Patents
Literature
Patsnap Eureka AI that helps you search prior art, draft patents, and assess FTO risks, powered by patent and scientific literature data.

50results about How to "Improve generation accuracy" patented technology

Text-to-structured query language conversion method based on large language model

A text-to-structured query language conversion method based on a large language model comprises the steps that database outline information is cut, and tables and columns needed by query of a non-target structured query language SQL are eliminated through correlation scoring and Steiner tree problem solving; screening context samples, predicting a similarity score from a text to a structured query language Text-to-SQL by using a cross encoder language model, and selecting a sample most related to a query problem; and injecting the cut database table and column information and the screened context sample into a cue word template, and generating an SQL query statement by using a large language model. The method can be used for optimizing the automatic conversion process from the natural language to the SQL and improving the conversion accuracy, and is used for providing a quick and convenient database retrieval mechanism for a non-SQL user and providing applications such as a data access mode in a wider form for mechanisms such as big language model retrieval enhancement generation RAG and the like.
Owner:TSINGHUA SHENZHEN INTERNATIONAL GRADUATE SCHOOL

Advertisement marketing method and system based on data analysis

The invention discloses an advertisement marketing method and system based on data analysis, relates to the technical field of advertisement marketing, and realizes a complete closed loop from user data acquisition and analysis to advertisement accurate matching and delivery by optimizing the whole process of advertisement delivery. Through scientific division of software categories and user browsing formats and collection of content information and behavior data based on user authorization, the generation precision of user interest labels is remarkably improved, and a solid foundation is laid for personalization and precision of advertisement putting; according to the mode, delivery of invalid advertisements is effectively reduced, the integrating degree of advertisement content and user requirements is improved, the advertisement receiving experience of the user is optimized, meanwhile, the natural language processing technology and the algorithm model are combined, matching of interest labels and advertisement information is converted into screening of optimal advertisements, and maximization of the advertisement delivery efficiency is achieved.
Owner:TAIDOU TECH GRP CO LTD

Depressive disorder detection report generation system based on retrieval enhancement generation technology and large model fine tuning

The invention discloses a depressive disorder detection report generation system based on a retrieval enhancement generation technology and large model fine tuning, which is applied to the technical field of depressive disorder detection, and comprises a data generation module used for rewriting social media data into psychological counseling scene dialogue data by utilizing prompt guidance GPT-4 based on a psychological counseling dialogue prompt project; the AI psychological counselor construction module is used for finely adjusting the large language model and constructing an AI psychological counselor based on the psychological counseling scene dialogue data and the clinical interview data; and the AI psychologist construction module is used for performing similarity matching on dialogue contents of the user and an AI psychologist and a DSM-V judgment standard by utilizing a retrieval enhancement generation technology, guiding GPT-4 by utilizing a prompt, generating a depressive disorder detection report according to a detection report prompt project, performing fine adjustment based on a large language model, and constructing an AI psychologist. The depressive disorder detection and detection report generation precision is effectively improved.
Owner:HEFEI UNIV OF TECH

Mobile game dynamic control optimization method based on adaptive deep learning

The invention discloses a mobile game dynamic control optimization method based on adaptive deep learning, and the method comprises the following steps: S1, collecting player operation data and scene background parameters, and generating feature data of a unified scale; s2, generating a self-adaptive spatial-temporal feature sequence through the spatial-temporal feature self-adaptive convolutional network; s3, constructing an adaptive memory network, recognizing a recent high-frequency operation mode through a short-term focusing mechanism, and generating operation intention features; s4, generating a self-supervised learning task through a self-supervised learning multi-layer perceptron, and adaptively adjusting the feature weight to generate a control parameter; s5, a feedback score is generated in real time through a feedback scoring mechanism, and the learning rate is adjusted; s6, establishing a satisfaction evaluation mechanism based on feedback scores and data analysis; and S7, recording operation feedback and scoring results in real time, and realizing adaptive optimization based on feedback circulation. According to the method, mobile game control is dynamically optimized through the adaptive deep learning method, and high adaptability and excellent user experience are achieved.
Owner:BEIJING AIPULEI TECHNOLOGY CO LTD

Method and equipment for improving Text-to-SQL (Structured Query Language) conversion accuracy, medium and program product

PendingCN119621746AThe effect of weakening the hallucinationsavoid missingRelational databasesEnergy efficient computingTable (database)Engineering
The invention provides a method and equipment for improving Text-to-SQL (Structured Query Language) conversion accuracy, a medium and a program product. The method comprises the following steps: connecting a target database to obtain DDL (Double Data Language) information of all tables in the target database; extracting effective information of all tables in the database based on the acquired data to establish a table association relation graph, and performing near-synonym expansion on DDL information; when a user inputs a natural language question, an entity contained in the question is extracted and combined with the question to perform RAG retrieval, reachability analysis is performed based on a table association relation graph, and DDL information is used as a context to be transmitted to a large model after a correlation table is complemented, so that the large model generates an SQL (Structured Query Language); performing correctness verification on the generated SQL by using the table association relation graph; if the verification is passed, the generated SQL is returned to the user, and if the verification is not passed, automatic correction is tried, and the corrected SQL is returned to the user. According to the method and the device, the SQL generation accuracy can be improved.
Owner:NO 30 INST OF CHINA ELECTRONIC TECH GRP CORP

Intelligent regulation and control task chain generation method and device, electronic equipment and medium

The invention discloses an intelligent regulation and control task chain generation method and apparatus, an electronic device and a medium. The method comprises the steps of obtaining content texts of multiple types of regulation and control security regulations and regulation and control operations; inputting the content text into a large language model agent for content extraction so as to extract regulation and control operation sequences and calling contents under various conditions; intelligently generating a knowledge graph in a decision tree form in the knowledge graph through an intelligent agent according to the regulation and control operation sequence and the calling content under various conditions; acquiring a text instruction, performing intention recognition by an intelligent agent, intelligently generating a task chain for regulating and controlling operation logic in the generated knowledge graph after autonomous understanding, and autonomously generating a thinking chain for task execution; and according to the generated thinking chain, the intelligent agent intelligently calls related modules for regulation and control so as to realize intelligent regulation and control. According to the method, the quality and efficiency of obtaining the text information can be improved, the regulation and control thinking chain logic architecture is established by using the decision tree type knowledge graph, the accuracy is improved, and the use threshold of operation is reduced.
Owner:SOUTH CHINA UNIV OF TECH

Complex shape three-dimensional building model generation method based on frame field deep learning

The invention relates to a method for generating a complex-shape three-dimensional building model based on frame field deep learning. The problems that in the prior art, when a three-dimensional building model is generated, training difficulty is large, the three-dimensional building model is easily affected by smooth corners, and artifacts are generated are solved. The method comprises the following steps: S1, acquiring remote sensing data and point cloud data; s2, preprocessing the acquired image data; s3, carrying out data normalization processing; s4, designing a frame field and optimizing details; s5, optimizing the activity skeleton model; and S6, generating a multi-shape contour and a three-dimensional model. The method has the advantages that the generation precision and the generation efficiency of the three-dimensional building model are improved, buildings with complex shapes can be processed, and the method has a wide application prospect.
Owner:AIKUN (SHANGHAI) ARCHITECTURAL DESIGN CO LTD

Table-to-text conversion method, system and device and medium

The invention provides a table-to-text conversion method, system and device and a medium, and the conversion method comprises the steps: obtaining a first dictionary recording key query information, a to-be-converted data table and a first prompt word corresponding to the key query information; the first prompt word is used for indicating table type data corresponding to the key query information in the retrieval data table; inputting the data table and the first prompt into a language model, and obtaining target table type data through the language model; according to the target table type data, performing storage updating on the first dictionary to obtain a second dictionary; and inputting the second dictionary and a second prompt into the language model to obtain a target generated text, the second prompt being used for indicating to retrieve dictionary structure data corresponding to the key query information in the second dictionary. According to the conversion method, computing resources and reasoning time consumed by converting the table into the text can be reduced, and the generation accuracy of converting the table into the text is improved. The invention relates to the technical field of artificial intelligence.
Owner:CHINA TELECOM ARTIFICIAL INTELLIGENCE TECHNOLOGY (BEIJING) CO LTD

Data comparison reconciliation method and system in cross-network data exchange transmission scene

The invention provides a data comparison and reconciliation method and system in a cross-network data exchange and transmission scene, and relates to the technical field of big data. A prepositive data reconciliation system receives service data, generates verification information containing a prepositive verification code, a service data format, a data volume and a unique identification code, and sends the verification information to a server; and after secondary packaging, the data is sent to a rear data reconciliation system through a one-way transmission channel. And the rear system de-encapsulates and generates corresponding verification information, compares the verification information with the received information, and marks a verification state according to a result. If verification fails, an abnormal classification and traceability mechanism is triggered, automatic retransmission or manual intervention is executed, the verification code generation model based on the multi-modal confrontation Hash network can extract multi-modal features, a Hash path is dynamically selected according to information entropy to generate a verification code, the verification accuracy and rate of the multi-modal data are improved, and the verification efficiency is improved. And the abnormal classification covers content tampering, protocol abnormity and the like, and transaction records can be searched in the alliance chain according to the transmission sequence number during tracing, so that the problem positioning and solving capabilities are improved.
Owner:FUZHOU PUBLIC SECURITY BUREAU +1

Satellite neighbor cell determination method and device, equipment, storage medium and program product

The invention relates to a satellite neighbor cell determination method and device, equipment, a storage medium and a program product. The method comprises: receiving interface configuration information sent by an adjacent base station of a target satellite base station, the adjacent base station being a base station within a preset range of the target satellite base station, the interface configuration information comprising non-ground network NTN auxiliary information of a cell of the adjacent base station; determining the distance between the cell of the target satellite base station and the cell of the adjacent base station according to the NTN auxiliary information of the cell of the adjacent base station and the NTN auxiliary information of the cell of the target satellite base station; and according to the distance between the cell of the target satellite base station and the cell of the adjacent base station, adjacent cell relation information of the target satellite base station is determined, and the adjacent cell relation information is used for indicating whether the cell of the adjacent base station is the adjacent cell of the cell of the target satellite base station. By adopting the method, the accuracy of the generated neighbor relation can be improved.
Owner:SHANGHAI SPACECOM SATELLITE TECH LTD

Instruction generation method and device, vehicle-mounted terminal and vehicle

The invention is suitable for the technical field of automobiles, and provides an instruction generation method and device, a vehicle-mounted terminal and a vehicle, and the method comprises the steps: responding to a motorcade management instruction, and determining a management mode corresponding to a motorcade; obtaining driving information of each vehicle in the motorcade; based on the driving information and the management mode of each vehicle, determining a management mode corresponding to each vehicle; and generating a management instruction corresponding to each vehicle based on the management mode of each vehicle, and sending each management instruction to the corresponding vehicle. Compared with the prior art, the method can accurately determine the management mode of each vehicle in combination with the driving information of each vehicle and the management mode of the motorcade, and further accurately generate the management instruction matched with each vehicle, thereby improving the generation accuracy of the instruction, enabling the management of each vehicle to be better matched with the management mode of the motorcade, and improving the user experience. And thus, the motorcade management accuracy is improved.
Owner:GREAT WALL MOTOR CO LTD

Text generation method and device, equipment, storage medium and program product

The invention provides a text generation method and device, equipment, a storage medium and a program product. The embodiment of the invention can be applied to text generation scenes such as voice recognition, voice text conversion and the like; the method comprises the steps of performing framing processing on a to-be-processed voice signal to obtain a plurality of voice frames of the to-be-processed voice signal; performing feature extraction on each voice frame to obtain a frame feature corresponding to each voice frame; fusing the plurality of frame features to obtain a voice feature of the to-be-processed voice signal; and based on the voice features, performing text conversion on the to-be-processed voice signal to obtain a matching text corresponding to the to-be-processed voice signal. According to the invention, the generation accuracy of the text content corresponding to the voice signal can be improved.
Owner:TENCENT TECHNOLOGY (SHENZHEN) CO LTD

Automatic generation method of test case and related device

The invention discloses an automatic generation method of a test case and a related device, and relates to the field of software test.The automatic generation method comprises the steps that feature vector extraction is conducted on a target uncovered code and a full-quantity code group through a preset feature extraction model, and a calling link feature vector and a calling feature vector of the target uncovered code are obtained; performing feature fusion on the calling link feature vector and the calling feature vector to obtain a calling information feature vector of the target uncovered code, inputting the calling information feature vector into a preset encoder so as to enable the preset encoder to output a weight code set, inputting the weight code set and a test case template into a preset decoder, and outputting the target uncovered code. And obtaining the test case of the target uncovered code. According to the method and the device, the test case of the target uncovered code is generated by utilizing the preset decoder based on the feature weight code carrying the weight representing the association degree of the calling information and the calling target uncovered code and the test case template, so that automatic generation of the accurate test case is realized.
Owner:AGRICULTURAL BANK OF CHINA

Generation method of three-period minimal curved surface and computer terminal

The invention provides a three-period minimal curved surface generation method and a computer terminal. The generation method comprises the following steps: S1, manufacturing a solid STL model with a required shape; s2, extracting contour coordinate data of the solid STL model; s3, determining a spatial range of the solid STL model; s4, performing grid division on the spatial range, and obtaining coordinate data of each node in a grid; s5, storing each node in the solid STL model as a feature point; s6, calculating an objective function value corresponding to each feature point; and S7, an isosurface is extracted through an isosurface function, and a curved surface is drawn and repaired. According to the method for generating the three-period minimal curved surface, voxel division can be carried out on the solid STL model with the needed shape, gradient change of porosity is achieved through the contour surface extraction algorithm, and the generation precision of the three-period minimal curved surface structure model is improved.
Owner:CHENGDE MEDICAL UNIV

A method for generating bird's-eye view images of reservoir landscapes based on a stable diffusion large model

This invention relates to the field of image generation technology, and more particularly to a method for generating bird's-eye view images of reservoir landscapes based on a Stable Diffusion large model. The method includes the following steps: preprocessing acquired reservoir landscape images and semantically annotating them based on water conservancy terminology to construct an image-text pairing dataset; training a pre-trained Stable Diffusion model, incorporating LoRA adaptation parameters into its UNet attention layer, using the image-text pairing dataset to generate a target feature image; and updating the LoRA adaptation parameters based on the difference between the target feature image and the reservoir landscape image to obtain a LoRA model for the reservoir landscape bird's-eye view image. This invention achieves structurally controllable and semantically consistent reservoir landscape generation by fusing structural reference constraints and semantically weighted alignment and introducing LoRA modulation.
Owner:GUANGDONG ZHURONG ENG DESIGN CO LTD

System and method for intelligent operations and visitor experience optimization for museums and art exhibitions

ActiveCN121032134BImprove matching accuracyAccurately reflect differences in resource requirementsInstrumentsEngineeringIntelligent management
The application discloses a museum and art exhibition intelligent operation and visitor experience optimization system and method, relates to the technical field of intelligent management of museums and art exhibitions, and comprises the following steps: obtaining interactive behavior data and identity data of each target object; performing group projection on the interactive behavior data and the identity data, generating group structure data, and performing hierarchical attribution division on the group structure data from an identity dimension, to generate exhibition participation data of each group; and performing scheduling analysis on the correlation degree of the target objects contained in the corresponding group according to the exhibition participation data, to generate behavior-identity deviation data. Through the hierarchical allocation and dynamic reallocation mechanism, the scheme can accurately reflect the resource demand differences of different group levels, and solves the problem of insufficient display strategy adaptability caused by ignoring the behavior differences between group levels in the prior art.
Owner:XIAMEN OCEAN VOCATIONAL & TECH COLLEGE

Defect type detection method and device, electronic equipment and storage medium

The invention provides a defect type detection method and device, electronic equipment and a storage medium. The detection method comprises the following steps: determining a defect area in a to-be-detected image of a to-be-detected object; extracting image features corresponding to a defect area in the to-be-detected image; based on image features corresponding to a defect area in the to-be-detected image, utilizing a trained aggregation projection module to generate a first aggregation feature; and determining the actual defect type of the to-be-detected object based on the first aggregation feature. The trained aggregation projection module is obtained by training based on the first global feature and the first defect feature of the training image, so that the first aggregation feature output by the trained aggregation projection module also pays attention to the global feature of the to-be-detected image to a certain extent, and the generation precision of the actual defect type can be improved.
Owner:SUZHOU MEGAROBO TECH CO LTD

Process page display method and device, electronic equipment and computer readable medium

Embodiments of the present disclosure disclose a process page display method and device, electronic equipment and a computer readable medium. A specific embodiment of the method comprises: in response to receiving a process page rendering request, obtaining code information and a set of business processing criteria information; generating initial full business process information; verifying the execution relationship between the initial full business process information, obtaining full business process information; rendering the full business process information and rendering and displaying it on the process page; in response to selecting selection information, generating a matching business batch as a first business batch; and displaying the corresponding business process processing tree on the process page. This embodiment can accurately render the corresponding business process information on the process page after receiving the business data related to the business process generation uploaded to the process page, and can set various page filtering operations for the business process information, greatly improving the rendering efficiency of the business process.
Owner:中信证券股份有限公司

Virtual building model generation method and device, equipment, medium and program

The invention provides a virtual building model generation method and device, equipment, a medium and a program, and relates to the technical field of games. The method comprises the following steps: acquiring at least one group of building structure parameters, wherein the building structure parameters at least comprise a space contour parameter, a structure control parameter and a placement parameter of a virtual connected entity model; generating a space geometry by adopting the space contour parameters; performing structure division on the space geometry by adopting the structure control parameters to obtain a plurality of space regions; generating a virtual segmentation entity model at a target position of a first space region in the plurality of space regions to obtain a basic model of each space region; according to the placement parameters of the virtual connected entity model, the virtual connected entity model is loaded to the basic model of the second space area in the multiple space areas, and a virtual building model corresponding to the building structure parameters is obtained. Therefore, the building model generation precision is improved, and the building model generation scene is expanded.
Owner:NETEASE (HANGZHOU) NETWORK CO LTD

Video retrieval method, device, equipment and storage medium

The present application relates to the field of computers, and in particular to the field of artificial intelligence, and provides a method, apparatus, device and storage medium for video retrieval to solve the problems of low quantization efficiency and low accuracy. The method includes: extracting features of the image features of the video to be retrieved to obtain a first quantized feature, and then based on the first quantized feature, obtaining a second candidate video with a high category similarity to the video to be retrieved, and finally using the second candidate video with a high content similarity to the video to be retrieved as the target video. The quantization control parameters are adjusted according to the texture feature loss value corresponding to each training sample, so that the target quantization processing sub-model learns the sorting ability of the target texture feature sub-model, ensuring that the sorting effects of the two sub-models tend to be consistent, and the end-to-end model architecture enables the target quantization processing sub-model to obtain the corresponding quantized features based on the image features, thereby improving the generation accuracy of the quantized features and the recall performance of the video retrieval.
Owner:TENCENT TECHNOLOGY (SHENZHEN) CO LTD

Method and device for generating driving path

The invention discloses a driving path generation method and device, and relates to the field of automatic driving and data processing. The method comprises the following steps: determining starting point pose information of a vehicle at a driving starting point and terminal point pose information of the vehicle at a driving terminal point; determining a target rotation angle from the driving starting point to the driving ending point based on the starting point pose information and the ending point pose information; based on the target rotation angle, at least one spiral line path is generated, and curvatures of different positions on any spiral line path are different; and generating a target driving path of the vehicle based on the at least one spiral line path. According to the method and the device, the technical problem that the efficiency of generating the path with the curvature and the curvature change rate meeting the requirements in the related technology is low is solved.
Owner:EACON TECHNOLOGY CO LTD

Energy storage energy management method, device and system and medium

The invention discloses an energy storage energy management method, device and system and a medium, the energy storage energy management system comprises a local energy management system, a local energy management system and a local energy system, the local energy management system obtains operation data of each device in the local energy system, the model prediction precision of a locally deployed energy prediction model is improved; based on the data type of the operation fusion data, performing data fusion processing on all the operation data to obtain the operation fusion data; calling an energy prediction model to perform energy prediction on the operation fusion data to obtain a strategy prediction result; when the model prediction precision reaches the target precision, obtaining a corresponding local pre-stored strategy according to the operation fusion data, and adjusting the local pre-stored strategy according to a strategy prediction result to obtain a target control strategy; according to the invention, the accuracy of energy management and control can be improved, so that the reliability of an energy storage energy management system is improved.
Owner:SHENZHEN CLOU ELECTRONICS

Method and system for generating candidate sets of video action segments based on common sense enhancement

The present invention belongs to the technical field of video behavior segment retrieval. A method and system for generating a candidate set of video behavior segments based on common sense enhancement is provided, which obtains the semantic feature representation of the query statement and the video to be retrieved; interacts with the semantic representation of the visual feature and the semantic representation of the text feature through a cross-modal interaction module to fuse multimodal information; predicts the probability of each video unit being retained, and retains the video units with high information validity based on the probability; maps the features of the retained video frames to the input space of the large image-text pre-training model through the visual adaptation layer; fine-tunes the large image-text pre-training model by inserting the adaptation layer, and constructs an instruction to instruct the model to complete the task of generating a candidate set of video behavior segments. The present invention introduces a large image-text pre-training model to utilize the rich external knowledge therein to improve the understanding of the visual content, while taking into account the generation speed and accuracy of the candidate set of video behavior segments.
Owner:SHANDONG UNIV +3

A signal processing method for Doppler frequency offset interference of height-finding radar

The present application relates to radar detection technology field, especially to a kind of signal processing method of Doppler frequency offset interference of height-finding radar, the method is by monitoring the maneuverability of height-finding radar and dynamically adjusting frequency modulation period and frequency modulation bandwidth of symmetrical frequency modulation signal;Based on the frequency modulation period and frequency modulation bandwidth after adjustment generate positive slope frequency modulation signal and negative slope frequency modulation signal and carry out periodic alternate emission, collect and process and difference calculation first echo signal and second echo signal after reflection, obtain after eliminating Doppler frequency offset interference ranging frequency difference, further avoid the height-finding error caused by Doppler shift amount, to obtain accurate target height based on ranging frequency difference calculation.This application is by using the Doppler shift polarity opposite characteristics of positive and negative slope frequency modulation signal, by frequency spectrum line pairing and difference calculation offset Doppler shift component, finally based on the target height of corrected frequency information inversion, to realize high-precision height-finding.
Owner:SHENZHEN ZKX SCI & TECH DEV CO LTD

A sentence generation method and related device

ActiveCN116306612BImprove generation accuracy
The application relates to the field of artificial intelligence and discloses a word and sentence generation method, which comprises the following steps: obtaining a target character string sequence, the target character string being input by a user in an input method tool; generating a target word and sentence corresponding to the target character string sequence through a target neural network according to the target character string sequence, wherein the target neural network comprises an encoder and a decoder, the encoder is used for obtaining an embedding vector according to the target character string sequence, the decoder is used for generating the target word and sentence according to the embedding vector, the target neural network is obtained by training a training sample, and the training sample comprises a character string sequence and a corresponding word and sentence; and presenting the target word and sentence in an interface of the input method tool. The application inputs a character string without error correction and word segmentation into the target neural network, reduces the error superposition influence of a word and sentence precision caused by an error correction model and a word segmentation model, and improves the generation precision of the word and sentence.
Owner:HUAWEI TECH CO LTD

Power audit SQL statement generation method based on large model

The invention discloses an electric power auditing SQL statement generation method based on a large model, relates to the technical field of information, and solves the technical problem that in the prior art, text data is directly generated into SQL statements, so that some text data with irregular formats are generated into wrong SQL statements, and the efficiency of electric power auditing work is low. The method comprises the following steps: identifying and verifying input data; calculating a data normativity coefficient according to the input data; when the data normativity coefficient is greater than a normativity coefficient threshold value, inputting the input data into an SQL generation model to obtain an SQL statement; otherwise, generating a prompt signal, re-inputting the data, interacting through diversified data, and carrying out standardized verification on the input data before generating the SQL statement, so that the statement generation times of nonstandard information are reduced, and meanwhile, fine adjustment is carried out on the large model in the field of electric power auditing, so that the generation accuracy of the SQL statement is improved.
Owner:国网安徽省电力有限公司综合服务中心 +2

Defect intelligent identification method and system for spunlace nonwoven fabric

The application relates to the field of image processing, in particular to a spunlace non-woven fabric defect intelligent identification method and system, which comprises the following steps: dividing a standard non-defect spunlace non-woven fabric image into multiple local image blocks according to spatial positions; constructing a local spectrum prior atlas corresponding to the image spatial positions; performing block processing on a to-be-detected image to obtain multiple original image blocks, and dynamically generating a personalized frequency domain mask for each original image block based on the local spectrum prior atlas to suppress the frequency band corresponding to normal textures; constructing a preliminary defect residual graph and a spectrum graph thereof; in response to the proportion of the medium-low frequency region energy in the total frequency domain energy in the spectrum graph exceeding a preset threshold, updating the mask parameters, applying the updated mask parameters to the personalized frequency domain mask, and obtaining a final defect residual graph; performing threshold segmentation and / or connected domain analysis on the final defect residual graph, and outputting a defect position. The application has the effect of improving defect detection accuracy.
Owner:HANGZHOU HANFORD TECH CO LTD

An algorithm for accurately importing PCB component coordinates based on multi-point marking

The present invention discloses a PCB component coordinate precision import algorithm based on multi-point marking, S1, first set four marking points and two coordinate systems, the PCB original coordinate system is a plane rectangular coordinate system (xOy), the angle of which is 90 degrees, the SMT working coordinate system is a plane oblique coordinate system (x'O'y'), the angle of which is θ, the conversion space system of any point N from the (xOy) coordinate system to the (x'O'y') coordinate system can be described as N'(x', y') = F(N(x, y)), the present invention relates to the technical field of SMT placement machine equipment. The PCB component coordinate precision import algorithm based on multi-point marking effectively counteracts the linear errors brought by the placement machine system and the PCB board itself, improves the generation accuracy of the SMT working coordinates, and thus improves the placement accuracy. Since the high-precision coordinates are automatically generated, no manual modification and calibration is required in the later stage, which improves the generation efficiency and reduces the labor cost.
Owner:HUAWEI SILKROAD

Method for generating mouth shape expression through audio of large model based on emotion reference and discrimination

PendingCN121962371AImprove generation accuracyResolve technical issues that prevent the expression of complex emotional statesSpeech analysisCharacter and pattern recognitionPattern recognitionComputers technology
The embodiment of the invention provides an audio generation mouth shape expression method of a large model based on emotion reference and discrimination, and relates to the technical field of computers, and the method comprises the steps: obtaining audio data and corresponding text data, emotion category labels and emotion reference videos; extracting audio features from the audio data through the large voice model, and extracting text features from the text data through the large voice model; extracting an emotion feature and an emotion reference feature from the emotion category label and the emotion reference video; performing facial parameter prediction based on the audio features, the text features, the emotion features and the emotion reference features by using a facial parameter generation model to obtain predicted facial parameters; the face parameter generation model is a large model obtained by performing emotion discrimination and content discrimination on the basis of historically predicted face parameters and performing model training according to a discrimination result; and generating a facial mouth shape expression according to the predicted facial parameters. According to the method, the expression generation accuracy in the complex emotional scene can be improved.
Owner:BEIJING JULI DIMENSION TECH CO LTD