100 results about "Speech recognition performance" patented technology

An exception dictionary creating device, an exception dictionary creating method, and a program therefor allowing creating an exception dictionary are provided for affording high speech recognition performance while reducing the size of the exception dictionary, as well as a speech recognition device and a speech recognition method capable of recognizing a speech with high accuracy of recognition by using the exception dictionary. To achieve this, a text-to-phonetic symbol converting unit (21) of an exception dictionary creating device (10) creates converted phonetic symbol sequence by converting text sequence of vocabulary list data (21) into phonetic symbol sequence. A recognition degradation contribution degree calculating unit (24) calculates a recognition degradation contribution degree when the converted phonetic symbol sequence is not identical to a correct phonetic symbol sequence registered in a database or word dictionary (50). An exception dictionary registering unit (41) registers in the exception dictionary (60) the text sequence and the phonetic symbol sequence registered in the text sequence of the vocabulary list data (12) with a high degree of recognition degradation contribution degree to the recognition so as not to exceed data limitation capacity indicated by exception dictionary memory size content (71).

A speech processing apparatus which, in the process of performing echo canceling by using a pseudo acoustic echo signal, continuously uses an impulse response used for the previous frame as an impulse response to generate the pseudo acoustic echo signal when a voice is contained in the microphone input signal, and which uses a newly updated impulse response when a voice is not contained in the microphone input signal.

The invention discloses a speech recognition device and a speech recognition method. The speech recognition device comprises a unit which is configured to obtain speech inputted by a current user, a unit which is configured to split obtained speech and output at least two voice command segments, a unit which is configured to recognize a first predefined voice command from the voice command segments through using an acoustic model unrelated to a speaker, a unit which is configured to calculate a transformation matrix for the current user based on a voice command segment which is recognized as the first predefined voice command, a unit which is configured to select an acoustic model for the current user from acoustic models registered in the speech recognition device based the calculated transformation matrix, and a unit which is configured to recognize a second voice command from the voice command segments through using the selected acoustic model. According to the speech recognition device and the speech recognition method of the invention adopted, speech recognition performance can be improved through using the selected acoustic model (AM).

The invention discloses a speech recognition processing method and device. The method comprises: according to speed sample data of all areas of the country, training is carried out on a preset processing model to generate a universal mandarin acoustic model; and on the basis of the speech sample data of all provinces, adaptive training is carried out on the universal mandarin acoustic model respectively to generate mandarin acoustic models with dialectal accents, wherein the mandarin acoustic models corresponding to all provinces. Therefore, on the basis of the accent difference of users at different areas, mandarin acoustic models with dialectal accents are established, so that the speech recognition performance is improved.

A distributed speech recognition system and method thereof in accordance with the present invention enables a word and a natural language to be recognized using detection of a pause period in a speech period in an inputted speech signal, and various groups of recognition vocabulary (for example, a home speech recognition vocabulary, a telematics vocabulary for a vehicle, a vocabulary for call center, and so forth) to be processed in the same speech recognition system by determining the recognition vocabulary required by a corresponding terminal using an identifier of the terminal since various terminals require various speech recognition targets. In addition, various types of channel distortion occurring due to the type of terminal and the recognition environment are minimized by adapting them to a speech database model using a channel estimation method so that the speech recognition performance is enhanced.

Provided are an image display apparatus able to improve voice recognition performance by decreasing the volume of an audio signal output from the image display apparatus to a predetermined level or less when the image display apparatus recognizes user voice, and a method of controlling the same. The image display apparatus enabling voice recognition includes a first voice input unit to receive a user-side audio signal, an audio output unit to output an audio signal processed by the image display apparatus, a first voice recognizer to recognize the user-side audio signal received through the first voice input unit, and a controller to decrease the volume of the audio signal output through the audio output unit to a predetermined level if a voice recognition start command is received.

The invention provides a voice data processing method and apparatus of a mobile terminal. The method includes: acquiring voice data input by users; recognizing the voice data; obtaining a keyboard corresponding to the voice data and identification information of a controlled electronic device according to a recognition result, and performing classification according to goals of the users in the voice data; searching a first historical voice control event matched with the keyboard and a classification result from historical voice control events, wherein an execution result of the controlled electronic device corresponding to the first historical voice control event is execution success; and obtaining a corresponding historical control instruction from the first historical voice control event, and sending the historical control instruction to the controlled electronic device corresponding to the identification information to control the controlled electronic device to execute an operation corresponding to the historical control instruction. By employing the above scheme, the speech recognition performance can be enhanced.

The invention provides a far-field pickup equipment device and a method for acquiring a human voice signal in the far-field pickup equipment device. The far-field pickup system comprises an device equipment main body and a microphone pickup unit which are separated from each other. And tThe microphone pickup unit collects echoes of the user voice and the a sound signal played by the equipment devicemain body after spatial propagation and sends the echoes back to the device equipment main body for processing. Wherein tThe equipment device main body comprises a playing signal source; , a synchronization signal generator; , a horn; , a delay determination unit; and an echo cancellation unit. According to the embodiment of the invention, the problem that the microphone signal and the echo reference signal cannot be synchronized in the prior art can be solved, and the speech recognition performance is improved.

A method and an in-vehicle system having a speech recognition component are provided for improving speech recognition performance. The speech recognition component may have multiple vocabulary dictionaries, each of which may include phonetics associated with commands. When the in-vehicle system receives speech input, the speech recognition component may determine whether the received speech input includes a speech access command. If the received speech input is determined to include a speech access command, then a dictionary changing component may transition a currently-used dictionary of the speech recognition component to a vocabulary dictionary associated with the determined speech access command. Otherwise, the dictionary changing component may transition the currently-used dictionary to a first vocabulary dictionary. A command included in the received speech input may then be recognized by the speech recognition component using the transitioned currently-used dictionary.

An apparatus and method for improving the performance of voice recognition in a portable terminal are provided. The apparatus includes a voice recognition management unit, and a controller. After recognizing a user's voice and extracting at least one voice parameter, the voice recognition management unit determines if the extracted at least one voice parameter meets a criterion for determining one of success and failure of voice recognition. The controller analyzes a result of the determination by the voice recognition management unit and outputs a result of the analysis.

A weight coefficient generation device, a speech recognition device, a navigation system, a vehicle, a vehicle coefficient generation method, and a weight coefficient generation program are provided for the purpose of improving a speech recognition performance of place names. In order to address the above purpose, an address database 12 has address information data items including country names, city names, street names, and house numbers, and manages the address information having a tree structure indicating hierarchical relationships between the place names from wide area to a narrow area. Each of the place names stored in the address database 12 is taken as a speech recognition candidate. A weight coefficient calculation unit 11 of a weight coefficient generation device 10 calculates a weight coefficient of the likelihood of the aforementioned recognition candidate based on the number of the street names belonging to the lower hierarchy below the city names.

The present invention relates to a method and apparatus for improving spontaneous speech recognition performance. The present invention is directed to providing a method and apparatus for improving spontaneous speech recognition performance by extracting a phase feature as well as a magnitude feature of a voice signal transformed to the frequency domain, detecting a syllabic nucleus on the basis of a deep neural network using a multi-frame output, determining a speaking rate by dividing the number of syllabic nuclei by a voice section interval detected by a voice detector, calculating a length variation or an overlap factor according to the speaking rate, and performing cepstrum length normalization or time scale modification with a voice length appropriate for an acoustic model.

Provided are an intelligent voice recognition method, a voice recognition device, and an intelligent computing device. A method of intelligently recognizing a voice in a voice recognition device includes obtaining an ambient noise signal from the microphone detection signal when a voice is detected from the microphone detection signal; updating a previously learned noise cancellation model based on the ambient noise; and removing the ambient noise signal from the microphone detection signal. Therefore, a voice recognition performance of the voice recognition device can be maximized. At least one of the voice recognition device, the intelligent computing device and the server of the present disclosure may be associated with an Artificial Intelligence module, a drone (Unmanned Aerial Vehicle, UAV), robot, Augmented Reality (AR) device, virtual reality (VR) device and a device related to the 5G service.

A miniature microphone enclosure is provided for use with a wireless headset. The microphone enclosure provides superior noise and wind cancellation without any circuitry. Two sound ports and wind suppression material help minimize any noise attributable to wind or air. Moreover the microphone enclosure has a directivity such that the microphone does not have to be placed directly in front of a user's mouth.

A method and system are provided in which a decision tree-based model (“general model”) is scaled down (“trim-down”) for a given task. The trim-down model can be adapted for the given task using task specific data. The general model can be based on a hidden markov model (HMM). By allowing a decision tree-based acoustic model (“general model”) to be scaled according to the vocabulary of the given task, the general model can be configured dynamically into a trim-down model, which can be used to improve speech recognition performance and reduce system resource utilization. Furthermore, the trim-down model can be adapted/adjusted according to task specific data, e.g., task vocabulary, model size, or other like task specific data.