91 results about "Language speech" patented technology

In some speech recognition applications, not only the language of the utterance is not known in advance, but also a single utterance may contain words in more than one language. At the same time, it is impractical to build speech recognizers for all expected combinations of languages. Moreover, business needs may require a new combination of languages to be supported in short period of time. The invention addresses this issue by a novel way of combining and controlling the components of the single-language speech recognizers to produce multilingual speech recognition functionality capable of recognizing multilingual utterances at a modest increase of computational complexity.

The invention discloses a cross-language end-to-end speech recognition method for the low resource Tujia language. The method comprises the following steps: preprocessing the Tujia language data; constructing a cross-language Tujia language corpus; establishing a unified coding dictionary of Chinese international phonetic alphabets and national international phonetic alphabets; establishing a cross-language end-to-end Tujia speech recognition model; and using a join temporal classification model and performing decoding under the action of the coding dictionary so as to obtain the recognition result. The recognition model with higher generalization is constructed by taking advantage of sufficient major language data and combining the idea of transfer learning so as to improve the accuracy of Tujia language speech recognition.

A machine translation apparatus includes a translating unit that translates a source language document into a translated document described in a target language, and creates an ambiguous portion that is a word or a sentence having an ambiguity occurred during translation; a storing unit that stores the translated document and the ambiguous portion; a speech receiving unit that receives a speech in the source language; a recognition unit that recognizes the speech received and creates a source language speech sentence as a recognition result; a translation updating unit that updates the translated language document by retranslating a part of a speech content of the source language speech sentence to which the ambiguous portion corresponds, when the source language document includes the ambiguous portion; and a display control unit that displays the updated translated document on a display unit.

The speech recognition apparatus ( 1 ) is equipped with the garbage acoustic model storage unit ( 110 ) storing the garbage acoustic model which learned the collection of the unnecessary words; the feature value calculation unit ( 101 ) which calculates the feature parameter necessary for recognition by acoustically analyzing the unidentified input speech including the non-language speech per frame which is a unit for speech analysis; the garbage acoustic score calculation unit ( 111 ) which calculates the garbage acoustic score by comparing the feature parameter and the garbage acoustic model; the garbage acoustic score correction unit ( 113 ) which corrects the garbage acoustic score calculated by the garbage acoustic score calculation unit ( 111 ) so as to raise it in the frame where the non-language speech is inputted; and the recognition result output unit ( 105 ) which outputs, as the recognition result of the unidentified input speech, the word string with the highest cumulative score of the language score, the word acoustic score, and the garbage acoustic score which is corrected by the garbage acoustic score correcting means.

The invention provides a speech recognition interface unit and a speech recognition method thereof. The speech recognition interface unit comprises a model parameter estimator and a speech recognizer. The model parameter estimator extracts the acoustic information of a target language and a native language from an exterior speech database and acquires an acoustic model of the target language and an acoustic model of the native language through respective training; and then the module parameter estimator applies model self-adaptive technology respectively to the two acoustic models and obtains a background model through the application of gauss component consolidation technology. The speech recognizer respectively receives the background model inputted from the model parameter estimator and exteriorly inputted non-native language speech and recognizes the inputted non-native language speech based on the background model. The speech recognition interface unit and the speech recognition method can improve the recognition accuracy of the non-native language speech.

The invention discloses a multi-to-multi speaker conversion method based on STARGAN and an x vector, which comprises a training stage and a conversion stage, wherein a speech conversion system is achieved by combining the STARGAN and the x vector, the personality similarity and quality of the converted speech can be greatly improved, particularly, for the short-time utterance, the x vector has better characterization performance and better speech conversion quality can be achieved, meanwhile, the problem of over-smoothing in C-VAE can be overcome, and a high-quality speech conversion method isachieved. In addition, the method can achieve the speech conversion under the condition of non-parallel text, the training process does not need any alignment process, the universality and practicability of a speech conversion system are improved, and the method can also achieve that the conversion system with multiple source-target speaker pairs is integrated in a conversion model, namely, the multi-speaker-to-multi-speaker conversion is achieved, and the system has a better application prospect in the fields of cross-language speech conversion, film dubbing, speech translation and the like.

The present invention relates to a method for modeling a common-language speech recognition, by a computer, under the influence of multiple dialects and concerns a technical field of speech recognition by a computer. In this method, a triphone standard common-language model is first generated based on training data of standard common language, and first and second monophone dialectal-accented common-language models are based on development data of dialectal-accented common languages of first kind and second kind, respectively. Then a temporary merged model is obtained in a manner that the first dialectal-accented common-language model is merged into the standard common-language model according to a first confusion matrix obtained by recognizing the development data of first dialectal-accented common language using the standard common-language model. Finally, a recognition model is obtained in a manner that the second dialectal-accented common-language model is merged into the temporary merged model according to a second confusion matrix generated by recognizing the development data of second dialectal-accented common language by the temporary merged model. This method effectively enhances the operating efficiency and admittedly raises the recognition rate for the dialectal-accented common language. The recognition rate for the standard common language is also raised.

A system for translating speech from at least two source languages into another target language provides direct speech to target language translation. The target text is converted to speech in the target language through a TTS system. The system simplifies speech recognition and translation process by providing direct translation, includes mechanisms described herein that facilitate mixed language source speech translation, and punctuating output text streams in the target language. It also in some embodiments allows translation of speech into the target language to reflect the voice of the speaker of the source speech based on characteristics of the source language speech and speaker's voice and to produce subtitled data in the target language corresponding to the source speech. The system uses models having been trained using (i) encoder-decoder architectures with attention mechanisms and training data using TTS and (ii) parallel text training data in more than two different languages.

The invention provides an anthropomorphic oral translation method with a man-machine communication function. The method comprises the following steps of conducting intelligent speech recognition of source language speech and obtaining source language text; processing the source language text and a communication scene, and conducting anthropomorphic man-machine communication; conducting machine translation to obtain a translation result. The invention further provides an anthropomorphic oral translation system with the man-machine communication function. Through the adoption of the system, man-machine communication with a user needs to be conducted according to a translation task if necessary, the information used for obviously improving translation experiences of the user in a complex application scene is obtained accurately, and the accuracy of translation semantics is improved.

The invention discloses cross-language speech recognition method and device, and relates to the technical field of speech processing. A concrete implementation mode of the method comprises the steps of acquiring cross-language sample data, training the sample data as input data of a preset neural network model, and obtaining a language category discriminator; inputting audio to be recognized intothe language category discriminator, and segmenting the audio to be recognized according to a language category determined by the language category discriminator; and utilizing a recognition engine corresponding to the determined language category for recognizing the segmented audio to be recognized. According to the implementation mode, an existing speed recognition engine has no need to be modified, so that the cost is low, the recognition rate is high, and the accuracy is high.

The invention relates to a TV (television) system with multi-language speech translation and a realization method of the TV system. The realization method is characterized by comprising the following steps: encoding, decoding and processing a speech of the TV system by using a DSP (digital signal processor) module in the TV system, so as to obtain character information; then distributing the character information to at least two speech translation modules according to a process set by the system; and sending translated speech data to an earphone assorted with a wireless receiving module through a corresponding wireless transmission module by the speech translation modules. With the adoption of the TV system provided by the invention, the speech in a TV can be translated into various languages simultaneously. The TV system is more suitable for humanized design. With the adoption of the TV system provided by the invention, people can more conveniently know about economical and cultural information in the other nations well. The TV system is more suitable for globalized development, and has better practical values.

The invention provides a simultaneous interpretation method and device, and computer equipment. The simultaneous interpretation method comprises the steps: acquiring a source language voice signal tobe translated; carrying out speech recognition on the source language speech signal to generate a source language vocabulary sequence and a source language pinyin sequence; inputting the source language vocabulary sequence and the source language pinyin sequence into corresponding encoders respectively, and obtaining a vocabulary vector sequence and a pinyin vector sequence corresponding to the source language voice signal; inputting the vocabulary vector sequence and the pinyin vector sequence into a decoder; generating a target language sequence corresponding to the source language voice signal; and due to the fact that the source language pinyin sequence generally does not make mistakes, determining the target language sequence corresponding to the source language voice signal by combining the source language pinyin sequence, so that part of errors in the source language vocabulary sequence can be corrected, and the simultaneous interpretation efficiency is improved, and the fault-tolerant capability of voice recognition errors is improved.

The embodiment of the invention discloses a speech translation method and device, electronic equipment and a storage medium. The method comprises the steps that source voice corresponding to a language to be translated is acquired; acquiring a specified target language; inputting the source voice and indication information matched with the target language into a pre-trained voice translation model, wherein the voice translation model is used for translating languages in a first language set into languages in a second language set, the first language set comprises a plurality of languages, thefirst language set comprises the language to be translated, the second language set comprises a plurality of languages, and the second language set comprises the target language; and obtaining translation voice output by the voice translation model and corresponding to the target language, the language to be translated being different from the target language. According to the embodiment of the invention, multi-language speech translation can be realized, and the efficiency and accuracy of speech translation are improved.

The invention discloses a multi-speaker and multi-language speech synthesis method. The method comprises the following steps: extracting speech acoustic features; processing the texts in different languages into a unified representation mode, and aligning the audio with the texts to obtain duration information; constructing a speaker space and a language space, generating a speaker id and a language id, extracting a speaker vector and a language vector, adding the speaker vector and the language vector into the initial speech synthesis model, and training the initial speech synthesis model byusing the aligned text, duration information and speech acoustic features to obtain a speech synthesis model; processing the to-be-synthesized text to generate the speaker id and the language id; andinputting the speaker id, the text and the language id into a speech synthesis model, outputting speech acoustic features and converting the speech acoustic features into audio. A system is also disclosed. According to the method, unentanglement of the characteristics of the speaker and the language characteristics is realized, and conversion of the speaker or the language can be realized only bychanging the id.

The invention discloses a system for realizing real-time speech mutual translation. The system comprises an intercom device and a mobile communication device; the intercom device acquires a speech signal, converts the speech signal into speech data, and sends the speech data to the mobile communication device; the mobile communication device or a cloud server converts the speech data into statements and characters after performing speech recognition on the speech data, translates the statements and characters into a text of a target language, performs speech synthesis processing and generatesplayable speech data for playing in the mobile communication device; the mobile communication device obtains the speech signal, converts the speech signal into speech data, converts the speech data into the statements and characters after performing speech recognition on the speech data, translates the statements and characters into the text of the target language, performs the speech synthesis processing and generates playable speech data; and the playable speech data are sent to the intercom device for playing. By adoption of the system provided by the invention, a language speech can be translated into a speech signal in another language in real time, which facilitates two-way communication between people in different languages.

The invention discloses a multilingual speech recognition model training method, and relates to the field of artificial intelligence, and the method comprises the steps: carrying out the training of aspeech recognition model through a first language, and obtaining an initial speech recognition model; building an adaptive network function, and embedding the adaptive network function into a hiddenlayer of the initial speech recognition model to obtain an initial multilingual speech recognition model; performing model training on the initial multilingual speech recognition model through the speech data of the second language to obtain a training result; and iteratively updating the initial multilingual speech recognition model until the training result falls into a preset standard trainingresult range, and outputting the multilingual speech recognition model. In addition, the invention also relates to a blockchain technology, and the voice data of the first language and the voice dataof the second language can be stored in the blockchain. According to the invention, the adaptive network function is embedded into the hidden layer of the initial speech recognition model, so that thetraining efficiency of the multi-language speech recognition model can be improved.