510 results about "Subvocal recognition" patented technology

The embodiment of the invention discloses a speech recognition method and a related device, and at least relates to a speech recognition technology in artificial intelligence, speech data to be recognized are used as input data of a time delay neural network in an acoustic model, and an output layer of the time delay neural network comprises acoustic modeling units corresponding to a plurality of syllables respectively, so that the speech recognition efficiency is improved. And the syllable probability distribution corresponding to the voice frames included in the voice data can be obtained by taking the syllables as the recognition granularity through the time delay neural network. When syllable recognition is carried out through the output layer, auxiliary judgment can be carried out on the syllables to which the voice frames belong on the basis of pronunciation rules in combination with front and back syllable information of the voice frames, so that more accurate syllable probability distribution is output. Moreover, since the syllables are generally composed of one or more phonemes, the method has higher fault-tolerant capability, not only can more accurately determine the speech recognition result based on the probability distribution of the syllables, but also has low requirements for the quality of the speech data to be recognized, and effectively expands the application scenarios of the speech recognition technology.

This invention provides a speech recognition method, system, device, and medium for elevator entrapment scenarios. The method includes: acquiring speech data in an elevator scenario and preprocessing the speech data to obtain a first speech feature; using the first speech feature as input to a deep neural network to output recognized text and an entrapment probability value; determining entrapment based on the recognized text and the entrapment probability value, and outputting an entrapment determination result. By using a deep neural network and a specific encoding and decoding process, this method can more accurately process speech data in elevator scenarios, thereby improving the accuracy of entrapment detection. By comparing the entrapment probability value with a threshold and combining it with the recognized text for comprehensive judgment, this method can reduce false detections and false negatives, improving the accuracy of entrapment determination.

Systems and methods are provided for accessing a factorized neural transducer comprising a first set of layers for predicting blank tokens and a second set of layers for predicting vocabulary tokens. The first set of layers comprises a blank predictor, an encoder, and a joint network and the second set of layers comprising a vocabulary predictor which is a separate predictor from the blank predictor. A context encoder is added to the factorized neural transducer which encodes long-form transcription history for generating a long-form context embedding, such that the factorized neural transducer is further configured to perform long-form automatic speech recognition, at least in part, by using the long-form context embedding to augment a prediction of vocabulary tokens.

The application provides a clothes drying machine voice recognition test method and system, comprising: when a clothes drying machine voice recognition test instruction is received, obtaining a voice test script corresponding to the clothes drying machine voice recognition test instruction and a voice response script corresponding to the voice test script; accessing an interface of a to-be-tested clothes drying machine and obtaining a function list stored in advance by the to-be-tested clothes drying machine, wherein the function list comprises a plurality of function instructions of the clothes drying machine; comparing the voice test instruction with the function instruction, constructing a voice test instruction set according to a comparison result, playing a voice test instruction in the voice test instruction set, and obtaining a response test instruction corresponding to the to-be-tested clothes drying machine; comparing the response test instruction with the voice response instruction, and generating a clothes drying machine voice test result. The application can effectively solve the problem that voice test scripts are not compatible due to function differences of different models of clothes drying machines.

Disclosed herein is a Voice-Enabled AI Chat Agent, optimized for use within web browsers. This advanced AI system facilitates natural, voice-driven interactions, allowing users to engage in conversations through speech instead of traditional text input. It features a sophisticated voice recognition module, a dynamic natural language processing engine, and a versatile role adaptation mechanism, enabling it to assume various user-defined roles such as a tutor, doctor, or counsellor. Designed with an emphasis on accessibility and user-friendliness, this AI chat agent represents a significant advancement in human-computer interaction, making digital communication more intuitive and accessible for a wide range of users.

The application provides a speech recognition model training method, a speech recognition method and device, and relates to the technical field of speech processing. The method comprises the following steps: in each iteration process, a training sample set is obtained, the training sample set comprises multiple training samples, each training sample comprises a sample speech signal, a text corresponding to the sample speech signal and a label of the text, the label is used for indicating whether the text is a complete sentence, for each training sample in the training sample set, taking the acoustic feature of the sample speech signal in the training sample as the input of a speech recognition model, outputting the speech recognition text and the predicted label of the sample speech signal, adjusting the parameters of the speech recognition model according to the speech recognition text and the predicted label of the sample speech signal obtained in each iteration process, and the text corresponding to the sample speech signal and the label of the text, until a stop training condition is met, and a trained speech recognition model is obtained. The accuracy of continuous speech recognition can be improved.

We provide the system. [Solution] A means of receiving voice instructions from a user and converting them into text using speech recognition technology, A means for generating an information gathering task based on the converted text and distributing it to an information sharing device, A means of collecting input from an information sharing device and analyzing the data through statistical processing, Means for providing analysis results to users, An information processing system that includes this.

The application provides a speech emotion recognition model training method, a recognition method and device, and relates to the technical field of speech recognition. The method comprises the following steps: when training a speech emotion recognition model, a plurality of speech sample pairs can be obtained, each speech sample pair comprising a first speech sample with a speech emotion label and a second speech sample without a speech emotion label, and the first speech sample and the second speech sample belong to different languages; the speech features corresponding to the speech sample pairs are input into an initial speech emotion recognition model to obtain the prediction results corresponding to the first speech sample and the second speech sample in the speech sample pairs; and the model parameters of the initial speech emotion recognition model are updated according to the first prediction results, the second prediction results and the speech emotion labels corresponding to each speech sample pair. The speech emotion recognition model obtained through the training can accurately recognize the speech emotions of different languages, thereby improving the accuracy of the recognition results.

This invention relates to the field of smart healthcare technology and discloses a method and system for verifying clinical nursing information based on speech recognition. The method includes the following steps: an edge device acquires a nursing item script; responds to a voice wake-up command, driving the terminal to read the script item by item; collects real-time voice recordings and performs beamforming noise reduction, using semantic vectors to accurately extract assessment information and fill in the slots; intelligently identifies missing items and triggers dynamic follow-up questions; summarizes and generates structured confirmation information for repetition; if a modification command is detected, a pointer callback is triggered to jump to re-record, and after confirmation, the information is encrypted and uploaded to the cloud. This invention, through a cloud-edge-device collaborative architecture and breakpoint protection technology, achieves contactless entry and closed-loop verification of nursing records. This solution supports interrupted resume transmission and non-linear logic jumps, effectively solving the pain points of easily interfered clinical operations and cumbersome manual recording. While ensuring aseptic operation, it significantly improves the efficiency and accuracy of clinical document generation.

We provide the system. [Solution] A means of acquiring knowledge from elderly people as voice or text data using speech recognition technology, A method for analyzing the acquired data using natural language processing algorithms and classifying it into general themes, A means of storing classified data on a digital information storage medium and providing it to users as needed, A means of providing an interface for sharing knowledge about the elderly among caregivers, A system that includes this.

This application discloses an AI-based intelligent diagnostic and prescription recommendation system based on famous prescriptions from renowned doctors throughout history. The system includes a human-computer interaction interface, a natural language processing module, a semantic analysis module, a speech recognition module, an image processing module, a medical terminology processing module, a feature extraction and pattern analysis module, and a prescription screening module. The speech recognition module converts captured speech signals into text information. The natural language processing and semantic analysis modules perform semantic understanding on the text information and extract structured TCM syndrome feature elements. The image processing module acquires and processes facial and tongue images of patients, recognizing visual features. The medical terminology processing module connects to a standard TCM terminology database, standardizing and mapping TCM syndrome feature elements and visual features. The feature extraction and pattern analysis module uses a deep learning model to perform diagnostic logic reasoning based on the standardized mapped features and outputs the diagnostic results.

This invention discloses a secure voice-based intelligent customer service system and interaction method based on a multi-agent architecture. The system includes an ASR (Automatic Speech Recognition) module, multi-layered security barriers, a main retrieval agent, a specialized retrieval agent cluster, a parallel retrieval and result aggregation module, an LLM (Limited Language Modeling) synthesis engine, and a TTS (Text-to-Speech) module. Through multi-agent system architecture and methodological design, this invention achieves modular processing, parallel retrieval tasks, and closed-loop security control. By enabling independent calls between input, output, agents, and inference stages via multi-layered security barriers, the system effectively mitigates risks such as prompt word injection, indirect injection attacks, harmful content generation, and information leakage. Sensitive information identification, access permission verification, and content compliance review are embedded in each stage of input, retrieval, information synthesis, and output to prevent sensitive data leakage and meet the application requirements of high-security business scenarios (such as finance, telecommunications, and government services).

A method for audio processing performed by an electronic device is provided. The method includes detecting at least one of a first voice activity near a first device and a second voice activity near a second device using a voice recognition module associated with the first device, comparing the first voice activity and the second voice activity to determine whether the first voice activity and the second voice activity exceed a predetermined threshold, and outputting the at least one of the first voice activity or the second voice activity through the second device in response to determining that the first voice activity and the second voice activity exceed the predetermined threshold.

This invention relates to an interactive method and system for energy storage systems based on a voice-controlled model. The method includes: receiving user voice commands, performing voice recognition and semantic parsing, and extracting control intentions and parameters; matching or generating system operation strategies based on the parsing results; and decomposing the strategies into device control commands and issuing them for execution. The system includes modules for voice input, processing, escape extraction, strategy escape, strategy decomposition, and output, supporting multilingual recognition, local offline operation, strategy customization, and security verification. This invention lowers the barrier to entry for new energy storage systems, improves interactive intelligence and security, and is suitable for residential, commercial, off-grid, and mobile energy storage scenarios.

We provide the system. [Solution] A means of acquiring audio information and converting it into text information using speech recognition, A means of analyzing textual information and generating a summary, A method for automatically extracting work items and action items from a summary and registering them in the task management system, A means for translating summaries and work items into multiple specified natural languages, A means of saving the generated data and making it accessible in real time, A method for generating natural language meeting minutes in real time and making them immediately available to meeting participants. A system that includes this.

1. The name of the design product: graphical user interface (AI translation) of real-time translation application of electronic device. 2. The use of the design product: for an electronic device. 3. The design points of the design product: the graphical user interface content displayed by the electronic device. 4. The picture or photo that best indicates the design points: front view. 5. The use of the graphical user interface: for providing AI translation. 6. The human-computer interaction mode of the graphical user interface: clicking, staying, and sliding the relevant function buttons on the screen. 7. The change state description of the graphical user interface: the front view is the main interface of the real-time translation application, the AI translation module in the front view includes earphone mode, external speaker mode, and face-to-face translation, clicking the earphone mode of the AI translation module in the front view enters the interface change state diagram 1; clicking the button at the bottom of the interface change state diagram 1 to start collecting voice, after single collection is completed, it is converted into interface change state diagram 2; clicking the translation language selection button at the top center of the interface change state diagram 2 enters the interface change state diagram 3 for user to select the recording language; clicking the translation language in the interface change state diagram 3 enters the interface change state diagram 4 for user to select the translation language; wherein, “XXXX” in the front view shows the connected earphone model, “AAAAAAAA” shows the MAC address, and the area covered by the gray translucent mask is the variable content picture; “ZZZ” in the interface change state diagram 2 is the text content to be translated by voice recognition, “YYY” is the translated text content, and clicking the clear button at the upper right corner of the interface change state diagram 2 can clear the text content in the interface.

This disclosure relates to a subtitle processing method and apparatus. The method includes: during the editing of a multimedia material segment, obtaining subtitle text corresponding to the audio and timestamp information of audio segments corresponding to each text element in the subtitle text through speech recognition; determining the material segment matching the text element in the multimedia material segment based on the timestamp information of the audio segment corresponding to each text element; and then synthesizing each text element with the matching material segment within the specified time to obtain a target multimedia material with a subtitle text appearing word by word in an animation effect. The solution of this disclosure can achieve a subtitle animation effect where the corresponding text subtitle appears when a certain word is spoken; furthermore, user input commands can automatically generate dynamic subtitles, simplifying user operation and improving user experience.

The application provides a voice false wake-up processing method, device and equipment and a storage medium, and relates to the technical field of smart home. The method comprises the following steps: determining the response state of each smart device in a target space receiving a voice signal; determining a pre-wake-up signal of a smart device identified as a pre-wake-up state and a non-wake-up device identified as a non-wake-up state; determining that the voice signal in the target space is a valid wake-up voice based on the pre-wake-up signal and the non-wake-up device; and waking up a target smart device in the target space based on the valid wake-up voice. The application solves the defects of low voice recognition accuracy and high false wake-up voice frequency in the prior art, reduces the dependence on wake-up audio training in a specific space environment, and improves the voice recognition accuracy by determining the valid wake-up voice.

The application relates to the technical field of artificial intelligence recommendation, and specifically discloses a medical cooperation partner intelligent matching method and system based on multi-source data fusion, which collects multi-source business data such as telephone recording, instant messaging text and documents in business communication, constructs a multi-dimensional feature vector of cooperation partners through voice recognition, text analysis and information extraction, carries out vector similarity calculation and basic matching degree comparison on the features of cooperation partners and business project features, forms a candidate set, adopts multi-dimensional weighted scoring and normalization processing to generate a precise recommendation list, and dynamically adjusts the model weight according to the operation feedback of business personnel to realize online optimization. The application can effectively integrate unstructured communication data, improve the mining depth and matching accuracy of cooperation partners, reduce the artificial screening cost and compliance risk, form a self-learning intelligent recommendation closed loop, and significantly improve the efficiency of medical academic popularization and business promotion and the rationality of resource allocation.

The application relates to an embedded fuzzy voice control method and system based on language operator quantification, and relates to the field of embedded voice control, which comprises the following steps: collecting a voice instruction of a user; performing voice recognition on the voice instruction to generate corresponding text instructions; performing semantic analysis on the text instructions to extract control variables, action directions and language operators; determining a basic fuzzy set based on the control variables and the action directions; determining a deformation operator according to the language operator; combining the deformation operator and the basic fuzzy set to generate an input fuzzy quantity, and collecting a current physical state value; and obtaining an aggregated fuzzy set based on the input fuzzy quantity, the current physical state value and a fuzzy rule library; performing defuzzification calculation on the aggregated fuzzy set to determine a physical control signal value, and driving a physical device to perform corresponding actions according to the physical control signal value. The application has the effect of meeting the actual needs of users for the delicacy of device adjustment.

Embodiments of the present application provide a speech recognition method and device, electronic equipment and storage medium, at least applied to the field of artificial intelligence and speech recognition, wherein the method comprises: performing vector coding processing on the audio feature vector of the speech to be recognized to obtain an audio coding vector; performing classification processing on the audio coding vector to obtain a prediction probability distribution of each predicted character in a preset vocabulary corresponding to each speech frame in the speech to be recognized; performing pruning processing on the audio coding vector based on the prediction probability distribution to obtain a pruned audio coding vector; and performing speech recognition on the speech to be recognized based on the pruned audio coding vector to obtain a speech recognition result. Through the present application, the decoding calculation amount in the speech recognition process can be reduced, the decoding efficiency can be improved, and thus the speech recognition efficiency can be improved.

A speech-to-speech translation method comprises transcribing speech spoken in a source language into transcribed text data in the source language using an on-premises speech recognition model. The transcribed text data is translated into translated text data in a target language using a first on-premises machine translation model. The translated text data is reverse translated into retranslated text data in the source language using a second, different on-premises machine translation model. The transcribed text and the retranslated text are displayed on a screen. The method also involves synthesizing, using an on-premises speech synthesis model, translated speech data in the target language based on the translated text data and play back, in response to a user confirmation, translated speech in the target language based on the translated speech data in the target language.

Disclosed are a method and system for providing a speech conference service including an artificial intelligence-based speech interpretation function. According to one embodiment, the method for providing a video conference service may comprise the steps of: setting two or more interpretation bots participating as virtual participants in a video conference service; interpreting a speech of a first language input by participants of the video conference service into a speech of at least one other language different from the first language through processes of speech recognition, translation, and synthetic speech generation between the two or more interpretation bots and artificial intelligence; and providing the interpreted speech.

The present application relates to a sound box volume intelligent adjustment method combining environmental noise monitoring and voice recognition, comprising the following: obtaining a first decibel value and a second decibel value of adjacent sampling intervals without human voice interference respectively; calculating the difference between the second decibel value and the first decibel value to obtain a decibel change value; judging the environmental noise change based on the decibel change value, if the environmental noise change is larger or smaller, triggering the voice recognition function, obtaining the feedback voice of the user and based on this to adaptively adjust the sound box volume. The present application uses the difference between the second decibel value and the first decibel value without human voice interference to approximately estimate the environmental noise change, when the environmental noise changes, the voice recognition function is triggered, the voice interaction with the user is carried out, and the sound box volume is adaptively adjusted, on the one hand, the volume can be adjusted in time when the noise fluctuates to adapt to the user's listening to the sound box content, on the other hand, the interaction with the user increases the user's experience.

Embodiments of the present application disclose a streaming speech recognition method, a method and device for training a speech recognition model. The method comprises: obtaining a speech audio stream; inputting continuous first audio blocks obtained by dividing the speech audio stream by a first time unit into a first speech recognition model to obtain recognition results of the first audio blocks for display; obtaining hidden vectors of each frame obtained by encoding the speech audio stream, and predicting a first sequence corresponding to the speech audio stream by using the hidden vectors, the first sequence comprising weight values of each frame in the speech audio stream; dividing the speech audio stream by using the first sequence to obtain continuous second audio blocks; inputting the continuous second audio blocks into a second speech recognition model to obtain recognition results of the second audio blocks, and updating the displayed recognition results of the corresponding first audio blocks by using the recognition results of the second audio blocks. The present application improves the display effect of real-time speech recognition and enhances user experience.

A system and method for assessing gaze stability and vestibular function using a mobile device are disclosed. The system includes a mobile application configured to execute gaze-stability testing protocols, a display module for presenting visual targets, a head-movement and eye-tracking module utilizing real-time video captured via the device's camera, a speech-recognition module for processing verbal responses, and a data-processing module to analyze head-movement and visual-acuity data. The system enables remote patient assessments and includes protocols such as static visual acuity, visual processing, and mobile gaze stabilization tests to evaluate metrics like peak head velocity and visual acuity. Results can be processed in real-time and can be securely transmitted to clinicians for remote evaluation. The disclosed system provides a cost-effective, user-friendly telehealth solution for vestibular function assessment, eliminating the need for specialized equipment or clinical visits.

This utility model relates to a fixed emergency call device, comprising: a housing with an operating hole on its side; a manual alarm assembly including an operating rope, a trigger block, a trigger switch, and a reset component, wherein one end of the operating rope is connected to the trigger block, and the other end is exposed; the trigger block moves along a preset linear path, and a trigger position is set on the preset path; the trigger switch is located inside the housing and at the trigger position; the reset component keeps the trigger block normally at the starting point of the preset path; a functional module group located inside the housing, including a voice recognition module, a GPS positioning module, an alarm module, and a communication module; wherein the voice recognition module includes an alarm voice receiving module and an alarm voice information processing module; and a main control module electrically connected to the trigger switch, the functional module group, and a monitoring platform server and / or control terminal. This application employs multiple simple and quick alarm methods, making it easy for middle-aged and elderly people to quickly master and avoiding accidental activation.

This invention relates to the field of voice interaction in smart devices, and discloses a cascaded voice interaction method for moxibustion devices. The method collects user voice, preprocesses it to obtain effective voice frames and acoustic feature maps; inputs the feature maps into a first-level voice processing unit, runs a low-power wake-word detection and short command recognition model, and directly outputs control commands when short commands are recognized; if no short command is recognized, it calculates a complex intent tendency score, and if the score exceeds a threshold, it wakes up a second-level voice processing unit; the second-level unit runs a locally deployed lightweight large language model to perform end-to-end voice recognition and semantic understanding, and generates interactive commands based on a moxibustion knowledge base; finally, the commands are executed and the results are broadcast. This invention, through a two-level cascaded architecture, balances low power consumption and high intelligence, achieves offline complex semantic understanding, and improves the voice interaction experience in moxibustion scenarios.

This invention relates to the field of lifting platform technology and discloses a lifting platform with a voice recognition system. The platform includes a vehicle base, an inner scissor-type hydraulic mechanism, a support base fixed to the upper end of the scissor-type hydraulic mechanism, and a protective barrier fixed to the outer side of the support base. A guide groove is formed inside the support base, and a through groove is formed at the lower end of the protective barrier corresponding to the guide groove. An L-shaped mounting frame is slidably connected between the guide groove and the through groove. A telescopic cylinder is installed inside the L-shaped mounting frame, and a mounting support is fixed to the telescopic end of the telescopic cylinder. A support shaft is fixed to the inner side of the mounting support, and an arc-shaped support block is rotatably connected to the outer side of the lower end of the support shaft. A hollow support arm is fixed to the outer side of the arc-shaped support block, and a microphone is installed at the end of the hollow support arm away from the arc-shaped support block. In this invention, the overall structure enables non-contact operation, improving operational convenience and efficiency, and allowing direct control of the lifting and moving of the support base via voice commands.

The application provides an oral medical order intelligent processing method based on an AI chest badge recorder system and related devices, collects audio data of an emergency scene obtained by a medical staff wearing an AI chest badge, and respectively obtains a doctor's dictation voice signal and a nurse's recitation voice signal through noise reduction enhancement and human voice separation processing; the doctor's voice recognition matching and identity authorization determination are completed relying on a pre-stored voiceprint feature library, and after authorization, the two voice signals are respectively transcribed into medical order text and recitation text; consistency verification is carried out on the two texts, and after the verification is passed, structured medical order data is generated and delivered to an execution terminal to assist nurses to standardize medical operations. The application discards the traditional mode of artificial memory, oral check and after-the-fact paper supplement, improves the accuracy of oral medical order checking, the efficiency of circulation and the standardization of execution, ensures the safety of emergency diagnosis and treatment, realizes the traceability of the whole process of medical order, and adapts to the application requirements of emergency and emergency rapid treatment.