219results about How to "Improve speech recognition rate" patented technology

A mobile terminal including an input unit configured to receive an input to activate a voice recognition function on the mobile terminal and a memory configured to store multiple domains related to menus and operations of the mobile terminal. It further includes a controller configured to access a specific domain among the multiple domains included in the memory based on the received input to activate the voice recognition function, to recognize user speech based on a language model and an acoustic model of the accessed domain, and to determine at least one menu and operation of the mobile terminal based on the accessed specific domain and the recognized user speech.

A method for detecting a character or a word emphasized by a user from a voice inputted in a mobile terminal to refer it as meaningful information for a voice recognition, or emphatically displaying the user-emphasized character or word in a pre-set format when the inputted voice is converted into text, and a mobile terminal implementing the same are disclosed. The mobile terminal includes: a microphone to receive a voice of user; a controller to convert the received voice into corresponding text and detect a character or a word emphatically pronounced by the user from the voice; and a display unit to emphatically display the detected character or word in a pre-set format when the converted text is displayed.

A system and method are provided for recognizing a user's speech input. The method includes the steps for detecting the user's speech input, recognizing the user's speech input by comparing the speech input to a list of entries using language model statistics to determine the most likely entry matching the user's speech input, and detecting navigation information of a trip to a predetermined destination, where the most likely entry is determined by modifying the language model statistics taking into account the navigation information. A system and method is further provided that takes into account navigation trip information to determine the most likely entry using language model statistics for recognizing text input.

Disclose is a mobile terminal and control method thereof for inputting a voice to automatically generate a message to be sent during conversation using a mobile messenger, and it may include a microphone for inputting a user's voice, a display unit for displaying a mobile messenger; and a controller for inputting and recognizing a user's voice when a mobile messenger is implemented and then converting into a message to display the message on a message input window of the mobile messenger, and sending the displayed message to the other party which has been preset, and displaying the message sent to the other party and a message received from the other party in the sending and receiving order on a send / receive display window of the mobile messenger.

Embodiments of the present invention provide a voice identification method, which includes: obtaining voice data; obtaining a confidence value according to the voice data; obtaining a noise scenario according to the voice data; obtaining a confidence threshold corresponding to the noise scenario; and if the confidence value is greater than or equal to the confidence threshold, processing the voice data. An apparatus is also provided. The method and apparatus that flexibly adjust the confidence threshold according to the noise scenario greatly improve a voice identification rate under a noise environment.

The invention disclose an updating method, a device, and a system of a voice recognition device, which relate to voice recognition technology, and are invented for improving voice recognition rate. The method comprises the following steps: receiving voice input signals; utilizing a local voice recognition device to conduct voice recognition for the voice input signals, so as to obtain a local voice recognition result; acquiring an optimal recognition result as a final voice recognition result from the local voice recognition result and a cloud voice recognition result, wherein the cloud voice recognition result is obtained in a way that when the local voice device conducts voice recognition for the voice input signals, and meanwhile a cloud voice recognition device is utilized to conduct voice recognition for the voice input signals; combining acquired user feedback information, and the final voice recognition result to define whether the reliability of the local voice recognition result satisfies requirements; utilizing the cloud voice recognition device to conduct updating for the local voice recognition device when the reliability of the local voice recognition result is sure not to satisfy the requirements.

The present disclosure relates to an electronic device and a method for operating same, which can divide a recognition section for voice recognition, process different types of voice recognition corresponding to the divided recognition sections, and improve a voice recognition rate. According to an embodiment, a method for recognizing a voice by an electronic device may include: detecting a voice input; determining the direction of the voice and a beamforming direction. Voice recognition is based on the voice when the direction of the voice and the beamforming direction correspond to each other. Different types of voice recognition may be performed for voices input in a first section and a second section, based on divided types of voice input sections, when the direction of the voice and the beamforming direction do not correspond to each other.

The embodiment of the invention discloses a method and system for dividing sentences in audio and an automatic caption generation method and system for video files. The method for dividing sentences in audio includes the steps of identifying first pause, identifying a first sentence, identifying second pause, determining whether the audio is finished, and if not, repeating the above sentence / pause identification step until the audio is finished, wherein the pause has a minimal length restriction, the sentence has a minimal length restriction and a maximal length restriction. The speech recognition rate is thus increased, which makes full automatic caption production possible.

The invention discloses a voice interaction control method and device for intelligent equipment, and the method comprises the steps: monitoring and collecting a voice signal, sent by an intelligent equipment user, in real time; carrying out the voice recognition of the collected voice signal; wakening the corresponding functions of the intelligent equipment according to the voice recognition result of the voice signal, and determining whether to control the intelligent equipment to move or not: obtaining the position of the user if the intelligent equipment is controlled to move, controlling the intelligent equipment to move towards the user, shortening the distance between the intelligent equipment and the user, and recognizing another voice signal sent by the user; or else, controlling the intelligent equipment to carry out the corresponding operation according to the voice recognition result. According to the technical scheme of embodiment of the invention, the method controls the intelligent equipment to move towards the user according to the voice recognition result when a voice instruction of the user cannot be heard clearly because of long distance, thereby achieving the near-field voice interaction, ironing out the defect that a far-field voice recognition effect is poor, and providing natural and smooth interaction experience for the user.

A speech recognition method comprises model selection step which selects a recognition model based on characteristic information of input speech and speech recognition step which translates input speech into text data based on the selected recognition model.

Voice operation device includes: voice recognition dictionary for storing plurality of groups of synonyms provided for plurality of functions of devices to be operated and each includes at least one word; voice recognition unit that checks voice data from voice taking unit against words stored in voice recognition dictionary to recognize word corresponding to voice; device control unit that controls devices to be operated based on word recognized by voice recognition unit; recognition history storage unit that sequentially stores words recognized by voice recognition unit; and dictionary update unit that updates voice recognition dictionary in such way that words which are determined to have been recognized at low frequencies in the past, based on recognition history stored in recognition history storage unit, are deleted except at least one of word which is left in each group of plurality of groups of synonyms in order to be checked.

A method for speech recognition can include generating a context-enhanced database from a system input. A voice-generated output can be generated from a speech signal by performing a speech recognition task to convert the speech signal into computer-processable segments. During the speech recognition task, the context-enhanced database can be accessed to improve the speech recognition rate. Accordingly, the speech signal can be interpreted with respect to words included within the context-enhanced database. Additionally, a user can edit or correct an output in order to generate the final voice-generated output which can be made available.

The operation of a device by a user is detected, and one or more speech input executing commands corresponding to the operated device are provided to the user, e.g., by speech or by being displayed on a screen. Speech input guidance may be stopped if it would interfere with an audio or image output of an operated device, or if a count of guidance speech outputs exceeds a predetermined number.

The embodiment of the invention provides a voice identification method which comprises the following steps of: acquiring voice data; acquiring a first confidence level value according to the voice data; acquiring a noise scene according to the voice data; acquiring a second confidence level value corresponding to the noise scene according to the first confidence level value; and if the second confidence level value is greater than or equal to a pre-stored confidence level threshold value, processing the voice data. The invention also discloses a device. According to the noise scene, the method and the device can be used for flexibly adjusting the confidence level value, so that the voice identification rate in the noise environment is greatly improved.

The invention provides a noise reduction method, a noise reduction device and a terminal, wherein the noise reduction method comprises the steps of determining noise properties of received voice information; determining a target noise reduction manner from a plurality of pre-set noise reduction manners according to the noise properties of the voice information, and reducing noise of the voice information, wherein the plurality of pre-set noise reduction manners comprise a hardware noise reduction manner and a software noise reduction manner. By means of the technical scheme provided by the invention, the noise reduction manner suitable for a current scene is determined according to the noise properties of the voice information, and therefore, input of the voice information can be enhanced; the voice recognition rate is increased, and further, the best noise reduction and voice recognition effects are achieved.

VoIP phones according to the present invention include a microphone, which may be internal or external, and allow the user to communicate unobtrusively, check voice mail and conduct other activities in an environment which can be noisy in general and extremely noisy sometimes. Speech recognition functionally may also be used to generate and send touch tone or DTMF tones such as in response to call trees or voice recognition functionality used by airlines, credit card companies, voice mail systems, and other applications. A system and method of audio processing which provides enhanced speech recognition is provided. Audio input is received at the microphone which is processed by adaptive noise cancellation to generate an enhanced audio signal. The operation of the speech recognition engine and the adaptive noise canceller may be advantageously controlled based on Voice Activity Detection (VAD).

The invention discloses a voice recognition method, device and apparatus and a storage medium. The method includes: when a sounding event is triggered, receiving a voice signal sent by a microphone and acquired by a user during the execution of the sounding event and an image signal including a lip; performing feature extraction on the voice signal to generate a voice feature signal, and performing feature extraction on the image signal including the lip to generate a lip-language feature signal; sending the voice feature signal and the lip-language feature signal to a server to instruct the server to match the voice feature signal with a preset voice signal to generate a voice recognition result and to match the lip-language feature signal with a preset lip-language signal to generate a lip-language recognition result; if the similarity between the voice recognition result and the lip-language recognition result is greater than or equal to a similarity threshold, generating a recognition feedback result according to the voice recognition result and sending the recognition feedback result to a terminal. The embodiment of the invention achieves improved voice recognition rate.

The invention relates to a voice recognition method, a voice recognition device and a voice recognition system. The voice recognition method comprises the following steps: preprocessing picked voice data to obtain preprocessed voice data; extracting voice features in the preprocessed voice data; matching the voice features with a local voice feature database, and if the matching is unsuccessful, sending a first voice recognition request to a target server; receiving a first voice recognition response which includes the matching result of the voice features and the local voice feature database through the target server returned by the target server and sending a second voice recognition request to the target server when the first voice recognition response presents that the matching is unsuccessful, and receiving a second voice recognition response returned by the target server, wherein the second voice recognition response comprises a voice recognition result obtained after the preprocessed voice data is sent to a manual translation terminal to be manually translated. According to the technical scheme of the embodiment of the invention, the voice recognition rate of an intelligent terminal is greatly improved.

The invention discloses an acoustic model based on a depth bidirectional long short-term memory (DBLSTM) recurrent neural network (RNN). The DBLSTM network is mainly divided into three parts: in a full-connection part of the DBLSTM, Maxout nerve cells are used for replacing original Sigmoid nerve cells to solve the problems of gradient disappearance and explosion commonly appearing in RNN; simultaneously, Dropout regularization training algorithm is used for preventing the neural network from causing overfitting in the training process. In the multi-layer BLSTM part, in order to adapt to the bidirectional dependency of DBLSTM on each time step length, a Context-sensitive-chunk Back-propagation through time (CSC-BPTT) algorithm is provided for training the network. A selection link layer is adopted on the back of the multi-layer BLSTM part, and is used for carrying out conversion on the output of the DBLSTM to obtain the input of the full-connection part. According to the acoustic model, higher voice recognition rate can be obtained.

A processing unit is provided which executes speech recognition on speech signals captured by a microphone for capturing sounds uttered in an environment. The processing unit has: an initial reflection component extraction portion that extracts initial reflection components by removing diffuse reverberation components from a reverberation pattern of an impulse response generated in the environment; and an acoustic model learning portion that learns an acoustic model for the speech recognition by reflecting the initial reflection components to speech data for learning.

The embodiment of the invention discloses a voice recognition method and a mobile terminal, and the voice recognition method can comprise the steps: carrying out the voice recognition of a received voice message, and obtaining an unrecognized voice segment; judging whether a local voice library stores a voice sample matched with the unrecognized voice segment or not; and determining a recognition result of the unrecognized voice segment according to the meaning marked by the matched voice sample if the local voice library stores the voice sample matched with the unrecognized voice segment. According to the embodiment of the invention, the method can search the matched sample from the local voice library for recognizing the voice segment which cannot be recognized through a conventional method, thereby effectively improving the voice recognition rate.

The invention relates to the field of personal mobility technology and vehicle-mounted Bluetooth technology and discloses a Bluetooth voice control method and device and an intelligent terminal. The method comprises the steps of conducting Bluetooth matching and connection with a peripheral, sending a request signal to the peripheral, attempting connection with an RFCOMM channel set by the peripheral, receiving an RFCOMM channel establishment confirmation signal returned by the peripheral, establishing a Socket port based on an RFCOMM, receiving a VTCP data packet transmitted from the peripheral through the Socket port, recognizing a voice instruction in the data packet and converting the voice instruction into a specific control instruction for operation of the intelligent terminal. According to the Bluetooth voice control method, device and intelligent terminal, a VTCP is adopted to transmit 16-KHz voice signals through the RFCOMM channel, so that the voice recognition rate is effectively increased; furthermore, efficient and accurate voice control over the intelligent terminal is achieved, and safety and convenience in the process of controlling the intelligent terminal by a user are guaranteed to the maximum extent.

The invention provides a voice recognition method which is applied to electronic equipment. The method comprises the steps that voice information inputted by a user is acquired; the voice information is recognized by using a first voice recognition method so that a first voice recognition result is obtained, and the voice information is recognized by using a second voice recognition method so that a second voice recognition result is obtained, wherein the first voice recognition method and the second voice recognition method operate in parallel; and the first voice recognition result and the second voice recognition result are displayed according to the preset rules. The invention also provides a voice recognition system. With application of the voice recognition method and system, the second voice recognition method can be utilized to assist the first voice recognition method to recognize the voice information of the user so that the voice recognition rate can be enhanced.

The invention discloses a speech recognition method and apparatus. The speech recognition method includes the steps of obtaining the position information of a user through a client, loading corresponding dialect speech base corresponding to a regional dialect according to the position information, obtaining speech information input by a user, calling the dialect speech base, and parsing and identifying the speech information by using a speech parsing algorithm corresponding to the dialect speech base. According to the technical scheme disclosed in the invention, the speech recognition rate and speech parsing accuracy of different regional dialects can be improved, the use experience can be improved, and user group can be expanded.

The invention discloses a human-computer interaction system and method controlling an IP set-top box through smart phone voice. The system comprises a smart terminal and the controlled IP set-top box, wherein the smart terminal and the controlled IP set-top box are connected through a communication network, the smart terminal comprises a voice acquiring module, a voice identifying module, a voice instruction screening module and a voice controlling module, and the voice identifying module is connected with a voice cloud computing platform through the Internet. According to the human-computer interaction system and method controlling the IP set-top box through the smart phone voice, voice control over the set-top box is achieved through a smart phone, operation is simple, and cost is low; voice identifying is performed through the cloud computing platform which collects a large amount of data of voice, vocabularies and the like, and the voice identifying rate is high; voice instructions are screened, the effects caused by accents, dialects, noise and the like are effectively weakened, the instruction expected by a user is hit, and accuracy of the voice control is improved; semantic identifying can be achieved through a custom voice input function, the voice expected by the user can be mapped to corresponding operation, and the accuracy of the voice control is further improved.

The embodiment of the invention provides an information processing method and electronic equipment and relates to the field of electronic terminals. The information processing method and the electronic equipment can improve voice recognition rate and improve user experience. The method comprises the steps that after a trigger instruction is obtained, the trigger instruction is responded to control a voice recognition engine unit to be switch into a normal work state from a low power consumption state; when the voice recognition engine unit is in the normal work state, a voice collecting unit can collect voice data, and a display unit can display first prompt information which is used for displaying that the electronic equipment is collecting voice; then parameter information of the voice data is obtained; output contents are decided according to the parameter information; the output contents are used for prompting a user to carry out input advise of voice interaction on the electronic equipment; the output contents are controlled to be output. The information processing method and the electronic equipment are used for voice input in a voice collecting state.

The speech recognition system of the present invention includes: a sound source separating section which separates mixed speeches from multiple sound sources; a mask generating section which generates a soft mask which can take continuous values between 0 and 1 for each separated speech according to reliability of separation in separating operation of the sound source separating section; and a speech recognizing section which recognizes speeches separated by the sound source separating section using soft masks generated by the mask generating section.

The invention relates to a vehicle-mounted speech control method and a vehicle-mounted speech control device. The method includes the following steps that: a corresponding relationship between first speech and a control instruction is established, wherein the first speech contains customized speech information stored in a first speech sample library, wherein the first speech sample library is stored in a vehicle-mounted storage unit of a vehicle; second speech is obtained, wherein the second speech is speech information inputted by a user; and if the first speech machined with the second speech is contained in the first speech sample library, the control instruction corresponding to the first speech is transmitted to a vehicle-mounted electronic control device, so as to realize control operation for the vehicle. With the method adopted, the user can customize speech training samples, the customization of the speech samples can be realized, the speech recognition rate of a vehicle-mounted speech control system can be improved; and since the user can customize the speech samples, situations that speech information inputted by the user cannot be matched with any speech information in the speech sample library can be decreased, interaction with a background cloud computing server can be reduced, the response time of the speech control system can be reduced, and the performance of the system can be improved.