47results about How to "Implement speech recognition" patented technology

The invention relates to the technical field of voice recognition, provides a voice recognition method and device and electronic equipment, which aim to solve the problem of relatively low voice recognition accuracy. The method comprises the following steps of acquiring a to-be-recognized voice, performing feature extraction on the to-be-recognized voice to obtain voice feature information, determining a target character sequence corresponding to the voice feature information according to the target acoustic model and the target language model, wherein the target language model comprises a first language model and a second language model, the first language model is obtained by performing language model training through a command word training text of a first scene, and the second languagemodel is obtained by performing language model training through a first text training set. In the voice recognition process, two language models are adopted, and the first language model is obtainedby performing language model training through the command word training text of the first scene, so that the recognition capability of the first language model for related command words in the first scene can be enhanced, and the voice recognition accuracy can be improved.

The invention relates to a recording realization method and device in conversation process. The method comprises the following steps: determining telephone number of the current communication is a preset telephone number; if the telephone number is the preset telephone number, determining whether the voice content of the current communication comprises a recording start keyword; and if the voice content comprises the recording start keyword, controlling a user terminal to start recording the current communication. The technical scheme can effectively control the user terminal to start the recording function automatically according to the communication content to record user-required information in the conversation process.

The invention relates to a general remote controller and aims at providing a multifunctional general remote controller which supports multiple modes and can be defined according to user requirements. The general remote controller comprises a processor, an operation system, an infrared emitting tube, a radio frequency module, a ZigBee module, a radio frequency module group, a gyroscope, a gravity sensor, a displayer, a thin film transistor (TFT) capacitor screen, a lithium ion battery, a player and a sound pickup. The operation system is installed in the processor, and the infrared emitting tube, the radio frequency module, the ZigBee module, a radio frequency module group, the gyroscope, the gravity sensor, the displayer, the TFT capacitor screen, the lithium ion battery, the player and the sound pickup are respectively connected with the processor through serial input/output ports.

The embodiment of the invention provides a speech recognition model training method, a speech recognition method and related devices. The training method comprises the steps of: determining a trainingcurrent mixed language audio, obtaining the training initial acoustic features, utilizing a first language module to obtain a training first time sequence position acoustic feature, utilizing a second language module to obtain a training second time sequence position acoustic feature, performing fusion and text coding on the training first time sequence position acoustic feature and the trainingsecond time sequence position acoustic feature to obtain a training current fusion text feature, obtaining a first training current prediction text feature according to the training current fusion text feature and a previous reference text feature, obtaining first loss according to the first training current prediction text feature and a current reference text feature, then obtaining model loss, and adjusting the parameters of a speech recognition model according to the model loss until the trained speech recognition model is obtained. According to the voice recognition model training method,the voice recognition method and the related devices provided by the embodiment of the invention, the voice recognition accuracy can be improved.

The invention discloses a smart voice cell phone or smart voice tablet computer and relates to the technical field of electronic devices. The smart voice cell phone or smart voice tablet computer comprises a cloud server and an electronic device. A voice wake-up module, a voice recognition module, a voice command module, a semantic analysis module, a voice synthesis module and a program control module are arranged in the electronic device. The voice recognition module is connected with the cloud server through a wired communication module. The voice recognition module is connected with the voice command module through an information analysis and processing and signal conversion and transmission module. The voice command module is connected with the program control module through an information processing and signal transmission module. By providing the voice wake-up module, the voice recognition module, the voice command module, the semantic analysis module, the voice synthesis moduleand the program control module in the electronic device, the smart voice cell phone or smart voice tablet computer achieves voice interaction, directly wakes up programs through the voice wake-up module without manually opening the applications, realizes software application through a voice interaction function easy to operate.

A method for identifying voice of mobile communication terminal includes establishing video ¿C audio model by utilizing triphone, registering telephone number, forming triphone data for registered telephone number name or firm name and storing it, inputting user voice, comparing inputted voice data with stored triphone data, dialing out telephone number of inputted voice if identification is successful.

The embodiment of the invention provides a voice recognizing method and a voice recognizing device based on a block chain, a medium and an electronic device. The voice recognizing method based on theblock chain comprises the following steps: storing labelled historical voice data in the block chain; and if a new block of the current voice information is generated in the block chain, triggering anacoustic model and a linguistic model finishing through training to recognize the current voice information, and outputting the current text information corresponding to the current voice information, wherein the acoustic model and the linguistic model are obtained through training by adopting the historical voice data labelled in the block chain. According to the technical scheme of the embodiment of the invention, the voice data of the user is stored by adopting the block chain technology, and the recognition for the current novice is carried out based on the stored historical voice data.

The invention discloses an intelligent financial counseling robot convenient to use and relates to the technical field of financial counseling equipment. The intelligent financial counseling robot convenient to use comprises a bottom plate. The top of the bottom plate is fixedly connected with a crate. A transmission case is connected between the two sides of the inner wall of the crate in a sliding mode. The bottom of the inner wall of the transmission case is fixedly connected with a motor through a connecting block, and the outer surface of an output shaft of the motor is fixedly connectedwith a first bevel gear. A bidirectional threaded rod is rotationally connected between the two sides of the inner wall of the transmission case through bearings, and the outer surface of the middle of the bidirectional threaded rod is fixedly connected with a second bevel gear adapting to the first bevel gear. According to the intelligent financial counseling robot convenient to use, people can use the robot quite conveniently, the height of a touch screen can be automatically adjusted according to the height of a client and the distance to the client, the purpose that the robot is suitable for different clients is well achieved, and accordingly, the client can use the intelligent financial counseling robot to handle the financial counseling business quite conveniently.

The invention discloses an automatic voice data labeling method and system for voice recognition, and particularly relates to the field of voice recognition. The system comprises a mute detection module, a volume screening module, a length screening module, a voice recognition module, a recognition result judgment module and a manual proofreading module, the mute detection module splits each voiceinto a plurality of voice segments through a mute detection algorithm, and the volume screening module is used for screening out voices meeting the requirements through a volume threshold value and removing voices not meeting the requirements. The invention discloses a combined system of multiple modules. According to the system, speech preprocessing and speech recognition are carried out, by a public cloud mode, recognition result judgment manual proofreading are carried out, voice data annotation is constructed, after multiple times of iteration of the processes, a new corpus is continuously trained, high-quality corpus data is obtained, manpower is reduced, the voice data annotation quality is improved, and the problems that the manual annotation period is long, the cost is high and the efficiency is low are solved.

The invention discloses a Chinese voice interaction non-inductive control system based on Raspberry Pi edge calculation and a method thereof, and the system comprises an edge end, a mobile end, an external control module, and an edge calculation detection and scheduling module. According to the invention, the sizes of a speech recognition model and a speech synthesis model are small; the edge calculation and offline work can be realized, the equipment is directly deployed in mobile terminal equipment without depending on a network, the functions of voice synthesis and voice recognition can berealized under an offline condition, the problems of voice recognition and interaction functions under severe conditions such as unsmooth network or attack are solved, and the function of Chinese voice recognition interaction under a severe environment is realized.

The invention discloses an intelligent speech recognition recovery circuit and a recovery method, the recovery circuit comprises a speech control module, an audio processing module and an output module, the speech control module sends speech data to the audio processing module and receives the speech data; the audio processing module receives the speech data sent by the speech control module and further processes the speech data; and the output module receives and outputs the speech data further processed by the audio processing module. Preferably, the recovery circuit further comprises a sound effect algorithm processing module which is used for receiving the speech data sent by the speech control module, carrying out sound effect processing and returning the processed speech data to thespeech control module. According to the intelligent speech recognition and recovery circuit and the recovery method provided by the invention, the speech recognition function of a non-intelligent speech recognition product and an intelligent speech recognition product can be realized by saving the sampling circuit and the ADC conversion circuit.

The invention is suitable for the technical field of communication, and provides a voice recognition method applied to a user terminal and terminal equipment. The method includes the steps that voiceinformation input by a user is obtained, and corresponding character information is recognized according to the voice information; an interface type of a current display interface is determined and includes a list interface and a playing interface; when the current display interface is the list interface or the playing interface, the character information is matched with a regular expression correspnidng to the interface type of the current display interface. Semantic analysis is conducted at the user terminal, and will not be conducted in a semantic server, the process of semantic analysis can be simplified, and the time for semantic analysis is shortened; according to the determined regular expression selectively corresponding to the current display interface, regular expressions for matching in semantic analysis of the user terminal can be reduced, matching time is accordingly shortened, the time for semantic analysis of the user terminal is further shortened, the efficiency of recognizing voice is improved, and user experience is promoted.

The invention relates to a multifunctional intelligent wireless charging operation method controlled by a computer. A charging device, a transmitting coil device, a power storage device, a monitoringdevice, an audio device, a central control device, a storage device, a communication device and an interaction device are included. The monitoring device collects the state data of a power bank in real time and stores the state data to the storage device, and the current working state is adjusted according to an optimal resource regulation and control model after the state data is processed by thecentral control device. According to the present invention, the intelligent voice control and audio playing functions are realized, and the interaction device can check the state of the multifunctional intelligent wireless power bank and control the multifunctional intelligent wireless power bank remotely.

The invention discloses a paint spraying robot voice recognition method based on a multi-scale enhanced BiLSTM model. The method comprises the following steps: 1) acquiring common spraying sound instructions by using a signal acquisition system, wherein NI-9234 is selected as a data acquisition card; 2) repeatedly adding Gaussian white noise to the collected audio signal for 100 times, generating a noisy signal, solving a corresponding Mel spectrum sequence, and then solving an average sequence of 100 Mel spectrum sequences; 3) performing feature extraction on the average Mel spectrum sequence by using a multi-scale convolution filter, and then performing further mining on the extracted features by using a BiLSTM model to obtain corresponding output; 4) splicing outputs of the BiLSTM model together, then inputting the spliced outputs to a full connection layer and a Softmax layer for processing, and finally realizing speech recognition in combination with a CTC algorithm. and 5) embedding the model obtained through training in the steps 1) to 4) into a spraying robot, so that corresponding spraying tasks are intelligently achieved. According to the model, the intelligent voice recognition function of the spraying robot can be realized, and the model has very high practical application value.

The invention relates to the field of voice recognition and particularly relates to a method, a system and a device for realizing voice age and/or gender recognition service and a medium, and aims to solve technical problems of remote accurate calling and simple and convenient deployment of an existing voice age and/or gender recognition model. Therefore, a terminal calls a server through a serialized voice age/gender identification request under a predefined GRPC framework, and the server identifies the age/gender through a set age/gender voice identification service; the corresponding voice age/gender recognition deep neural network model is accurately selected to decode and determine the age and/or gender information of the target object, and the age and/or gender information is returned to the terminal. Due to the fact that the age and/or gender service mode and the remote calling architecture are set, the corresponding model is called after the type of the model is determined, calling is more accurate and does not need to depend on a fixed frame, the method is more flexible, expandability is high, the resource utilization rate is high, concurrency is high, and meanwhile iterative updating of the algorithm model is facilitated.

The invention discloses a voice recognition method and system based on comparative predictive coding, and belongs to the technical field of voiceprint recognition, and the method is characterized in that the method comprises the following steps: S1, collecting A voice files of each voice category, carrying out the preprocessing of each voice file, and obtaining PCM-coded voice time sequence data; s2, constructing a pairing data set of the voice time sequence data; s3, constructing a paired fragment data set; s4, constructing an artificial neural network; s5, training a speech recognition network formed by the first converter, the second converter and the one-dimensional convolutional neural network; and S6, voice recognition is carried out through the voice recognition network. According to the method, a large amount of insufficient voice data acquired by a background is fully utilized, the voice data are regarded as time sequence data, end-to-end conversion is directly realized, and the extraction of voice time sequence data characteristics is not needed.

The invention relates to the technical field of speech recognition, and discloses a speech recognition method and device, equipment and a computer readable storage medium. According to the invention, if first voice information is monitored to be received, acoustic feature extraction is carried out on the first voice information to obtain first acoustic feature information, and then a decoder is utilized to decode the first acoustic feature information to obtain a decoding recognition result; and mode matching is performed on word strings in the decoding recognition result and word strings in a preset proper noun vocabulary to obtain a matching result, and the matching result is output to realize voice recognition. The problem that the accuracy of speech recognition is poor in the prior art is solved.

The invention discloses auxiliary computer hardware and particularly provides an anti-theft wireless USB flash disk with a voice recognition function. The anti-theft wireless USB flash disk comprises a storage end and a receiving end, wherein the storage end comprises a storage end shell, a flash memory unit, a storage end microprocessor, a storage end wireless receiving-transmitting module and a rechargeable battery; the receiving end comprises a receiving end shell, a receiving end USB interface, a receiving end wireless receiving-transmitting module and a receiving end microprocessor; and the anti-theft wireless USB flash disk further comprises a voice recognition device, a photocell, a photoelectric charger, a displacement sensor, an audible and visual alarm and a photosensitive sensor. According to the anti-theft wireless USB flash disk, the voice recognition can be realized, and photoelectric charging, automatic formatting and leakage prevention after stealing, and abnormal photoelectric alarm also can be realized, so that the information safety can be brought to a user; and the anti-theft wireless USB flash disk is simple to operate and is very convenient to use, and the individualized requirements of markets can be met.

The invention discloses a playing method of a sound control vehicular player. First of all, a parallel table of player control driving signals and entry codes is established; then, a database of sound instruction sample characteristic values and the entry codes is established; next, a BP network model is established, the BP network model is enabled to input the sound instruction sample characteristic values and output the entry codes, data in the database of the sound instruction sample characteristic values and the entry codes is input into the BP network model, and the BP network model is trained until network errors of the BP network model are smaller than and equal to a preset error threshold; and during work, a voice instruction of a driver is acquired, after the voice instruction of the driver is preprocessed, a characteristic value is calculated and then input to the BP network model, a corresponding entry code is obtained, retrieval is carried out in the parallel table, and a corresponding driving control signal is found for driving a player to work. According to the invention, the application is convenient, the driver does not have to move for operating the player, and thus the driving safety is improved.

The invention discloses a transforming device from speech identification to video, which comprises: an identifying code establishing module which is used for establishing a corresponding identifying code according to the types of video resource when a media server is started; an audio stream receiving module which is connected with the identifying code establishing module and is used for establishing a connecting channel of the audio stream and receiving the audio stream after the media server receives the request of an application server; a speech identifying module which is connected with the audio stream receiving module and is used for identifying audio data and outputting the identified data to a transformation processing module; the transformation processing module which is connected with the speech identifying module and the identifying code establishing module and is used for transforming the received data of the speech identifying code and making comparison between the transformed data with an identifying code established by the identifying code establishing module, thus realizing video transformation; a video stream output module which is connected with the transformation processing module and used for outputting the transformed video stream to a terminal unit through network.