15 results about "Voice data" patented technology

A method and device for voiceprint recognition, include: establishing a first-level Deep Neural Network (DNN) model based on unlabeled speech data, the unlabeled speech data containing no speaker labels and the first-level DNN model specifying a plurality of basic voiceprint features for the unlabeled speech data; obtaining a plurality of high-level voiceprint features by tuning the first-level DNN model based on labeled speech data, the labeled speech data containing speech samples with respective speaker labels, and the tuning producing a second-level DNN model specifying the plurality of high-level voiceprint features; based on the second-level DNN model, registering a respective high-level voiceprint feature sequence for a user based on a registration speech sample received from the user; and performing speaker verification for the user based on the respective high-level voiceprint feature sequence registered for the user.

A network telephone system is provided with a distributed network, a network call processor, with the call processor connected to the network. A telephone line network interface is connected to a telephone line and is connected to the network for receiving packets from the network and sending packets to the network including packets with telephone voice data. A plurality of network telephones are part of the network telephone system with each network telephone connected to the network. Each network telephone has a display for displaying information and each network telephone is capable of engaging in a concurrent telephonic communication. Each network telephone has an I/O device in electrical communication with the network for receiving and sending packets to other devices connected to the network, an input device for producing audio signals from an input local to the device and a packet controller in electrical communication with the I/O device and the input device. The packet controller generates packets from the audio signals received by the input device, forwards the generated packets to the I/O device for transmission to the network and combines packets received by the I/O device to produce an audio signal with the combined packets and the audio signals from the local input device. Each network telephone displays information on the display corresponding to an identity of a source of packets combined whereby the identity may be selected for dropping a source form a concurrent telephonic communication.

A method of transmitting voice frames, via a transmission channel reserved for voice data, by a transmitting terminal generating voice frames using a voice signal is proposed. Such a method includes steps of: obtaining non-voice data; selecting voice data from the voice frame according to configuration data obtained beforehand relative to the transmission of non-voice data on the reserved channel; constructing of a degraded voice frame by replacing selected voice data with non-voice data; transmitting the degraded voice frame via the reserved channel to a receiving terminal. On the receiving terminal side, a method of managing voice frames coming from the transmitting terminal is proposed including the steps of: detecting a non-voice data header included in the voice frame; extracting, from the voice frame, of non-voice data, according to configuration data read in the header; transmitting of extracted non-voice data to a processor of non-voice data.

The invention relates to a fraud detection method and device for video data, computer equipment and a storage medium. The method comprises the following steps: acquiring to-be-detected video data; extracting image data of each video frame from the to-be-detected video data, and dividing the image data into a plurality of image data sets according to the time sequence of each video frame, the imagedata sets which comprises image data corresponding to continuous video frames; inputting each image data set into a pre-trained image feature extraction model to obtain an image feature vector; extracting voice data from the to-be-detected video data, and obtaining a voice feature vector of the voice data; performing cascade splicing on the image feature vector and the voice feature vector to obtain a multi-modal feature vector; and inputting the multi-modal feature vector into a pre-trained fraud detection model to obtain a fraud detection result corresponding to the to-be-detected video data output by the fraud detection model. By adopting the method, the characteristic information amount can be increased, the comprehensiveness and diversity of the characteristic information are improved, and the accuracy of video data fraud detection is effectively improved.

The present invention is to provide a voice-controlled display device configured such that the inputted user's speech is compared with the identification voice data assigned to each of the execution unit areas on a screen displayed through a display unit and, if there exists identification voice data corresponding to the user's speech, an execution signal is generated to the execution unit area to which the identification voice data is assigned to resolve the inconvenience that the user needs to learn the voice commands stored in the database and to apply the convenience and intuitive simplicity of user experience (UX) of the conventional touchscreen control to the voice control, and a method of voice control of the above display device

The invention discloses a children smartwatch and wechat client talkback method, comprising a wechat server and a transfer platform; the transfer platform comprises a transfer server; the talkback for a children smartwatch and a wechat client comprises the following steps: for downlink voice, the wechat client stores downlink voice data to the wechat server after recording; the wechat server sends a voice message notification to the transfer server; and the transfer server obtains the downlink voice data from the wechat server and then sends the downlink voice data to the corresponding children smartwatch; and for uplink voice, the children smartwatch uploads uplink voice data to the transfer server after recording; the transfer server pushes the uplink voice data to the wechat server; and then the wechat server sends the uplink voice data to the corresponding wechat client. By utilizing the method disclosed by the invention, parents can communicate with children very conveniently via wechat.

A data simplifying and merging method for a voice decoding memory system is disclosed. The method includes the steps of: reading a voice data from a non-volatile memory in a memory system; performing logic operation on the voice data in order to obtain an index; fetching corresponding decoded voice data in a table of the memory system in accordance with the index; and adding the decoded voice data to the voice data in order to obtain an original voice data.

The invention provides a conference voice data processing method and system, and relates to the field of voice recognition. The method comprises the following steps: arranging a plurality of acquisition modules near different participants, and acquiring identity information and initial voiceprint features of the corresponding participants according to the different acquisition modules so as to acquire speech of the corresponding participants; identifying and judging whether the voice contents of the plurality of speech are the same, if so, analyzing the sound intensity of the plurality of voice contents, and selecting the voice content with the highest sound intensity; establishing a voice feature model of the plurality of participants according to the identity information and the initial voiceprint features, and inputting the selected speech into the voice feature model to obtain an identity matching result; and judging whether the identity information is matched with the identity matching result according to the acquisition modules, and if not, selecting the same voice content of the corresponding acquisition modules according to the identity matching result. The accuracy of voice acquisition of participants can be improved, and the conference recording effect is improved.

The invention is applicable to the technical field of Internet of Things, and provides a method and a system for controlling a display content of a table lamp. The method comprises the following stepsthat a server receives user data collected by the table lamp, wherein the user data comprises user voice data, and the server analyzes the user requirements based on the user voice data; the server sends a preset display content corresponding to the user requirements to a transfer station; the transfer station receives the preset display content which is sent by the server and is corresponding tothe user information, and detects the current use condition of the table lamp and the position of the table lamp; the transfer station sends the preset display content corresponding to the table lampto the table lamp; the table lamp collects the user data and sends the user data to the server; and the table lamp receives and displays the display content to be sent, so that the table lamp in a control system of the table lamp display content can display the different information according to the different user requirements and the actual use condition of the table lamp, and thus the functionsof the table lamp are expanded, and the user experience is improved.

The invention provides a spoken language evaluation method and device based on phonemes and electronic equipment, and the method comprises the steps: setting an evaluation model; training the coding sub-model and the alignment output sub-model, and then training the recognition output sub-model under the condition of keeping the coding sub-model unchanged; inputting the target voice data into an evaluation model, determining a feature vector and phoneme alignment information of the target voice data, and determining an identification result of each phoneme in the target voice data based on the identification output sub-model; and determining an evaluation result of the target voice data according to the standard answer and the recognition result of each phoneme in the target voice data. According to the spoken language evaluation method and device based on the phonemes and the electronic equipment provided by the embodiment of the invention, the characteristics of mass data can be learned, and the phoneme alignment effect is relatively good; and a large amount of high-quality data is not needed, so that the problem of insufficient high-quality data volume in the training process is effectively solved.

The invention discloses a portable voice dynamic monitoring system based on Beidou No.3, wherein the system comprises a plurality of convenient terminals, the convenient terminals are connected through a satellite unit, the convenient terminals are connected with a ground data center through the satellite unit, and the satellite unit is in communication connection with a ground master station unit. Then the ground data center is connected with the background command management system through a network, wherein a satellite positioning technology, a Beidou short message communication technology and a voice data compression technology are fused by arranging the portable terminal of the portable voice dynamic monitoring system, the background command management system, the satellite unit, the ground master station unit and the ground data center, so that the actual operability and the information transmission efficiency are greatly improved by utilizing the portable terminal, real-time positioning, personnel scheduling and emergency message transmission in special occasions.

The invention provides a voice data expansion method and system, electronic equipment and a storage medium, and the technical scheme of the method comprises a dialogue design step: designing dialoguecontents in a text form according to a preset scene; a text conversion step: converting the dialogue content into dialogue voice; and an audio mixing step: mixing the dialogue voice with the noise inthe preset scene to obtain analog audio data, and outputting the analog audio data. The problems that an existing voice data expansion method is complex and not high in reliability are solved.

The present invention relates to an automatic speech recognizer and a speech recognition method using a keyboard macro function, the automatic speech recognition method using the keyboard macro function comprising the steps of: a transcriptional data generation unit generating transcriptional data on the basis of speech data input through a microphone; a tagging unit that, when a word to be tagged that is preset for executing a macro function is included in the generated transcriptional data, tags the generated transcriptional data, and performs conversation and storage; and displaying the generated transcriptional data in a space, in which data can be input, on a practical program in a virtual keyboard input manner. The automatic voice recognizer and the voice recognition method can provide the following effects: voice can be directly recorded in a text form under the condition that an electronic medical record, a paper operation record and the like are difficult to manually record when an actual patient is diagnosed and treated on a medical site, and convenience is provided.