method can also be run in a user's mobile terminal, or in a server. In one embodiment of the present disclosure, the legged robot includes multiple legs, such as a quadruped robot or a bipedal robot, or other legged robots. The method comprises the following steps:

[0036] Step 110, obtaining audio information collected by the robot.

[0037] In one embodiment of the present disclosure, a microphone may be provided on the body of a robot, such as a legged robot, and audio is collected through the microphone to generate audio information. In other embodiments of the present disclosure, multiple microphones may be provided on the body of the robot, such as six microphones, two microphones respectively provided at the front and rear of the body, and four microphones respectively provided at both sides of the body. In the embodiments of the present disclosure, the robot, especially the legged robot, is in constant motion. Therefore, the legged robot may be far away from the user and its position is not fixed. In order to better receive the user's voice command, multiple microphones in different directions are provided. According to the signal-to-noise ratio of each microphone, the microphone with the lowest signal-to-noise ratio may be selected as the voice input. The microphone with a low signal-to-noise ratio may be a microphone that is closer to the user, or it may be a microphone facing the user, so the user's voice received by it is clearer.

[0038] In one embodiment of the present disclosure, the audio collected by the robot may contain the user's voice and environmental noise. At the same time, since the robot is constantly moving, there will also be noise generated by the robot's own movement. For example, the noise of the robot's own driving motor, the noise of the robot's joint movement, and the noise generated by the robot's feet colliding with the ground, etc. If the robot is in a quiet state, it can be detected by detecting features such as spectrum flatness, such as detecting the state change from a quiet state to the appearance of voice, so as to achieve judgment. However, for the robot, the noise generated by itself is not stable. Since its motion state is not fixed, the noise generated by itself is sudden and belongs to sudden interference (i.e., transient interference). This interference is very difficult to determine whether the collected audio information includes voice in the voice detection scenario.

[0039] Step 130: Perform speech enhancement on the audio information to generate enhanced audio information.

[0040] In one embodiment of the present disclosure, audio information can be input into a neural network model to generate enhanced audio information, wherein the neural network model is trained based on the noise generated by the robot itself. In the embodiments of the present disclosure, the structure of the neural network model is not limited, and neural network model structures such as DNN (Deep Neural Networks), RNN (Recurrent Neural Network), and CNN (Convolutional Neural Networks) can be implemented.

[0041] In the embodiments of the present disclosure, although the noise generated by the robot itself is sudden, the spectrum of the noise is fixed. For example, the noise of the motor drive in the robot, the noise of the joint rotation, and the collision noise generated by the robot's feet hitting the ground are basically fixed, so these noises can be sampled and entered into the neural network model as samples for learning. In other embodiments of the present disclosure, although different ground surfaces will produce different noises, the materials of the ground are limited, so different ground materials can be sampled to improve the neural network model.

[0042] In the embodiments of the present disclosure, performing speech enhancement on audio information to generate enhanced audio information means filtering the frequency spectrum of the noise generated by the robot itself learned by the neural network model from the audio information. Since the noise is filtered out, the speech is enhanced, so it is referred to as enhanced audio information herein.

[0043] Step 150: Generate a probability of existence of speech including speech in the audio information according to the audio information and the enhanced audio information.

[0044] In one embodiment of the present disclosure, since the noise (noise generated by the robot itself) of the audio information is removed through a neural network model, enhanced audio information can be obtained. The enhanced audio information is compared with the audio information to obtain the probability of the existence of speech in the audio information. For example, if there is only noise in the audio information, then the enhanced audio information will be very small relative to the audio information (because most of the noise has been filtered out), so it can be determined that the probability of the existence of speech including speech in the audio information is very low, or even zero, that is, it is all noise. Conversely, if the enhanced audio information does not change much relative to the audio information, it means that very little noise has been removed and a lot of speech is left, so the probability of the existence of speech including speech in the audio information will be very high.

[0045] According to the embodiments of the present disclosure, the audio information collected by the robot can be enhanced to generate enhanced audio information, and the probability of the existence of the voice including the voice in the audio information is generated according to the audio information and the enhanced audio information, thereby improving the accuracy of voice extraction. Through the method of the embodiments of the present disclosure, the user's voice can be effectively and accurately extracted from the audio information collected by the robot, thereby providing a basis for subsequent voice recognition, etc.

[0046] like Figure 2 FIG. 1 is a flowchart of a method for generating speech existence probability according to an embodiment of the present disclosure. This embodiment is only a method for realizing speech existence probability recognition. Those skilled in the art can refer to this method to propose other speech existence probability recognition methods, which should be covered by the protection scope of the present invention.

[0047] Step 210: Generate a first eigenvalue based on the audio information, and generate a second eigenvalue based on the enhanced audio information.

[0048] In one embodiment of the present disclosure, a root mean square operation is performed on the audio information to generate a first eigenvalue, and a root mean square operation is performed on the enhanced audio information to generate a second eigenvalue. In other embodiments of the present disclosure, the root mean square is only one implementation method, and the measurement of the audio information and the enhanced audio information can also be achieved through other operations.

[0049] Step 230: Generate a speech existence probability according to the first eigenvalue and the second eigenvalue.

[0050] In one embodiment of the present disclosure, it can be determined whether the first eigenvalue and the second eigenvalue are less than a first preset threshold. If both the first eigenvalue and the second eigenvalue are less than the first preset threshold, the audio information is determined to be noise. As described above, if both the first eigenvalue and the second eigenvalue are less than the first preset threshold, it means that both the audio information and the enhanced audio information are very small, so the audio information is noise, that is, the probability of speech existence is 0.

[0051] In another embodiment of the present disclosure, if the first eigenvalue or the second eigenvalue is greater than or equal to the first preset threshold, it means that there may be speech in the audio information, so further judgment is required. Specifically, the first eigenvalue and the second eigenvalue are first converted into the first non-negative eigenvalue and the second non-negative eigenvalue, and then the probability of speech existence is generated according to the first non-negative eigenvalue and the second non-negative eigenvalue. In one embodiment of the present disclosure, since the first eigenvalue and the second eigenvalue are negative values, they need to be converted to the non-negative integer domain. In one embodiment of the present disclosure, the first eigenvalue can be added with a preset value to generate a first non-negative eigenvalue, and the second eigenvalue can be added with a preset value to generate a second non-negative eigenvalue. In one embodiment of the present disclosure, the preset value can be 100. Of course, in other embodiments of the present disclosure, a larger or smaller preset value can be selected as needed.

[0052] In one embodiment of the present disclosure, the first non-negative eigenvalue may be subtracted from the second non-negative eigenvalue to generate a difference, and then a judgment is made based on the difference. If the difference is less than a second preset threshold value (for example, 0), it is judged that the audio information includes speech, that is, the probability of speech existence is 1. In this embodiment, if the second non-negative eigenvalue is larger than the first non-negative eigenvalue, the audio information is mainly speech. On the contrary, if the difference is greater than or equal to the second preset threshold value, the speech existence probability can be generated based on the difference and the second non-negative eigenvalue. Specifically, the difference is divided by the second non-negative eigenvalue to generate the probability that the audio information is noise, and then the speech existence probability is generated based on the probability that the audio information is noise, that is, 1 minus the probability that the audio information is noise, to obtain the speech existence probability.

[0053] In one embodiment of the present disclosure, after the probability of the audio information being noise is generated, boundary detection is also required to limit the probability of the audio information being noise to a range of 0 to 1. For example, if the probability of the audio information being noise is less than 0, it is set to 0, and conversely, if the probability of the audio information being noise is greater than 1, it is set to 1, thereby filtering out extreme cases and improving accuracy.

[0054] According to the embodiments of the present disclosure, the audio information collected by the robot can be enhanced to generate enhanced audio information, and the probability of the existence of the voice including the voice in the audio information is generated according to the audio information and the enhanced audio information, thereby improving the accuracy of voice extraction. Through the method of the embodiments of the present disclosure, the user's voice can be effectively and accurately extracted from the audio information collected by the robot, thereby providing a basis for subsequent voice recognition, etc.

[0055] In the embodiments of the present disclosure, since the voice in the robot audio information can be accurately recognized, the user's voice command can be timely and accurately recognized and responded to in a timely manner, thereby improving the response speed and response accuracy of the robot, especially the legged robot.

[0056] like Figure 3 , which is a structural diagram of a speech existence probability generating device according to an embodiment of the present disclosure. The speech existence probability generating device 300 includes an audio acquisition module 310, an enhancement module 320 and a probability generating module 330. The audio acquisition module 310 is used to acquire audio information collected by the robot. The enhancement module 320 is used to perform speech enhancement on the audio information to generate enhanced audio information. The probability generating module 330 is used to generate the speech existence probability of the speech included in the audio information according to the audio information and the enhanced audio information.

[0057] In one embodiment of the present disclosure, in the audio collected by the robot, there may be user voices and environmental noises. At the same time, since the robot is constantly moving, there will also be noises generated by the movement of the robot itself. For example, the noise of the robot's own driving motor, the noise of the robot's joint movement, and the noise generated by the robot's feet colliding with the ground, etc. If the robot is in a quiet state, it can be detected by detecting features such as spectrum flatness, such as detecting the state change from a quiet state to the appearance of voice, so as to achieve judgment. As for environmental noise, since the environmental noise is stable, speech recognition can be achieved through technologies such as IMCRA (Improved Minima Controlled Recursive Averaging, improved control-based recursive averaging). However, for the robot, the noise generated by itself is not stable. Since its motion state is not fixed, the noise generated by itself is sudden, which belongs to sudden interference (i.e., transient interference). This interference is very difficult to determine whether the collected audio information includes voice for the voice detection scene.

[0058] In one embodiment of the present disclosure, the enhancement module 320 includes a neural network model, wherein the neural network model converts audio information into the enhanced audio information, wherein the neural network model is trained based on the noise generated by the robot itself. In one embodiment of the present disclosure, the audio information can be input into the neural network model to generate enhanced audio information, wherein the neural network model is trained based on the noise generated by the robot itself. In the embodiments of the present disclosure, the structure of the neural network model is not limited, and neural network structures such as DNN (Deep Neural Networks), RNN (Recurrent Neural Network), and CNN (Convolutional Neural Networks) can be implemented.

[0059] In the embodiments of the present disclosure, although the noise generated by the robot itself is sudden, the spectrum of the noise is fixed. For example, the noise of the motor drive in the robot, the noise of the joint rotation, and the collision noise generated by the robot's feet hitting the ground are basically fixed, so these noises can be sampled and entered into the neural network model as samples for learning. In other embodiments of the present disclosure, although different ground surfaces will produce different noises, the materials of the ground are limited, so different ground materials can be sampled to improve the neural network model.

[0060] In the embodiments of the present disclosure, performing speech enhancement on audio information to generate enhanced audio information means filtering the frequency spectrum of the noise generated by the robot itself learned by the neural network model from the audio information. Since the noise is filtered out, the speech is enhanced, so it is referred to as enhanced audio information herein.

[0061] In one embodiment of the present disclosure, since the noise (noise generated by the robot itself) of the audio information is removed through a neural network model, enhanced audio information can be obtained. The enhanced audio information is compared with the audio information to obtain the probability of the existence of speech in the audio information. For example, if there is only noise in the audio information, then the enhanced audio information will be very small relative to the audio information (because most of the noise has been filtered out), so it can be determined that the probability of the existence of speech including speech in the audio information is very low, or even zero, that is, it is all noise. Conversely, if the enhanced audio information does not change much relative to the audio information, it means that very little noise has been removed and a lot of speech is left, so the probability of the existence of speech including speech in the audio information will be very high.

[0062] In one embodiment of the present disclosure, the probability generation module 330 includes an eigenvalue generation submodule 331 and a probability generation submodule 332. The eigenvalue generation submodule 331 is used to generate a first eigenvalue according to the audio information and to generate a second eigenvalue according to the enhanced audio information. The probability generation submodule 332 is used to generate a speech existence probability according to the first eigenvalue and the second eigenvalue.

[0063] In one embodiment of the present disclosure, the feature value generation submodule 332 performs a root mean square operation on the audio information to generate the first feature value, and performs a root mean square operation on the enhanced audio information to generate the second feature value.

[0064] In one embodiment of the present disclosure, the probability generation submodule 332 determines that the audio information is noise when both the first eigenvalue and the second eigenvalue are less than the first preset threshold. The probability generation submodule 332 converts the first eigenvalue and the second eigenvalue into the first non-negative eigenvalue and the second non-negative eigenvalue when the first eigenvalue or the second eigenvalue is greater than or equal to the first preset threshold, and generates the probability of speech existence according to the first non-negative eigenvalue and the second non-negative eigenvalue.

[0065] In one embodiment of the present disclosure, the probability generation submodule 332 obtains the difference generated by subtracting the second non-negative eigenvalue from the first non-negative eigenvalue, and when the difference is less than the second preset threshold, determines that the audio information includes speech, and when the difference is greater than or equal to the second preset threshold, generates the probability of the speech existence according to the difference and the second non-negative eigenvalue. In one embodiment of the present disclosure, the probability generation submodule 332 divides the difference by the second non-negative eigenvalue to generate the probability that the audio information is noise, and generates the probability of speech existence according to the probability that the audio information is noise.

[0066] In one embodiment of the present disclosure, a robot is also disclosed, comprising the device as described above.

[0067] In one embodiment of the present disclosure, an electronic device is also disclosed, including: at least one processor; and a memory in communication with the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the method described above. The electronic device can be a robot, such as a legged robot, or a user's mobile terminal, or a server, or other similar devices.

[0068] In one embodiment of the present disclosure, a non-transitory computer-readable storage medium storing computer instructions is also proposed, wherein the computer instructions are used to enable the computer to execute the method as described above.

[0069] According to the embodiments of the present disclosure, the audio information collected by the robot can be voice enhanced to generate enhanced audio information, and the probability of the existence of the voice including the voice in the audio information can be generated based on the audio information and the enhanced audio information, thereby improving the accuracy of voice extraction. Through the method of the embodiments of the present disclosure, the user's voice can be effectively and accurately extracted from the audio information collected by the robot, thereby providing a basis for subsequent voice recognition, etc. In addition, in the embodiments of the present disclosure, since the voice in the robot's audio information can be accurately identified, the user's voice command can be timely and accurately identified, and a timely response can be made. Thereby improving the response speed and response accuracy of the robot, especially the legged robot.

[0070] like Figure 4 , is a block diagram of an electronic device according to a method for generating a probability of speech existence according to an embodiment of the present application. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workbenches, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital processing, cellular phones, smart phones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples, and are not intended to limit the implementation of the present application described and / or required herein.

[0071] like Figure 4 As shown, the electronic device includes: one or more processors Y01, a memory Y02, and interfaces for connecting various components, including high-speed interfaces and low-speed interfaces. The various components are interconnected using different buses and can be installed on a common mainboard or installed in other ways as needed. The processor can process instructions executed in the electronic device, including instructions stored in or on the memory to display graphical information of the GUI on an external input / output device (such as a display device coupled to the interface). In other embodiments, if necessary, multiple processors and / or multiple buses can be used together with multiple memories and multiple memories. Similarly, multiple electronic devices can be connected, and each device provides part of the necessary operations (for example, as a server array, a group of blade servers, or a multi-processor system). Figure 4 A processor Y01 is taken as an example.

[0072] The memory Y02 is the non-transient computer-readable storage medium provided in the present application. The memory stores instructions executable by at least one processor to enable the at least one processor to execute the method for generating the probability of speech existence provided in the present application. The non-transient computer-readable storage medium of the present application stores computer instructions, which are used to enable a computer to execute the method for generating the probability of speech existence provided in the present application.

[0073] The memory Y02, as a non-transient computer-readable storage medium, can be used to store non-transient software programs, non-transient computer executable programs and modules, such as program instructions / modules corresponding to the method for generating the probability of speech existence in the embodiment of the present application. The processor Y01 executes various functional applications and data processing of the server by running the non-transient software programs, instructions and modules stored in the memory Y02, that is, implements the method for generating the probability of speech existence in the above method embodiment.

[0074] The memory Y02 may include a program storage area and a data storage area, wherein the program storage area may store an operating system, an application required by at least one function; the data storage area may store data created according to the use of the electronic device of the speech existence probability generation method, etc. In addition, the memory Y02 may include a high-speed random access memory, and may also include a non-transient memory, such as at least one disk storage device, a flash memory device, or other non-transient solid-state storage device. In some embodiments, the memory Y02 may optionally include a memory remotely arranged relative to the processor Y01, and these remote memories may be connected to the electronic device of the speech existence probability generation method via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0075] The electronic device of the speech existence probability generation method may further include: an input device Y03 and an output device Y04. The processor Y01, the memory Y02, the input device Y03 and the output device Y04 may be connected via a bus or other means, and FIG. Y takes the bus connection as an example.

[0076] The input device Y03 can receive input digital or character information, and generate key signal input related to user settings and function control of the electronic device of the speech existence probability generation method, such as a touch screen, a keypad, a mouse, a track pad, a touch pad, an indicator rod, one or more mouse buttons, a trackball, a joystick and other input devices. The output device Y04 may include a display device, an auxiliary lighting device (e.g., an LED) and a tactile feedback device (e.g., a vibration motor), etc. The display device may include, but is not limited to, a liquid crystal display (LCD), a light emitting diode (LED) display and a plasma display. In some embodiments, the display device may be a touch screen.

[0077] Various implementations of the systems and techniques described herein can be realized in digital electronic circuit systems, integrated circuit systems, dedicated ASICs (application specific integrated circuits), computer hardware, firmware, software, and / or combinations thereof. These various implementations can include: being implemented in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which can be a special purpose or general purpose programmable processor that can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.

[0078] These computer programs (also referred to as programs, software, software applications, or code) include machine instructions for programmable processors and can be implemented using high-level procedural and / or object-oriented programming languages, and / or assembly / machine languages. As used herein, the terms "machine-readable medium" and "computer-readable medium" refer to any computer program product, device, and / or means (e.g., disk, optical disk, memory, programmable logic device (PLD)) for providing machine instructions and / or data to a programmable processor, including a machine-readable medium that receives machine instructions as a machine-readable signal. The term "machine-readable signal" refers to any signal for providing machine instructions and / or data to a programmable processor.

[0079] To provide interaction with a user, the systems and techniques described herein can be implemented on a computer having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and pointing device (e.g., a mouse or trackball) through which the user can provide input to the computer. Other types of devices can also be used to provide interaction with the user; for example, the feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including acoustic input, voice input, or tactile input).

[0080] The systems and techniques described herein may be implemented in a computing system that includes back-end components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes front-end components (e.g., a user computer with a graphical user interface or a web browser through which a user can interact with implementations of the systems and techniques described herein), or a computing system that includes any combination of such back-end components, middleware components, or front-end components. The components of the system may be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: a local area network (LAN), a wide area network (WAN), and the Internet.

[0081] A computer system may include a client and a server. The client and the server are generally remote from each other and usually interact through a communication network. The relationship between the client and the server is generated by computer programs running on the corresponding computers and having a client-server relationship with each other. The server may be a cloud server, also known as a cloud computing server or cloud host, which is a host product in the cloud computing service system to solve the defects of difficult management and weak business scalability in traditional physical hosts and VPS services ("Virtual Private Server", or "VPS" for short).

[0082] It should be understood that the various forms of processes shown above can be used to reorder, add or delete steps. For example, the steps recorded in this application can be executed in parallel, sequentially or in different orders, as long as the expected results of the technical solution disclosed in this application can be achieved, and this document is not limited here.

[0083] The above specific implementations do not constitute a limitation on the protection scope of this application. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions and improvements made within the spirit and principles of this application should be included in the protection scope of this application.

Claims

1. A method for generating speech existence probability, characterized in that: include: Get the audio information collected by the robot; Performing speech enhancement on the audio information to generate enhanced audio information; as well as generating a probability of existence of speech including speech in the audio information according to the audio information and the enhanced audio information; The generating, according to the audio information and the enhanced audio information, a probability of existence of speech including speech in the audio information comprises: generating a first eigenvalue according to the audio information, and generating a second eigenvalue according to the enhanced audio information; Generating the speech existence probability according to the first eigenvalue and the second eigenvalue; The generating the speech existence probability according to the first eigenvalue and the second eigenvalue comprises: If the first eigenvalue or the second eigenvalue is greater than or equal to a first preset threshold, converting the first eigenvalue and the second eigenvalue into a first non-negative eigenvalue and a second non-negative eigenvalue; The speech existence probability is generated according to the first non-negative eigenvalue and the second non-negative eigenvalue.

2. The method according to claim 1, characterized in that The step of enhancing the audio information to generate enhanced audio information comprises: The audio information is input into a neural network model to generate the enhanced audio information, wherein the neural network model is trained based on the noise generated by the robot itself.

3. The method according to claim 1, characterized in that The generating a first eigenvalue according to the audio information and generating a second eigenvalue according to the enhanced audio information comprises: performing a root mean square operation on the audio information to generate the first eigenvalue; and A root mean square operation is performed on the enhanced audio information to generate the second eigenvalue.

4. The method according to claim 1, characterized in that The generating the speech existence probability according to the first eigenvalue and the second eigenvalue comprises: Determine whether the first characteristic value and the second characteristic value are less than a first preset threshold; If both the first eigenvalue and the second eigenvalue are smaller than the first preset threshold, the audio information is determined to be noise.

5. The method according to claim 4, characterized in that The generating the speech existence probability according to the first non-negative eigenvalue and the second non-negative eigenvalue comprises: Obtaining a difference value generated by subtracting the second non-negative eigenvalue from the first non-negative eigenvalue; If the difference is less than a second preset threshold, it is determined that the audio information includes speech; If the difference is greater than or equal to the second preset threshold, the speech existence probability is generated according to the difference and the second non-negative eigenvalue.

6. The method according to claim 5, characterized in that The step of generating the speech existence probability according to the difference and the second non-negative eigenvalue comprises: Dividing the difference by the second non-negative eigenvalue to generate a probability that the audio information is noise; The speech existence probability is generated according to the probability that the audio information is noise.

7. A device for generating a probability of speech existence, characterized in that: include: An audio acquisition module is used to obtain audio information collected by the robot; An enhancement module, configured to perform speech enhancement on the audio information to generate enhanced audio information; as well as A probability generation module, configured to generate a probability of existence of speech including speech in the audio information according to the audio information and the enhanced audio information; The probability generation module comprises: an eigenvalue generating submodule, configured to generate a first eigenvalue according to the audio information, and generate a second eigenvalue according to the enhanced audio information; A probability generation submodule, configured to generate the speech existence probability according to the first eigenvalue and the second eigenvalue; When the first eigenvalue or the second eigenvalue is greater than or equal to a first preset threshold, the probability generation submodule converts the first eigenvalue and the second eigenvalue into a first non-negative eigenvalue and a second non-negative eigenvalue, and generates the speech existence probability according to the first non-negative eigenvalue and the second non-negative eigenvalue.

8. The device according to claim 7, characterized in that The enhancement module includes a neural network model, wherein the neural network model converts the audio information into the enhanced audio information, wherein the neural network model is trained based on the noise generated by the robot itself.

9. The device according to claim 7, characterized in that The eigenvalue generation submodule performs a root mean square operation on the audio information to generate the first eigenvalue, and performs a root mean square operation on the enhanced audio information to generate the second eigenvalue.

10. The device according to claim 7, characterized in that The probability generation submodule determines that the audio information is noise when both the first eigenvalue and the second eigenvalue are smaller than the first preset threshold.

11. The device according to claim 10, characterized in that The probability generation submodule obtains the difference generated by subtracting the second non-negative eigenvalue from the first non-negative eigenvalue, and when the difference is less than a second preset threshold, determines that the audio information includes speech, and when the difference is greater than or equal to the second preset threshold, generates the probability of the existence of the speech based on the difference and the second non-negative eigenvalue.

12. The device according to claim 11, characterized in that The probability generation submodule divides the difference by the second non-negative eigenvalue to generate the probability that the audio information is noise, and generates the speech existence probability according to the probability that the audio information is noise.

13. A robot, characterized in that: Comprising a device as claimed in any one of claims 7 to 12.

14. An electronic device, characterized in that: include: at least one processor; as well as a memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the method according to any one of claims 1 to 6.

15. A non-transitory computer-readable storage medium storing computer instructions, characterized in that: The computer instructions are used to cause the computer to execute the method according to any one of claims 1 to 6.

Images