1723 results about "Conversion method" patented technology

The invention discloses a conversion method and system based on SysML and Modelica model semantic mapping. The method comprises the following steps that S1, a Modelica standard model library is imported into a SysML modeling tool; s2, completing the definition of behavior elements, and forming an SysML model instance; s3, exporting a Modelica model which can be executed by the Modelica simulator; s4, establishing a mapping relation table between SysML model elements and Modelica model elements; s5, when a change event occurs in the SysML model, updating the content of the corresponding Modelica model; s6, the Modelica model is imported into the SysML model, and a corresponding SysML model instance is generated; s7, simulation result mapping and behavior modeling generation are executed, a state transition diagram structure is formed, and structure updating of internal block diagrams in the system model is completed. According to the method, bidirectional conversion and joint simulation optimization between models are realized, and the method is suitable for cross-platform modeling, collaborative design and functional verification scenes of a complex system.

The invention discloses an MRI-PET mode conversion method and system based on a generative adversarial network, and belongs to the technical field of artificial intelligence medical image generation. And the multi-scale structure representation injection module injects multi-scale anatomical prior information at different stages of the encoder, and overcomes the limitations of insufficient utilization of prior information and single injection scale. And the adaptive semantic residual fusion module adopts semantic attention guidance and double-branch attention weighting, adaptively fuses fine-grained local features and global context information, harmonizes the difference between the fine-grained local features and the global context information in an abstract level and a semantic category, and solves the problems of feature conflict and semantic fuzziness in a bottleneck region. The direction sensing space-frequency discriminator realizes multi-dimensional and fine-grained adversarial supervision through a space, frequency and local image block multi-branch collaborative discrimination mechanism, and improves the structural fidelity and spectrum authenticity of a synthetic image. And the generated image is superior to the existing method in indexes such as structural similarity and peak signal-to-noise ratio, and has higher clinical practical value.

The invention belongs to the technical field of multi-modal intelligent data management, and discloses a multi-modal business data co-processing and intelligent conversion method, which comprises the steps of fusing multi-source data, coupling cross-modal time sequence association, detecting image-text semantic conflicts, and generating an RAG semantic unit and a space-time alignment semantic packet. Constructing a four-dimensional priority label to divide resource pool categories, dynamically allocating resource pool computing resources, generating a low-redundancy feature set, triggering a forced preemption strategy during resource competition, and synchronously generating a resource arbitration log; injecting a dynamic rule node according to a data type, executing confidence arbitration in combination with a conflict arbitration identifier, generating business decision content and generating a rule correction decision packet; constructing a full-link semantic thinking chain, synchronously capturing a system operation track, and aligning to generate a dual-channel tracing report; identifying residual sensitive fields, triggering hot update to load corresponding rules, quantizing effectiveness to generate compliance reports, reversely updating an industry rule base, and optimizing a resource allocation strategy.

The present application relates to a waveform file conversion method and apparatus for chip testing, and a computer readable storage medium. The method comprises: performing content analysis on a waveform file to be converted, to acquire period information, pin information, and state information of pins in each period of said waveform file; generating a PAT-format digital signal waveform file; and on the basis of a mapping relationship between pins and ATE channels, converting the PAT-format digital signal waveform file into a binary vector file for ATE testing. A waveform file in any format is converted into a PAT-format digital signal waveform file which uses periods as rows and pins as columns, so that each row vector not only contains current time information but also contains a logical state of each pin; in addition, to facilitate understanding, the PAT-format digital signal waveform file is further converted into a binary vector for ATE testing, so that efficient processing of waveform files in a plurality of different formats can be realized, thereby improving the universality of conversion.

The invention discloses an AI scene conversion method and system based on visitor behaviors. According to the method, visitor behavior data are collected in real time through a multi-mode sensor network, and a UWB positioning base station, a millimeter wave radar module, a high-definition camera, a microphone array and a pressure sensor matrix are included. The system adopts an edge computing architecture, is equipped with an NVIDIAJetsonAGXXavier processor and a 512-core VoltaGPU (Graphics Processing Unit), and operates three core AI models to realize intelligent analysis. The LSTM residence time prediction model adopts a three-layer bidirectional architecture, each layer is configured with 256 hidden units, and visitor behavior sequences within 30 minutes are analyzed to predict the residence time. The Transform interaction intention recognition model comprises six layers of encoders and eight multi-head attention modules, and visual, voice and behavior three-mode characteristics are fused to judge visitor interaction requirements. According to the technical scheme, self-adaptive adjustment of the exhibition hall environment is achieved, the visitor satisfaction degree is improved by 28%, the operation cost is reduced by 15%, and important technical support is provided for intelligent transformation of modern exhibition places.

The invention provides a structured data representation method, a structured data conversion method, a structured data scheduling method, a structured data conversion device, a structured data scheduling device, a structured data conversion system, a medium and electronic equipment. The representation method comprises the following steps: responding to an operation of dragging a module from a module library to a visual editing canvas by a user to form a node corresponding to the dragged module; responding to the operation that a user establishes visual connection between input ports and output ports of different nodes on the canvas to form edges, wherein the edges define a data flow and calculation dependency relationship between the nodes; a structured data representation of a machine-readable AI model structure is generated based on all the nodes on the canvas and user-configured parameters thereof, and a set of edges defining the edge interconnection relationship. According to the representation method, conversion from graphical representation to structured data of the artificial intelligence model is completed.

The invention provides an execution feedback and iterative optimization-based natural language to SQL (Structured Query Language) conversion method. The method comprises the following steps of: generating a plurality of candidate SQL statements; performing parallel execution on the candidate SQL statements, and capturing grammar error information in execution results and line number classification numbers of different candidate SQL return results; generating an error pattern abstract based on the grammar error information, converting the error pattern abstract into a natural language constraint instruction, and injecting the natural language constraint instruction into a prompt word generated in the next round; when the line number classification number is smaller than or equal to a preset threshold value, a self-consistency voting mechanism is adopted, a candidate SQL set with the largest number of returned same lines is selected, and the SQL with the shortest execution time is selected from the set to serve as final output. According to the method, the accuracy, robustness and execution efficiency of conversion from the natural language to the SQL can be remarkably improved by executing a feedback-driven multi-round iterative optimization mechanism.

According to the statement conversion method based on the fine tuning model, semantic information and query intention of the query text are analyzed from two aspects of slot positions and operation types through the multi-task model, so that the accuracy of the query statement is improved, and semantic completion is performed on the query statement by reconstructing the slot position relation graph, so that the accuracy of the query statement is improved. Therefore, a relation topology basis is provided for subsequent template matching, and the semantic analysis integrity of the complex query statement is improved. And finally, according to the slot position relation graph, an SQL template conforming to the query intention and semantic information corresponding to the original query statement is determined, and the target SQL statement is obtained by splicing and assembling the target SQL template based on the slot position relation graph, so that the writing difficulty of the query statement is reduced. When the method is applied to a data query function of a system in the financial business field or the medical field, the input convenience of query statements can be improved, and the statement conversion accuracy is improved.

The invention discloses a bidirectional reversible conversion method and system between a peptide molecule SMILES and a sequence expression. The core innovation lies in that a new sequence description syntax is defined to retain information of a polypeptide special bond and specific modification of amino acid; a main chain atom index and adjacency traversal topology identification algorithm is adopted, and end group and topology integrated detection and coding are carried out; a residue recognition algorithm for main chain cutting and template library matching is compatible with any standard or non-standard amino acid residues, an extensible end group library / monomer template library and an automatic increment mechanism, and automatic recognition and sequence annotation of S-S disulfide bonds; the invention relates to a high-fidelity assembly algorithm of HELM anchor points and topology aware cyclic peptide processing. The method solves the problems of incapability of supporting a complex polypeptide topological structure, poor reversibility, insufficient expansibility of a monomer library and the like in the prior art, can be widely applied to scenes of quantitative structure-activity relationship model construction, large-scale polypeptide data cleaning and the like, and has remarkable practicability and innovativeness.

The invention discloses an equipment agent processing program conversion method and device based on cooperation of a large model and a small model. An equipment agent is constructed by analyzing control systems, kinematics structures and process constraint information of a source machine tool and a target machine tool, and line-level instructions are abstracted into system-independent unified semantic representation based on a multi-system numerical control corpus. And performing grammar analysis and semantic mapping on the source program, and generating a conversion task in combination with grammar rules, kinematics accessibility and process security constraints of a target machine tool. A small model is finely adjusted through supervised learning to realize basic generation capability, a large model is introduced as a patch evaluator, candidate results are jointly scored from different dimensions, knowledge migration is realized based on improved near-end strategy optimization, and a small model generation strategy is continuously optimized. And finally, high-quality results are screened through confidence degree sorting, cross-system G / M instruction and multi-axis track generation and verification are completed, and high-precision automatic conversion of numerical control programs is achieved.

The invention belongs to the technical field of geospatial information systems, and particularly relates to a component-level catenary three-dimensional model lightweight three-dimensional tile conversion method, which comprises the following steps: S1, forming a component-level fine model data set with semantic annotations; s2, constructing a detail level division rule; s3, reconstructing a lightweight linear component model; s4, constructing a comprehensive evaluation function Q, deleting the body-shaped components of which the Q values are smaller than a preset threshold value, and retaining the body-shaped components of which the Q values are greater than or equal to the preset threshold value; s5, obtaining optimized body-shaped component model data; and S6, according to the detail level division rule in the S2, assembling and recombining the lightweight linear component model in the S3 and the optimized body-shaped component model data in the S5, generating a hierarchical component structure, converting each component into a three-dimensional instance model, and outputting a three-dimensional tile data structure. And efficient organization and visualization of the contact network BIM model in the GIS environment are realized.

The embodiment of the invention provides a real-time 2D-3D conversion method, device, product and system. The method comprises the following steps: acquiring a 2D video stream to be converted, wherein the 2D video stream comprises a plurality of 2D frames; and asynchronously and concurrently executing the following steps of depth reasoning and 3D rendering on the plurality of 2D frames: performing depth reasoning on the 2D frames by using a depth reasoning model to obtain a corresponding depth map; performing 3D rendering on the 2D frame based on the obtained depth map to obtain a 3D frame corresponding to the 2D frame; and outputting the obtained 3D frame in real time. Depth reasoning of the 2D frames and the 3D rendering process are decoupled, asynchronous parallel execution of depth calculation of the multiple 2D frames in the 2D video stream and the 3D rendering process is achieved, the end-to-end delay of converting the 2D video into the 3D video is greatly shortened, and the real-time performance of converting the 2D video into the 3D video is improved.

The invention provides a video dubbing language conversion method, a video dubbing language conversion system and related equipment. The method comprises the following steps: acquiring audio track data from a video to be converted; carrying out human voice extraction on the audio track data and classifying according to roles to obtain a single speaker audio of each role; performing voice-to-text conversion on the single speaker audio of each role to obtain an original language copywriting of each role; performing sound cloning on the single speaker audio of each role to obtain a timbre model of each role; performing target language translation on the original language copywriting of each role to obtain a translated copywriting of each role; based on the translation copywriting of each role and the tone model of each role, performing text-to-voice conversion to obtain a translation audio of each role; and performing replacement of each role translation audio on the audio track data in the to-be-converted video to obtain a dubbing conversion video. According to the technical scheme, language video dubbing conversion combined with the tone of the speaker is achieved, the video is more diversified, and the user requirements can be better met.

The invention discloses a lip language-to-voice conversion method and device, a computer storage medium and a terminal, and aims to solve the problems that a lip language-to-voice conversion technology cannot be deployed on terminal equipment and voice quality cannot meet application requirements. The method is combined with a neural vocoder which reduces calculation complexity and resource requirements, system configuration requirements are reduced while system parameters are reduced, a design basis is provided for deploying a lip language-to-speech method on terminal equipment, and context-related visual feature sequences and user audio embedding vectors are fused, so that the user audio-to-speech conversion efficiency is improved, and the user audio-to-speech conversion efficiency is improved. A Mel spectrum acoustic feature sequence used for being converted into a voice waveform is obtained, and a user audio vector fused with the Mel spectrum acoustic feature sequence is obtained, so that the output voice waveform is more consistent with the real voice of a user; according to the embodiment of the invention, technical support is provided for deploying and applying the lip language-to-speech method meeting the speech quality requirement on the terminal equipment.

The invention belongs to the technical field of data processing, and provides a scale detection and scale conversion method and system based on an ultrasonic image, and the method comprises the steps: constructing a discrimination function fusing multi-modal features according to the spectral basis functions of different scales, and determining the scale contour of the scales in the ultrasonic image; the design can dynamically adapt to the spectral characteristics and structural characteristics of the scale, can effectively cope with the interference of scale diversity, ultrasonic speckle noise and the like of equipment of different manufacturers, and ensures the stability of scale positioning and subsequent conversion; the method comprises the following steps of: establishing a biomechanical-image deformation coupling model by taking a scale contour of a scale in an ultrasonic image as a geometric reference and combining biomechanical characteristics of an ultrasonic probe and tissues to obtain a corrected pixel distance between adjacent scale points of the scale; and curvature compensation for eliminating refraction errors of ultrasonic waves in curved tissues is introduced to correct a scale pixel-millimeter conversion coefficient, so that the pixel-millimeter conversion error rate is reduced to 1.2%, which is far lower than the level of a traditional method.

The invention provides a document adaptive conversion method and device based on a multi-modal large model and a medium, and belongs to the technical field of data processing. The method comprises the following steps: inputting a multi-modal document into a pre-constructed multi-modal large model, wherein the multi-modal large model comprises a multi-modal joint framework fusing a ViT visual model and an LLM model; using a ViT visual model and an LLM model to respectively extract multi-scale visual features corresponding to non-text content and semantic features corresponding to text content, and using a cross-modal attention mechanism to bidirectionally align the multi-scale visual features and the semantic features to obtain a multi-modal document; carrying out self-adaptive blocking on the multi-modal document by using a self-adaptive blocking strategy, and modeling a relative position relationship between blocks; establishing a dynamic mapping rule from a document element to an HTML tag, and mapping the multi-modal document into a front-end interaction component in combination with a natural language instruction of a user; and dynamically rendering the front-end interaction component in an on-demand mounting and resource isolation mode. The problem of insufficient document restoration capability in the prior art can be solved.

The invention discloses an interface conversion device and circuit, electronic equipment and an interface conversion method, and relates to the technical field of computer system structures. Extracting a cache or memory transaction request from the first preset interface signal through a channel processing module, generating consistency transaction information, sending the consistency transaction information to a consistency transaction concurrent processing module, receiving response data of the consistency transaction concurrent processing module, and packaging the response data into a second preset interface signal; the consistency transaction concurrent processing module carries out processing according to the received consistency transaction information to obtain a concurrent signal; and the interface conversion module decodes and converts the concurrent signal into a first target protocol control signal, and / or converts the authorized transaction into a second target protocol control signal. Therefore, concurrent analysis, consistency transaction mapping and direct conversion are carried out on a multi-channel protocol through a configurable modular hardware architecture, and high-concurrency and low-delay protocol conversion and cache consistency maintenance between an inter-chip consistency interconnection protocol and an on-chip bus protocol are realized.

The invention relates to an unmanned end-to-end sensing method and system for an unstructured road scene, and belongs to the technical field of automatic driving. A point cloud supervision view angle conversion method is adopted, point cloud densification is realized by constructing a local subgraph, and multi-view angle depth estimation is optimized in combination with an optical flow consistency constraint; according to the method, long and short focus image features are fused through a cross attention mechanism, camera internal reference information is embedded, image BEV features are enhanced, and the long-distance perception capability is improved; according to the method, dynamic interference such as flying dust is effectively recognized and filtered through a space-time attention mechanism, a unified deep learning framework is adopted, the problems of information loss and error accumulation in a traditional modular framework are avoided, and unstructured scene perception is effectively improved. The problems of long-distance sensing, blocking of dust raising and the like, inaccurate space conversion and the like in an unstructured road scene are solved.

The invention provides an efficiency optimization control method of a permanent magnet synchronous motor and related equipment. The method comprises the following steps: acquiring an output current of an inverter; constructing an inverter loss model based on the output current of the inverter and the inverter loss coefficient; constructing a total loss model based on the electromagnetic loss model, the inverter loss model and an equivalent conversion method; constructing a current population by taking the plurality of d-axis current values as individuals; and performing optimization solution based on the total loss model, the current population, a PID search algorithm and an artemisinin optimization algorithm to obtain a global optimal solution and control the permanent magnet synchronous motor. According to the scheme, the driver loss is added into the total loss model, and efficiency optimization can be carried out on the whole motor driving system; in order to solve the problem that a traditional method is not suitable for an IPMSM, a PID search algorithm and an artemisinin optimization algorithm are combined, parameter search is carried out on a motor efficiency optimal point by simulating an incremental PID controller adjusting mechanism, a global optimal solution is obtained, and therefore optimal control over the click efficiency is achieved.

The invention discloses a quantum circuit conversion method and system, electronic equipment and a storage medium, and relates to the technical field of quantum circuit conversion, and the method comprises the steps: extracting a feature vector of a logic quantum circuit based on initial quantum bit mapping; the feature vector is input into a pre-trained decision tree, recommendation parameters of a line conversion algorithm are obtained, the decision tree is constructed by a training set, and the feature vector of each sample corresponds to the optimal parameter of the decision tree, namely the parameter with the minimum number of physical line gates; and configuring a line conversion algorithm by using recommended parameters, converting a logic line into a physical line, and ensuring that all two quantum bit gates meet quantum computer connected graph constraints and are equivalent in function. According to the method, adaptive parameter configuration can be achieved through the feature vectors and the decision tree, the line conversion quality is effectively improved, the number of physical line gates is remarkably reduced, and the quantum program execution efficiency is improved on the premise that function equivalence and connectivity are guaranteed.

The invention particularly relates to a Word document format conversion method based on Java. The Word document format conversion method based on Java comprises the following steps: analyzing the content of an original. Doc file, and extracting text paragraphs, tables, pictures, style information and document metadata; the extracted content is divided into different categories, and the XML node type corresponding to each category of elements in the target. Docx document is established; the method comprises the following steps of: constructing a pattern mapping rule base, constructing a new. Docx document structure by using an XWPF Document object model according to an Office Open XML (Extensible Markup Language) specification, sequentially inserting paragraphs, tables and pictures, and applying corresponding pattern configuration; and outputting and generating a. Docx file, detecting and processing abnormal conditions, and recording a conversion log at the same time. The Word document format conversion method based on Java is efficient and accurate, has good compatibility, expandability and cross-platform capability, is suitable for enterprise-level document management systems, cloud services and batch document processing scenes, and remarkably improves document compatibility and processing efficiency of office automation systems.

The invention discloses a protocol conversion method and device, a medium and a server system, and relates to the technical field of servers. Once an original protocol type is determined, a dynamic interface routing controller calls a corresponding conversion program stored in a mapping table; and converting the detected heterogeneous bus protocol command into a unified single-line serial bus protocol operation code. As the conversion program and the protocol characteristics are stored in the mapping table, the method has infinite protocol expansion capability, a new protocol can be supported by updating the mapping table and a detection engine without redesigning hardware or modifying codes of an editable logic device, the system design is greatly simplified, and the development cost is reduced. Therefore, the problem that the current programmable logic device can only realize single interface protocol conversion is solved.

The invention relates to the technical field of data processing, and discloses a test case file conversion method and device, computer equipment and a storage medium. The test case file conversion method comprises the steps of obtaining a to-be-converted file, and analyzing the to-be-converted file according to a file type of the to-be-converted file to obtain file analysis data; performing conversion processing on the file analysis data according to a preset intermediate conversion rule to obtain an intermediate conversion file; obtaining format selection information, and determining a target file format and a target file template corresponding to the target file format according to the format selection information; and according to the intermediate conversion file and the target file template, generating a target conversion file corresponding to the target file format. According to the embodiment, a unified intermediate file processing mechanism is established, meanwhile, automatic processing of file conversion is achieved, the file conversion efficiency and accuracy are improved, and different scene requirements can be met.

The invention discloses an industrial design-oriented drawing semantic analysis and structured conversion method and system. The method comprises the following steps: acquiring an input image, carrying out different-scale coding on image features through an image coding backbone network, selecting a specific layer to extract a multi-scale feature map, and fusing to generate multi-scale features; inputting the multi-scale features into a geometric primitive detection network and an other element detection network, respectively identifying geometric primitives and non-geometric primitives, and carrying out positioning and classification; analyzing mutual relations between geometric elements and other elements contained in the graph through an element relation graph network, and generating formalized language description based on a matching rule; a Prompt template is constructed, geometric information is supplemented by using a multi-modal large model, the rationality of the supplemented information is verified through a geometric attribute relationship verifier, and complete formalized language description is obtained. According to the method, the industrial drawing image containing the complex constraint relation can be efficiently and accurately converted into machine-readable structured data.

The invention discloses a Token conversion method and system for converting multi-modal information to a unified control instruction, and the method comprises the steps: obtaining a multi-modal perception information flow and instruction reference data, carrying out the feature deconstruction of the multi-modal perception information flow, recognizing a modal identification code, carrying out the cross-domain mapping calculation based on the modal identification code and the instruction reference data, generating an offset vector, and carrying out the Token conversion of the multi-modal information to the unified control instruction. Performing tensor synthesis on the offset vectors to construct a Token coding library; performing intention flow direction analysis according to the Token coding library to generate an intention propagation domain, performing semantic potential energy optimization on the propagation domain to form an optimal intention link, and generating an adaptive Token flow; the input analysis layer, the logical reasoning layer and the output mapping layer are positioned and recognized through semantic backtracking, semantic nodes are implanted, node balance adjustment is executed, and a cooperative processing chain is formed; feature reverse extraction, reasoning complementation processing and deep reasoning processing are executed through a cooperative processing chain, a three-layer control network is formed, a control Token sequence is generated, and efficient conversion from information to instructions is achieved.

The invention belongs to the technical field of industrial Internet of Things data processing, and particularly discloses a data protocol conversion method and system based on the industrial Internet of Things, and the method comprises the steps: extracting a feature vector through a protocol feature recognition model according to an original data stream of target equipment, and obtaining a matching result of a dynamic protocol fingerprint database; according to a matching result of the dynamic protocol fingerprint database, judging whether a protocol learning mode is triggered to obtain a newly added fingerprint and an analysis rule; according to the analyzed structured source data and the domain ontology library, semantic annotation is carried out through a semantic understanding engine, and semantic data with physical quantity tags are obtained; and according to the real-time resource state parameters, a conversion strategy is dynamically selected, an unloading point is decided and calculated, and a lightweight and full-amount depth conversion instruction is obtained. The invention aims to solve the problem of low protocol conversion efficiency caused by protocol isomerism, semantic difference and resource constraint in the industrial Internet of Things in the prior art.

The invention discloses a method for converting a natural language into a structured query language by considering accuracy and reasoning cost. The method comprises the following steps: firstly, performing task difficulty judgment on an input natural language query statement through a multi-dimensional scoring method; then diversified example data are generated through a data enhancement method, and an enhanced example library is formed; a semantic vector retrieval method is adopted, and a sample most relevant to input is selected; different cue words are generated according to different complexity levels of the query task, and large language models of different parameter scales are called for SQL generation; and finally, carrying out structure and semantic consistency verification on the generated SQL statement, if inconsistency exists, automatically constructing a correction prompt word, guiding the model to carry out correction, and finally outputting the SQL statement with a correct structure and accurate semantics. According to the method, the reasoning cost is remarkably reduced while the generation accuracy is improved, and the method has good practicability and expansibility.

The invention provides a Text-to-SQL (Structured Query Language) conversion method based on semantic modeling, which relates to the technical field of natural language processing and data analysis, and comprises the following steps: designing a structured semantic model comprising a fact table and a dimension table by adopting a dimension modeling method, and defining business field division, data model definition and model dictionary construction; performing word segmentation processing on the natural language question input by the user, matching the natural language question with dimensions, indexes and dimension values in the structured semantic model based on the model dictionary library, and generating Schema information containing database types, table names, field definitions and filtering conditions; calling the large model, and converting the natural language problem into a preliminary semantic SQL in combination with Schema information and a preset rule; calling the large model again, performing logic verification and grammar optimization on the preliminary semantic SQL, and generating a corrected semantic SQL; based on Schema information and data model definition, the corrected semantic SQL is converted into an executable database SQL statement.

The invention discloses an image scene conversion method and device, equipment and a storage medium. Firstly, an image to be processed and a conversion cue word used for indicating an image scene conversion direction and a specific attribute adjustment requirement can be obtained, the conversion cue word is input into a pre-trained scene noise optimizer, and structured guide noise accurately matched with the cue word is generated through noise optimization processing and used as target noise. Performing vector encoding on the to-be-processed image and the conversion prompt word through an image encoder and a text encoder to obtain an image feature vector and a text feature vector, and performing feature fusion on the image feature vector and the text feature vector by using a cross attention mechanism to generate a fusion vector; and finally, jointly inputting the fusion vector and the target noise into a pre-trained scene conversion model, and performing conditional decoding generation processing to obtain a scene conversion image corresponding to the to-be-processed image. According to the invention, disordered disturbance caused by noise injection is avoided, and drift of the image content in the conversion process is effectively inhibited.

A floating-point data precision conversion method includes determining a bit width of a second mantissa field based on a coded value of a first exponent field. The floating-point data precision method further includes determining a reserved coded value and a discarded coded value in a first mantissa field. The floating-point data precision method further includes, if the coded value of the first exponent field is greater than or equal to a first preset threshold, performing a rounding operation on the reserved coded value based on a coded value that starts from a most significant bit and whose bit width is a preset bit width in the discarded coded value, to obtain a coded value of the second mantissa field.