33results about How to "Improve intelligibility" patented technology

This invention provides a speech synthesis method, apparatus, electronic device, and storage medium. The method includes: acquiring target text; performing prosodic level prediction on the target text to obtain prosodic boundary information of the target text; determining each prosodic word in the target text based on the prosodic word boundaries in the prosodic boundary information; acquiring the dialectal phonology corresponding to the target text, and determining the dialectal tone sandhi pattern corresponding to each prosodic word based on the dialectal phonology, wherein the dialectal tone sandhi pattern is used to indicate the tone sandhi boundary in each prosodic word; and synthesizing the dialectal speech corresponding to the target text based on the prosodic boundary information and the dialectal tone sandhi pattern. The speech synthesis method, apparatus, electronic device, and storage medium provided by this invention can improve the accuracy and intelligibility of speech synthesis.

This invention discloses an individual motion behavior analysis system based on dynamic modal configuration and rule-guided fusion, belonging to the fields of intelligent perception, sports science, and human-computer interaction. It includes a data input module, a dynamic modal organization and configuration module, a multimodal feature modeling and fusion module, an indicator-level evaluation and behavior state prediction module, a statistical rule-guided analysis and result correction module, and a structured result generation and output module. These modules work together to form a complete closed loop for motion behavior analysis technology. This invention aims to address the problems in existing motion behavior analysis technologies, such as limited dimensionality of single-modal information, rigid multimodal fusion methods, insufficient stability and interpretability of analysis results, and strong dependence on large-scale manually labeled data.

The invention discloses a speech enhancement method based on a time sequence convolutional network, and the method comprises the steps: obtaining a data set, completing the preprocessing of the data, and generating a data pair for training a model; constructing a measurement standard of the system according to general industry knowledge rules and actual application scene requirements; constructing an encoder by using convolution short-time-distance Fourier transform, and using the encoder as a characteristic value of audio data and a model input item; a time sequence convolutional network TCN training model is introduced, and an encoder is decoded to obtain a voice enhancement processing result; according to the method, the effect of speech enhancement is analyzed according to a previously established measurement standard, and large-scale complex audio enhancement processing is quickly and effectively performed by introducing the time sequence convolutional network. The structure, the number of layers and parameters of the time sequence convolutional network can be learned and adjusted according to actual conditions, so that the model has flexibility and adaptability, the performance and the effect of the model are further improved, and stronger support is provided for realizing large-scale multi-type speech enhancement.

The application discloses a natural language and meta-model driven code generation system, which comprises an AI layer containing multiple intelligent agents working cooperatively, the AI layer being configured to receive natural language instructions and perform intent recognition, entity extraction, context analysis and constraint verification through the intelligent agents to generate structured metadata; a meta-model engine configured to construct and instantiate one or more meta-model objects according to the structured metadata generated by the AI layer; and a code template engine configured to match one or more corresponding code templates from a code template library according to the content and type of the meta-model objects, and generate executable code or configuration files in one or more target languages through rendering. According to the scheme, natural language is converted into a meta-model decoupled from technology through AI, and high-quality standardized code is automatically generated through a template engine, thereby improving development efficiency and enabling developers to focus on core business innovation.

This application relates to a method, apparatus, device, medium, and product for generating visualization results of an indicator. The method includes: acquiring a data query file and a visualization conversion file; the data query file includes target library table information for each indicator among at least one indicator, the target library table information indicating a library table used to store indicator data for each indicator; the visualization conversion file includes at least one conversion file; in response to an indicator query instruction carrying an indicator to be queried, determining the target library table information of the indicator to be queried based on the data query file, and determining a target conversion file matching the indicator to be queried from the at least one conversion file; and generating a visualization result of the indicator to be queried based on the target library table information and the target conversion file. This method can improve the efficiency of generating visualization results for the indicator to be queried.

The invention discloses a hearing aid and spoken language partner training integration method and system based on hearing loss and compensation, and belongs to the crossing field of hearing aid and artificial intelligence language learning. The method is cooperatively executed by a Bluetooth headset and a mobile terminal application program, and comprises the following steps: establishing a personalized hearing loss curve of a user; the equipment has an intelligent hearing aid mode and a spoken language partner training mode. And in the hearing aid mode, real-time compensation playing is carried out on the environment sound based on the curve. In the partner training mode, three partner training processes of real-time, preset scene and word generation scene are provided, and adaptation processing is performed on the generated partner training audio based on the same hearing curve, so that key teaching phonemes are ensured to be clear and audible. The core innovation lies in that the system dynamically adjusts the compensation parameter of the hearing loss curve by analyzing the pronunciation accuracy of the user in partner training, and realizes collaborative optimization and personalized improvement of the hearing aid effect and the spoken language ability.

The invention relates to the technical field of software engineering, in particular to a code detection method and device, electronic equipment and a storage medium, and aims to improve code detection efficiency and role perceptivity. Comprising the steps of obtaining a to-be-detected code and structured data of the to-be-detected code; the structured data is used for describing the function of the to-be-detected code; based on the task splitting agent and a plurality of preset detection angles, performing task splitting on the changed sub-codes and the structured data in the to-be-detected codes to obtain a plurality of detection sub-tasks corresponding to different detection angles; for any detection angle, executing a detection sub-task corresponding to any detection angle through a detection intelligent agent corresponding to any detection angle to obtain a detection sub-result corresponding to any detection angle; and inputting the plurality of detection sub-results and the information of the target detection role into the security field large model to obtain a target detection result of the target detection role corresponding to the to-be-detected code.

The application discloses a language decoding method and device, electronic equipment and storage medium. The method comprises the following steps: obtaining an electroencephalogram decoding result of a current time step, wherein the electroencephalogram decoding result comprises electroencephalogram decoding scores corresponding to each preset syllable; expanding a historical candidate sentence of a previous time step based on candidate words associated with the preset syllable to obtain a corresponding expanded sentence set; for each expanded sentence, determining an integrated language decoding score of the expanded sentence based on a historical integrated language decoding score of the previous time step, a first language probability score of the expanded sentence and the corresponding electroencephalogram decoding score; determining a candidate sentence sequence of the current time step based on the integrated language decoding scores of the expanded sentences; performing sentence end detection based on the candidate sentence sequence of the current time step, and determining language content to be output in the current time step according to a detection result of the sentence end detection. The application improves the accuracy, intelligibility and naturalness of interaction of the output language content.

The invention provides a PDMA-based ultra-long distance speech enhancement system, which comprises a hardware acquisition end and a signal processing end, and is characterized in that the hardware acquisition end comprises a parabolic reflector and a differential microphone array, the hardware acquisition end is used for acquiring an original mixed signal, the original signal passes through the parabolic reflector to obtain a pre-enhanced signal, and the pre-enhanced signal is used for processing the pre-enhanced signal; an original signal passes through the differential microphone array to obtain a differential signal, the obtained pre-enhanced signal and the differential signal are input to the signal processing end, and the signal processing end carries out fusion processing on the received pre-enhanced signal and the differential signal to realize secondary enhancement of a target and output a target voice signal. According to the invention, the voice signal can be effectively enhanced in an ultra-long distance scene of more than 20 meters, the voice quality and intelligibility can be obviously improved, the requirements of daily communication transmission are met, the complexity of the system is reduced, the declaration cost is controlled, and the performance of the system is improved.

The application belongs to the technical field of emergency help equipment, and provides an emergency help signal panoramic projection device, method and terminal, the device can continuously detect the spatial posture and rotating motion state of itself through the environment positioning module, generate corresponding posture data and direction data, the control module can obtain the data and determine the real-time projection pointing angle of the device, and then when the device rotates, the projection control instruction corresponding to the real-time projection pointing angle is generated and sent to the image projection module, so that the image generation unit in the image projection module drives the corresponding rotating transformation of the help image, and the light source in the image projection module irradiates the help image to project the visual help signal which is right and complete. The scheme realizes omnidirectional controllable intelligent projection, can make the help signal cover all directions, at the same time ensures that the projected image can be clearly observed, is more easily identified by the rescuer, reduces the delay caused by signal blur, and improves the probability of being rescued by the help seeker.

The application provides a kind of army diabetic patient intelligent remote management method, device, equipment and medium, comprising: obtaining the initial health data of army diabetic patient, the initial health data is desensitized, and the initial health data after desensitization is distributed in cloud security database using blockchain node network Storage;The initial health data after desensitization is input into the diabetes management large model for analysis, and a health management scheme is generated;According to the monitoring index in the health management scheme, the real-time health data of the patient is collected through the wearable device, and the diabetes management large model is input for risk analysis and updating the health management scheme;In response to the natural language interaction request of patient end or medical expert end, human-computer interaction is carried out through medical expert digital avatar, and visual medical advice is output, to solve the problem of intelligent, safe and closed-loop remote management of diabetic patients under the condition of high confidentiality requirement of army and lack of medical resources in border areas.

The invention discloses a front-end audio processing method in the technical field of audio data processing, and the method comprises the steps: carrying out the framing processing of an audio signal collected by a browser, and inputting an audio frame into a multi-stage audio processing pipeline constructed based on WebAssembly; sequentially completing audio preprocessing, deep learning noise reduction processing, echo cancellation processing and post-processing enhancement, performing adaptive adjustment on parameters of each processing stage according to environmental noise changes and processing states, and outputting processed audio signals for real-time communication or audio rendering; by means of the mode, noise and echo interference can be effectively reduced in a pure browser environment, meanwhile, low processing delay and operation stability are considered, and therefore the audio quality of real-time audio and video communication is improved.

The embodiment of the present application relates to the technical field of block chain, and discloses a secret key management method, which comprises the following steps: a request party sends a secret key negotiation request to a service party according to a request party first secret key; the request party first secret key is generated based on a request party smart contract; the service party receives the secret key negotiation request and returns a negotiation result according to a service party first secret key; the service party first secret key is generated based on a service party smart contract; the request party generates a business secret key according to the negotiation result; and the business secret key is used for encryption when the request party and the service party transfer information. Through the above method, the security of the secret key is effectively improved.

The invention relates to a mining area ecological carbon sequestration evaluation method, system and device and a medium. The method comprises the following steps: performing structured processing on original data of a mining area to generate a preprocessed data set; inputting the preprocessed data set into a preset ecological dynamic model, outputting a dynamic model calculation result, and generating a training data set; based on the training data set, training a preset machine learning model to obtain a carbon sequestration prediction model, inputting the to-be-predicted data set into the carbon sequestration prediction model for prediction, and outputting a prediction result and a feature importance weight; and based on natural language query of the user, target data is retrieved from a calculation result and a prediction result of the kinetic model, and an interpretive response is generated in combination with the feature importance weight. According to the method, the precision, the efficiency and the result interpretability of mining area ecological carbon sequestration evaluation are improved by fusing the ecological dynamics model and the machine learning model, meanwhile, the user interaction capability is enhanced, and specific query requirements can be quickly responded.

The invention relates to the technical field of oil and gas drilling, and discloses a multi-track comparison visualization method and system for real-time guiding of a well track, and the method comprises the steps: constructing a well track real-time guiding multi-track comparison visualization architecture which is integrated in a drilling control system and is fused with multi-source data to complete real-time track calculation and prospective prediction; a prediction track is introduced and compared with a design track and a real-time track on the same screen, so that an engineer can more intuitively pre-judge a future trend and adjust a guiding strategy in advance, synchronous presentation of a dynamic drilling tool and a tool face orientation is ensured, the understandability of a guiding key state is improved, and single dependence on experience is reduced; according to the method, a visual guiding solution capable of performing same-screen comparative analysis on a design track, a real-time track and a predicted track is provided, so that the intuition, continuity and real-time performance of guiding judgment are improved, and the problems that in the prior art, visualization and a control system are disjointed, the prediction capability is lacked, seamless integration is difficult, and guiding decision lag is caused are solved.

The invention discloses a clinical outcome machine learning prediction method and device based on multiple models and a storage medium, and the method comprises the steps: constructing a plurality of machine learning prediction models under a unified calculation framework, and combining with automatic parameter optimization, multi-dimensional performance evaluation, prediction probability calibration and model interpretation analysis. Systematic modeling and evaluation of a clinical outcome prediction task are realized, instability caused by dependence of a single model and manual parameter adjustment is effectively avoided, and the stability, generalization ability and repeatability of a prediction result are improved. Meanwhile, the reliability and practicability of the prediction result in risk assessment and clinical decision support are improved by performing calibration processing on the prediction probability and introducing clinical decision curve and clinical influence curve analysis. Through model explanation and visual output, the transparency and understandability of the prediction process are enhanced, so that the acceptability and application value of the model in medical research and clinical application are improved.

The invention discloses a data lake treatment method and device, electronic equipment and a program product, and relates to the technical field of big data. The method comprises the steps of obtaining original data from a heterogeneous data source, constructing a formal context and generating a concept lattice; semantic embedding and similarity calculation are carried out on the field names, clustering is carried out in combination with the concept lattice, unified field names are generated, and a unified data view is constructed; performing semantic enhanced metadata management on the unified data view, including constructing an attribute graph, mapping into a knowledge graph and expanding an implicit relationship thereof, and constructing an index to support semantic query; and dynamic access control and metadata integrity management are carried out on the unified data view, so that the safety and reliability of the data are ensured. According to the method, the automation degree and accuracy of data integration are effectively improved, and the discoverability, interoperability and safety of the data are enhanced.

The invention relates to an aviation search and rescue system load cooperative intelligent management and control system and method, and belongs to the technical field of aviation detection. The search load is used for acquiring target information of a search area and providing real-time position and attitude information; the interface equipment is used for matching the data of each search load and a control interface to realize receiving, conversion and sending of the data; and the display control equipment is used for receiving the search data of each search load and generating target identification information, realizing real-time management and control on each load, and issuing a control command according to a search and rescue task process. When a search rescue task is executed, depending on the established intelligent decision model library, the working modes of the search loads can be intelligently controlled in an automatic / manual control mode, target information is obtained, cross control over the different search loads is achieved according to the target information, and cooperative intelligent control over the loads is achieved when search rescue is carried out in a general scene.

The invention discloses a large language model-oriented super-long text three-dimensional compression method, which comprises the following steps of: receiving super-long text data, and carrying out word segmentation to obtain an initial text word sequence; performing three-dimensional feature extraction on the full text based on the sequence to obtain word frequency, statement and context dimension features; respectively combining the three-dimensional features to obtain corresponding word frequency, statement and context compression results; and integrating the three compression results, performing semantic integrity check on the obtained preliminary compressed text, and complementing missing information from the original text according to the check result to obtain a final compressed text. According to the method, through three-dimensional feature collaborative extraction and integrity constraint, the information fidelity and logic self-consistency of the compressed text are improved, and the problems of serious information loss and low availability caused by insufficient semantic understanding and logic fracture in an existing method are solved.

The invention is suitable for the technical field of data management, and provides a research and development full-process data management method and system for a high-Tg cover film, and the method comprises the following steps: generating a unique experiment batch ID for each film preparation experiment, and collecting the structural data of the experiment batch; taking the structured data as an experimental node, and storing the structured data in a graph database; establishing a change relation edge between the current experimental node and a historical experimental node serving as a father node of the current experimental node, wherein the change relation edge comprises change description of formula parameters and / or process parameters; on the basis of the multiple experimental nodes and the change relation edges accumulated in the graph database, constructing a visual experimental relation graph; and when an analysis instruction for a target experiment node is received, generating a correlation analysis result based on the experiment relation graph. According to the method, the formula, the process, the result and the iterative decision logic are completely associated together and are visualized in a graph form, so that the understandability and the traceability of data are enhanced.

The invention provides a hydropower engineering design data modeling system based on a BPRS method, which comprises a process object enumeration library, a design semantic library, a geometric feature library, a design parameter library, a business process library, a visual modeling module, a data modeling engine, a database structure generation module, a data exchange standard generation module and a collaborative interface module, unification of design data, business processes and system structures is realized, and multi-specialty collaboration, parameter association, semantic modeling and automatic generation of data standards are supported. According to the method, efficient collaboration, parameterized management, semantic modeling and standardized generation of hydroelectric engineering design data can be realized, the problems of data islands, flow breakage, semantic deficiency and the like in traditional design are solved, and the design efficiency and the intelligent level are improved.

This invention relates to the field of meteorological forecasting technology, specifically a meteorological forecasting adaptive optimization method and system based on multi-algorithm fusion. This method constructs a parameter optimization framework based on multi-algorithm fusion, embedding physical constraints, and possessing adaptive capabilities, aiming to systematically solve key problems in the parameter tuning process of traditional meteorological forecasting systems. This framework achieves efficient and intelligent optimization through a series of tightly linked steps, ensuring the effectiveness and stability of the method in operational scenarios. This process integrates preceding modules into a unified platform through a microservice architecture, automating data and control flows. The evaluation system includes historical backtesting, real-time forecast testing, and extreme weather-specific testing, quantifying forecast accuracy improvement, resource consumption optimization, and cross-scenario generalization capabilities. Results are fed back to a knowledge base to form a self-optimization loop, ultimately verifying the comprehensive advantages of this method in improving forecast efficiency, ensuring physical rationality, and operational applicability.

The invention relates to a voice signal processing method and device, electronic equipment and a storage medium. The method comprises the steps that content coding features, rhythm coding features and timbre coding features are acquired, the content coding features are used for representing the content of a voice signal, the rhythm coding features are used for representing the rhythm of the voice signal, and the timbre coding features are used for representing the timbre of the voice signal; content context construction is carried out on the content coding features based on the tone coding features and the rhythm coding features, and content context features are obtained; and performing voice reconstruction according to the timbre coding feature, the rhythm coding feature and the content context feature to obtain a target reconstructed voice signal. According to the technical scheme provided by the invention, high-quality speech reconstruction can be provided at a low code rate.

The application discloses an audio transmission method, an audio transmission device and a storage medium, and belongs to the technical field of audio transmission. The audio transmission method comprises the following steps: controlling a sending end to collect audio data and encoding the audio data to form an encoded frame; determining parameters of a network, and controlling the sending end to switch to a corresponding working mode according to the parameters of the network, wherein the working mode comprises a first mode and a second mode, the first mode is different from the second mode, and the parameters of the network comprise a network speed change rate and a bandwidth; packing the encoded frame into a data packet according to the current working mode of the sending end, and controlling the sending end to send the data packet to a receiving end; and controlling the receiving end to switch to a target mode, so that the receiving end receives the data packet, decodes the data packet according to the target mode in which the receiving end is located, obtains decoded data, and sends the decoded data outward, wherein the target mode is a mode that is the same as the working mode in which the sending end is currently located.

This application provides an anomaly risk identification method, apparatus, electronic device, medium, and program product, relating to the application of large-scale models in fintech scenarios, and applicable to the fields of big data technology, artificial intelligence technology, and fintech. The method includes: constructing a user profile based on historical transaction data of a target account; constructing an account transaction graph based on real-time transaction data of the target account; identifying a target transaction path corresponding to the target transaction based on the account transaction graph; and identifying anomaly risks of the target transaction based on the user profile and the target transaction path.

The invention provides a VHF and IP voice converged communication method and device for a special ship. The method comprises the following steps: monitoring environment temperature and electromagnetic interference intensity in real time; a pre-stored antenna parameter compensation library is matched based on the environment temperature, and radio frequency front end transmission parameters are dynamically optimized; and performing multi-modal noise suppression on the received VHF audio signal, wherein the multi-modal noise suppression comprises the following steps of: generating a time-frequency mask based on electromagnetic interference intensity by adopting a complex convolution attention network to suppress electromagnetic interference; and converting the processed signal into an IP voice data packet, and reversely converting the IP voice data packet into a VHF signal through a protocol conversion engine to realize a bidirectional communication link. The voice communication quality and reliability in the polar region environment are improved, and the navigation safety is guaranteed.

The utility model discloses an explanation emitter and a wireless explanation device. The explanation emitter comprises a housing, a control mainboard, a moving-coil microphone, a microphone protection net, a transceiver, a transmit-receive antenna and a battery. The shell is provided with a head part and a tail part which are away from each other; an accommodating cavity is formed in the shell; at least part of the surface of the shell forms a handle part; the control mainboard is arranged in the accommodating cavity; the moving-coil microphone is arranged at the head part of the shell and is electrically connected with the control mainboard; the microphone protective net cover is arranged at the periphery of the moving-coil microphone and is connected with the shell; the transceiver is arranged in the accommodating cavity and is electrically connected with the control mainboard; the transmitting-receiving antenna is arranged in the accommodating cavity and is far away from the area where the handle part is located, and the transmitting-receiving antenna is further electrically connected with the control mainboard; the battery is arranged in the accommodating cavity and is electrically connected with the control mainboard; by adopting the moving-coil microphone, the sound of an explainer can be picked up at a short distance, a large amount of picked environmental noise is reduced, and the voice definition and intelligibility are improved.

The application discloses a kind of fusion pushing mode and gesture recognition's underground unmanned vehicle control method, system and unmanned vehicle, comprising: with the tracking result of unmanned vehicle to guide personnel and gesture recognition result as input, the cycle control decision of unmanned vehicle working mode is carried out in combination with environmental perception information, comprising: in response to single control period, guide personnel station position and instruction intention are identified based on tracking result, gesture recognition result and environmental perception information;According to guide personnel station position and instruction intention, the working mode of unmanned vehicle is decided;In response to the working mode decided, the environment passability check result is calculated and the chassis control instruction of unmanned vehicle is output, and the working mode, direction intention, executability state and safety keeping instruction are synchronously controlled unmanned vehicle and visualized display is shown outside;The application can be based on guide personnel continuous tracking combined with the core control scheme of gesture recognition, realize the autonomous switching of unmanned vehicle between following, pushing, parking and other working modes.

The present application relates to the field of cash logistics and vault operation intelligence, and particularly relates to a regional cash service intelligent decision system and method, wherein the system comprises: an edge risk perception and prevention layer, which is used for identifying abnormal behaviors and generating risk event data; a data perception and fusion layer, which is used for collecting and fusing multi-source data, and constructing system state time sequence features; a causal inference layer, which is used for constructing a causal graph model and performing intervention effect estimation; a multi-agent reinforcement learning decision layer, which is used for generating decision actions according to the risk event data, the system state time sequence features and the intervention effect estimation results; a business rule constraint and explanation module, which is used for performing compliance verification on the decision actions and generating decision explanation information; and an autonomous evolution layer, which is used for generating cross-supervision signals according to mutual evaluation between the agents, and driving evolution update of the agent strategies. The present application realizes global collaboration, real-time risk control, interpretable decision and autonomous evolution of regional cash service.

The invention discloses an audio signal adjusting method and device, virtual reality equipment and a storage medium. The method comprises the following steps: determining a binaural hearing threshold value of a current user; determining a first target audio signal and a second target audio signal according to the binaural hearing threshold; and adjusting a left sound production unit of the virtual reality equipment according to the first target audio signal, and adjusting a right sound production unit of the virtual reality equipment according to the second target audio signal. According to the invention, the volume of the left and right sound channels of the virtual reality equipment can be intelligently and independently adjusted according to the hearing difference between the two ears of the user, so that the effective sound pressure levels received by the two ears are balanced, and the immersion and comfort of VR experience are remarkably improved.