447 results about "Semantic layer" patented technology

The invention discloses an unmanned aerial vehicle image small target detection method based on dynamic filtering and an adaptive sparse Transform. According to the method, an end-to-end target detection framework is adopted, a dynamic filtering module is introduced into a backbone network, global feature interaction is achieved through data-dependent frequency domain operation, and linear calculation complexity is maintained. For feature interaction in a scale, an adaptive sparse Transform module is introduced to enhance the capability of focusing key information on high semantic hierarchy features of a model, and noise interference and feature redundancy are effectively suppressed at the same time. Through the combination of dynamic filtering and adaptive sparse Transform, the model can extract image foreground information more effectively on the premise of not significantly increasing the calculation burden, and the problem that a traditional target detection model is susceptible to complex background interference is significantly relieved.

The invention provides an RAG knowledge base construction method and system based on hierarchical semantic indexes. The method belongs to the cross technical field of artificial intelligence and information retrieval. The method comprises the following steps: performing multi-level semantic analysis on an input original document set to generate document semantic hierarchical structure data; constructing a hierarchical semantic index tree based on the document semantic hierarchical structure data; performing dynamic knowledge graph initialization according to the hierarchical semantic index tree to generate an initial dynamic cognitive graph; and collecting real-time interaction data through a user feedback interface and a new data access module, and performing incremental updating on the initial dynamic cognitive map to form a knowledge representation system supporting life cycle evolution. Through multi-level semantic analysis and construction of a hierarchical semantic index tree (HSIT), deep semantic analysis can be performed on an original document set, the context integrity of knowledge is ensured, structured storage is realized, the knowledge can be expressed and stored more accurately, and information loss or semantic ambiguity is avoided.

The invention provides a semi-supervised image semantic segmentation method and system based on a visual basic model, and the method comprises the steps: constructing a multi-task model which comprises a visual basic model and a depth estimation basic model, and the visual basic model is connected with a task solution head, an adapter parameter efficient fine tuning module and a multi-modal cross fusion module; the task solution head comprises a semantic segmentation head and a depth estimation head; extracting semantic hierarchy features and a depth feature map of the RGB image, performing cross attention fusion on the semantic hierarchy features and the depth feature map, and inputting obtained fusion features into a semantic segmentation head and a depth estimation head respectively; semi-supervised learning is adopted to train a multi-task model, only parameters in the adapter parameter efficient fine tuning module and the multi-modal cross fusion module are trained, and a multi-task loss function is adopted. The image semantic segmentation model obtained through training can improve semantic segmentation performance, reduce training cost and is suitable for different tasks.

An illustrative method for querying multiple datasets may include generating, based on models each associated with and defining attributes of a different datasets stored in a plurality of data stores, a semantic layer defining relationships between the models and that provides a centralized application programming interface (API) for exposing the datasets by way of a common query language, receiving, by way of the centralized API, a query request for information that depends on data included in multiple datasets included in the plurality of datasets, querying, based on the query request and the relationships between the models defined by the semantic layer, the multiple datasets, and presenting, based on the querying, a query result representative of the information that depends on the data included in the multiple datasets.

The invention provides a code generation method and system based on a waterfall model and multi-agent collaboration, and the method comprises the steps: constructing a multi-agent collaboration framework which comprises a problem analysis agent, a solution agent, a pseudo-code agent, a coding agent and a restoration agent based on a software development waterfall model thought; and the interaction process of each agent is coordinated through a dynamic cooperation algorithm. Wherein the problem analysis intelligent body checks similar problems and solutions; the solution intelligent agent generates and evaluates candidate schemes; performing scheme conversion by the pseudo-code intelligent agent; the coding agent generates an executable code; and the repair agent performs fine-grained repair on the code in grammar, runtime and semantic levels through a two-dimensional repair mechanism. According to the method, the limitation of a single agent in a complex programming task is broken through, automatic code generation covering the whole process of demand analysis, scheme design, code implementation and test repair is realized, and the quality of generated codes and the reliability of operation are remarkably improved.

The invention belongs to the technical field of text processing, and particularly relates to a method and system for retrieving DOCX document content based on keywords, which comprises the following steps: analyzing an Office Open XML structure of a DOCX document, combining with multi-dimensional features such as style names, and utilizing a title classification score model to accurately distinguish a title and a text, so that a semantic hierarchical structure of the document is effectively reserved; and secondly, a multi-level semantic extension mechanism is introduced, and a Sension-BERT, a HowNet knowledge base and a Word2Vec model are fused, so that intelligent extension of synonyms and synonyms of keywords is realized, and the recall rate and semantic understanding ability of retrieval are remarkably improved. And in addition, a BM25 model is combined with paragraph length normalization and structure position weight to calculate a correlation score, so that retrieval results are sorted more accurately and reasonably. The construction of the reverse index is combined with the position coding and compression optimization strategy, and the retrieval efficiency and the storage performance are both considered.

The invention relates to the technical field of natural language processing, in particular to a context-based Prompt generative model optimization system, which comprises a context analysis module, a cue word generation module, a context optimization module, a semantic check module and a structure reconstruction module. According to the method, the context path and the semantic hierarchy information of the semantic unit are introduced, the fine degree of semantic matching degree recognition is improved, semantic guide deviation caused by statement template solidification is avoided, the cue words are recombined in combination with the semantic coherence weight and the logic dependency relationship, and the recognition accuracy is improved. The consistency and expression accuracy of the prompt content in the context are enhanced, the prompt word insertion sequence and connection mode are dynamically adjusted through a semantic conflict detection and structure rechecking mechanism, coherence and stability of a semantic structure and controllable generation of the prompt content are kept, semantic conflicts and expression chaos caused by static matching are effectively avoided in the generation process, and the generation efficiency is improved. And dynamic adaptation of prompt configuration and smooth optimization of language output are integrally realized.

The invention belongs to the technical field of artificial intelligence and multimedia processing, and discloses a multi-modal content conflict resolution method based on modal time sequence dependence modeling, which comprises the following steps of: constructing a modal representation vector by unifying time, space, semantics and priority information of modeling modal contents; and further structuring and quantifying complex dependency and conflict relationships among different modal contents by utilizing a modal time sequence dependency graph, and constructing a context relationship by combining a graph neural network, so that the system can identify conflicts of the contents in time, space and semantic levels, and realizes self-adaptive content reordering and display configuration based on conflict scores. According to the method, the limitation of traditional single-dimension conflict detection is overcome, the accuracy rate and the coverage range of conflict identification are improved, the method can be widely applied to multi-modal content-intensive scenes such as intelligent media generation, AI broadcasting, virtual navigation and XR interaction, information shielding and semantic redundancy are reduced, and the interaction experience of users and the system content quality are enhanced.

The invention belongs to the field of artificial intelligence, and particularly relates to a disaster risk early warning algorithm based on multi-modal fusion and a space-time propagation model, which comprises a multi-modal data management module, a cross-modal semantic alignment module, a space-time propagation graph modeling module, a propagation constraint deduction module, a graded early warning generation module and an online closed-loop calibration module. Depth alignment of multi-modal data on a semantic level, dynamic deduction of a disaster propagation process under constraint of a physical mechanism and operable conversion of an early warning result in a business scene are realized, and the whole system takes a unified space-time framework as a base, takes a propagation graph as a link and takes closed-loop calibration as a guarantee, so that the accuracy of the system is improved. The structural defects of extensive data splicing, propagation modeling deficiency and decision chain fracture in a traditional early warning method are fundamentally overcome.

The invention relates to the technical field of intelligent recommendation, and discloses an AI intelligent matching method based on a knowledge graph, and the method comprises the steps: constructing a quaternary knowledge graph containing a time dimension; identifying legal entities and semantic relationships by adopting an entity identification and relationship extraction technology; multi-level semantic features are extracted through a two-layer progressive semantic matching algorithm of a grammar layer, a semantic layer and a reasoning layer; constructing lawyer ability portraits based on a heterogeneous graph neural network and a time sequence perception graph convolution technology; progressive matching calculation is adopted, the optimal matching weight is learned through a multi-layer attention mechanism, and dynamically optimized intelligent matching is achieved. The technical problems of cold start, insufficient semantic understanding ability and poor timeliness processing ability in the existing legal consultation matching system can be solved, and the matching precision and the user satisfaction are improved.

The invention discloses an Internet of Things multi-source heterogeneous data management method based on artificial intelligence, and the method comprises the following steps: collecting voice data and text data in an Internet of Things system, and constructing a voice and text pair; inputting the voice data into an improved Whisper model, and outputting an enhanced transliteration text; inputting the enhanced transliteration text and the text data into a semantic encoder based on dynamic weighting to generate a fused semantic embedding vector; calculating semantic similarity between the fused semantic embedding vectors, and screening semantic consistent sample pairs; performing event aggregation on the semantically consistent sample pairs; performing semantic level coding and confidence adaptive fusion processing to generate a unified semantic fusion vector; and constructing a backmarking training set based on expert feedback, and executing periodic increment optimization on the updatable parameters. The voice and text multi-modal data fusion quality and semantic consistency recognition precision are improved, and the method has good expandability and is suitable for intelligent data governance tasks in a complex Internet of Things environment.

The invention relates to a pavement disease detection method, system and device based on a hybrid architecture, and a storage medium. The method comprises the following steps: obtaining pavement image data; the method comprises the steps that data are input into a re-parameterized feature extraction network, the network is based on an HGNetV2 architecture, a RCHGBlock module is formed by embedding a RepConv structure into HGBlock, and a plurality of modules are cascaded and stacked to construct a four-stage progressive feature pyramid structure; based on a high-level semantic layer of a feature pyramid, integrating a space edge perception enhanced attention mechanism, enhancing disease edge features through a dual-path complementary processing framework of an edge extraction path and a standard convolution path, and performing multi-scale feature fusion by using an improved C3K2-SEAM module as a feature fusion unit; and performing end-to-end disease detection based on the fusion features, and outputting disease types, positions and confidence information. Compared with the prior art, the method has the advantage that the disease detection precision and robustness in a complex scene are remarkably improved.

The invention relates to the technical field of visual navigation, in particular to a blind person navigation path planning method combining visual SLAM and semantic segmentation. The method comprises the following steps: acquiring an environment image and pose data; performing front-end tracking by using the pose data, and determining a key frame sequence; generating a depth map based on the environment image, and identifying a passable area according to the depth map; constructing a three-dimensional point cloud map according to the key frame sequence; semantic segmentation is carried out according to the three-dimensional point cloud map, and obstacle type labels are recorded; according to the obstacle type label, carrying out safety level layering on the passable area, and determining a path cost weight; performing global path planning based on the path cost weight, and generating a semantic enhancement navigation trajectory; and driving real-time voice guidance by using the semantic enhancement navigation trajectory, and calculating a path execution deviation corresponding to the real-time voice guidance. According to the method, obstacle types are distinguished through semantic segmentation based on a visual navigation technology, a semantic enhancement path is generated, semantic hierarchical navigation is realized, and the intelligence of path decision is improved.

The invention relates to the technical field of natural language processing, and discloses a question retrieval method and system, electronic equipment and a storage medium. The method comprises the steps of performing semantic understanding and decomposition processing on an original question to obtain a plurality of candidate sub-questions; constructing an initial problem tree structure by taking the original problem as a root node and the candidate sub-problems as sub-nodes; rationality scores of the child nodes are determined, and the rationality scores represent rationality of candidate sub-problems corresponding to the child nodes; determining a relevance score between the nodes, the relevance score representing a semantic repetition condition between the nodes; based on the rationality score and the relevance score, optimizing the initial problem tree structure to obtain a target problem tree structure; and performing retrieval based on the target question tree structure to obtain answer data for the original question. By means of the method, rationality and relevance evaluation can be conducted on the problem tree nodes, and architectural disassembly and semantic hierarchical management of complex problems are achieved.

The invention relates to the field of target detection, and particularly discloses an electric power work target detection method based on multi-modal large model knowledge distillation, which utilizes a vision-language multi-modal large model as a teacher model, and improves the target detection efficiency by expanding prompt word guidance. A high-quality pseudo label and a region-text pair are generated for an unlabeled electric power work image as a supervision signal, and on this basis, through joint optimization of detection loss, feature distillation loss, logic distillation loss and multi-modal contrast learning loss, a lightweight YOLO student model is guided to learn positioning and classification knowledge and to learn a multi-modal contrast learning loss. And deep alignment with the open vocabulary understanding ability of the teacher model is carried out on the feature space and semantic level, so that a semantic gap between closed category detection and open world perception is effectively bridged. Through the mode, the detection precision and generalization ability of the student model on common, rare and even unseen targets in the electric power work scene are remarkably improved.

The invention discloses a ship knowledge graph construction method based on a large model and a graph neural network, and the method comprises the steps: carrying out the format conversion and protection type partitioning processing of a professional document, and completing the recursive segmentation through a placeholder protection formula, a table and other structures according to the semantic hierarchy, and obtaining a text block suitable for the extraction of an entity and a relation; extracting domain entities by utilizing a large language model containing thinking chain cues, and positioning candidate entities in combination with an AC automaton so as to improve the coverage rate and accuracy of relation triple extraction; performing new entity backfilling and relation standardization on an extraction result, and constructing an initial knowledge graph with a consistent structure; and inputting the initial knowledge graph into a graph neural network model with directional expansion and a multi-scale decoder, and complementing a missing relationship through link prediction, so that node distribution and a relationship structure of the graph are more complete. According to the method, a continuous processing chain from text preprocessing, knowledge extraction to graph completion is formed.

The invention discloses a dual-machine target positioning method based on semantic prior and spatial intersection constraint. The method comprises the following steps: 1, acquiring images of a camera 1 and a camera 2, and determining the spatial orientation of a camera coordinate system relative to a reference world coordinate system; step 2, obtaining a target normalized coordinate under a coordinate system of the camera 1; step 3, obtaining a target normalized coordinate under a coordinate system of the camera 2; 4, converting the normalized coordinates of the target into geometric rays in a three-dimensional space; 5, constructing an optimization function with the minimum distance sum of squares; step 6, performing semantic-level foreground and background division on the image content through a YOLOv5 network, and constructing a three-dimensional space feature point library of a semantic background; and 7, performing single-view semantic neighborhood retrieval and depth compensation by using the three-dimensional space feature point library of the semantic background constructed in the step 6, and realizing target three-dimensional positioning in a high-altitude long-distance complex scene. According to the invention, the detection robustness and the positioning precision of the high-altitude long-distance target in the environment are improved.

The invention discloses an electric power system data anomaly prediction method based on deep learning, and relates to the field of data anomaly prediction.The method comprises the steps that firstly, a graph convolutional network and LSTM are used for deeply mining time series data of an electric power system and spatial-temporal characteristics of a topological structure, and high-precision anomaly detection and positioning are achieved; and when an exception is detected, an exception attribution mechanism is introduced, and exception signal features (such as exception fragments and positioning information) at a mathematical level are converted into a structured query object at a semantic level. And then, a barrier between numerical data and the unstructured operation and maintenance knowledge base is broken through a vector retrieval technology, and related maintenance regulations and historical cases are accurately recalled. And finally, an executable recommended disposal scheme is automatically generated based on recall knowledge in combination with the generation capability of a large language model, so that an intelligent closed loop from anomaly perception and knowledge matching to decision assistance is constructed, and the accuracy and efficiency of power system fault handling are remarkably improved.

The invention discloses a power distribution network semantic analysis and intelligent number asking method and system based on NL2DSL, and relates to the technical field of power distribution network data query. The method comprises the following steps: receiving a natural language question of the power distribution network, clarifying fuzzy terms through multiple rounds of asking, and outputting a clarified question; based on the historical dialogue and the terminology library optimization problem, generating a standardized and optimized problem; taking a semantic layer of the power distribution network as a unique true phase source, and converting an optimized problem into an initial DSL (Digital Subscriber Line) by combining RAG and Prompt engineering; a compliance DSL is output through double-layer protection of static rule verification and dynamic service verification; and analyzing the compliant DSL and associating the pre-calculation model to generate an optimized SQL, and outputting a structured result after execution. The method can improve the query accuracy and the data security, reduces the maintenance cost, and is suitable for the full-scene intelligent number asking demand of the power distribution network.

The invention discloses an identity authentication method, a program product, an electronic device and a storage medium, and relates to the technical field of cloud computing, the method is applied to an identity issuing server, and the method comprises the following steps: obtaining a logic identity label of a first computing workload in a first computing node; acquiring a hardware trust metric value in a trusted execution environment of the first computing node; converting the hardware trust metric value into a standardized metric description object by using a preset semantic mapping table; generating a verifiable certificate of the digital signature according to the logic identity identifier and the measurement description object; sending the verifiable credential to the first computing node; according to the method, the measurement values generated by different TEE technologies can be abstracted and standardized in a semantic level by utilizing the preset semantic mapping table, so that identity authentication of cross-domain access is realized, and the technical barrier of heterogeneous TEE is broken; and the logic identity of the computing workload is strongly bound with the hardware credible state of the computing node, so that the security of identity authentication is improved.

The invention discloses an intelligent number asking method, device and system based on a semantic data model and a large model, and belongs to the field of artificial intelligence. After a current question of a user is received, a task plan corresponding to the current question of the user is obtained based on the large model, then MQL information extraction is performed based on the task plan to obtain a target MQL, then a word semantic data model converts the target MQL to obtain an SQL query statement, query is performed based on the SQL query statement to obtain reply data, and the reply data is sent to the user. And finally, returning the reply data to the user. According to the technical scheme, the natural language is converted into the MQL through the large model, then the MQL is converted into the SQL through the semantic data model, and compared with a traditional scheme that the natural language is directly converted into the SQL, data training does not need to be carried out; and the large model can accurately identify the user intention, so that the obtained MQL unified semantic layer provides a standard caliber, and the reply accuracy is greatly improved.

The invention belongs to the technical field of unmanned aerial vehicle navigation, and relates to an unmanned aerial vehicle autonomous navigation method and device based on semantics, equipment and a medium. The method comprises the following steps: acquiring an image sequence and IMU data of an unmanned aerial vehicle, and constructing a three-dimensional geometric skeleton of an environment; forming a two-dimensional semantic map according to the image sequence of the unmanned aerial vehicle; marking corresponding semantic tags according to the three-dimensional geometric skeleton and the two-dimensional semantic map, and taking the semantic tags and the geometric information as observation results; fusing observation results from different visual angles to obtain a semantic-geometric coupling map; according to the semantic-geometric coupling map, different nodes are generated, and a dynamic three-dimensional scene graph is obtained by taking a relationship between the nodes as an edge for connecting the nodes; and obtaining an unmanned aerial vehicle instruction, and generating a flight path according to the dynamic three-dimensional scene graph to realize autonomous navigation of the unmanned aerial vehicle. According to the invention, instructions can be understood in a semantic level, so that autonomous navigation of the unmanned aerial vehicle is realized.

The invention provides an AI-based ChatBI platform design and implementation method, which comprises the following core steps of: constructing a first semantic layer database, receiving a natural language problem of a user, and generating an executable SQL (Structured Query Language) statement through an NL2DSL engine; extracting keywords and performing vector similarity calculation on the keywords and the first semantic layer database to generate a second semantic layer knowledge base; judging query complexity according to the quantity of the knowledge bases, and scheduling corresponding processing engines; checking the user access authority, and automatically generating a chart configuration template and recommending an adaptive chart type when the user access authority exists; based on user selection, data interpretation and a core conclusion are generated through a large model in combination with SQL statements and a knowledge base, and an output result is integrated. According to the invention, by optimizing the platform architecture, accurate data insight can be obtained by non-technical personnel in a dialogue manner, the use threshold is reduced, terms are unified, and the result accuracy is improved; meanwhile, operation efficiency is improved, and waiting time is shortened; and the data security is enhanced through permission setting.

The invention relates to an academic paper abstract generation method based on an image-text fusion index, and belongs to the technical field of artificial intelligence and multi-modal information processing. The method comprises the following steps: acquiring academic paper data from a public data source; according to academic papers in the academic paper data, segmenting paper contents according to logic, establishing association between fragments and images, and constructing an image-text fusion index; performing semantic retrieval on four dimensions in the academic paper to obtain a final retrieval content set of all the dimensions; and generating an image-text fusion abstract based on the retrieval content set of the four dimensions. According to the method, by constructing the image-text fusion index, the image-text association strength and the chapter context coherence are enhanced, deep coupling of the image and the text in the semantic level is achieved, and the image-text collaborative expression ability of the abstract is improved; by designing a structured dimension template and a content retrieval mechanism driven by a standard problem, the structured coverage of key information is ensured, and the logic continuity and readability of the abstract are remarkably improved.

The invention belongs to the technical field of enterprise services, and particularly relates to a deep learning-based enterprise and policy information matching recommendation system and method.Enterprise feature vectors are constructed and converted into semantic embedding, so that enterprise attributes form continuous expression in a deep space; semantic unit division and context-aware vector generation are carried out on the policy text, and the policy semantic center is adjusted by fusing the overall distribution trend of the enterprise, so that the reality adaptability of the policy text is improved; a multi-view representation mechanism is introduced, and a policy support direction is described from different dimensions; and finally, generating a comprehensive matching score in combination with the confidence coefficient of semantic distance conversion and the explicit condition satisfaction degree. According to the method, the limitation of keyword matching is broken through, the alignment of enterprises and policies on the semantic level is realized, the discovery capability and matching accuracy of potential applicable policies are remarkably improved, and more enterprises which conform to conditions but are not completely consistent in expression form can be effectively identified and corresponding policies can be recommended.

The invention discloses an advertisement creativity matching method based on multi-modal content generation, and relates to the technical field of digital media content generation, and the method comprises the following steps: building a cross-modal time anchoring belt facing advertisement creativity matching, carrying out metaphor level decomposition on input text information, marking a symbol axis for image information, and carrying out data processing on the image information; obtaining an initial semantic boundary list; and constructing a culture fingerprint database according to the initial semantic boundary list, and mapping the territory taboo information and the brand symbol information into constraint tags to obtain a semantic guardrail set. According to the method, through cross-modal time anchoring and semantic boundary control, accurate correspondence of the text and the image in time and semantic levels is achieved, and it is ensured that generated content is clear in semantic meaning and adaptive in culture. In combination with breathing type phase traction and cultural fingerprint dynamic adjustment, multi-modal content rhythm and emotion are coordinated and unified, brand expression is kept stable, and the overall consistency and propagation effect of advertisement creativity are improved.

The invention discloses a hierarchical visual language navigation memory enhancement system in a cross-floor scene, relates to the technical field of intelligent navigation, and adopts the technical scheme that the hierarchical visual language navigation memory enhancement system comprises three core parts, namely a hierarchical visual semantic model construction module, a memory enhancement algorithm module and a navigation system integration module. The hierarchical visual semantic model construction module is used for constructing a basic visual feature layer, a floor semantic layer and a cross-floor semantic association layer; the memory enhancement algorithm module comprises a short-term memory module, a long-term memory module and a memory fusion and update strategy module; and the navigation system integration module is used for integrating the hierarchical visual semantic model and a memory enhancement algorithm into the system. Three-dimensional space characterization misalignment and path planning error accumulation can be effectively avoided, the accuracy, environment adaptability and decision-making efficiency of cross-floor navigation of the intelligent agent are remarkably improved, and the method has important technical innovation value and application prospects.

The invention discloses a double-domain RAG-driven multi-omics fusion pathology analysis system, and belongs to the technical field of artificial intelligence of medical data. Pathology image feature data and structured multi-omics data of a patient are fused in a semantic layer through a multi-modal fusion module, a semantic layer fusion result is obtained, and a comprehensive representation vector of the patient is generated; the double-domain retrieval module obtains internal reference evidence corresponding to a hospital case knowledge base and external reference evidence corresponding to an external medical literature knowledge base; the consistency gating fusion module analyzes the consistency of the internal reference evidence and the external reference evidence, and fuses the internal reference evidence and the external reference evidence to obtain a fused credible evidence; and the report generation module generates a medical auxiliary report with an evidence chain based on a large language model according to the semantic layer fusion result and the credible evidence. According to the embodiment of the invention, the interpretability and credibility of the diagnosis conclusion can be enhanced.

The invention belongs to the technical field of information retrieval, and particularly relates to a library document abstract generation method based on deep learning model construction, which comprises the steps of data preprocessing, model construction, semantic understanding and core viewpoint extraction, abstract generation and optimization and quality evaluation. According to the library literature abstract generation method based on deep learning model construction, the mixed labeling technology is combined with rules and deep learning, the advantages of the rules and the deep learning are fully played, the rules provide a basic framework, the deep learning makes up for the deficiencies of the rules, and domain term meanings can be more accurately understood; the domain entity knowledge base ensures accurate recognition and relation understanding of entities in literatures through a rigorous extraction and disambiguation method, a foundation is laid for high-quality abstract generation from the semantic level, and in addition, through accurate understanding of domain terms and effective capture of long text semantic association, the abstract generation efficiency is improved. The accuracy, the integrity and the readability of the generated abstract are improved, so that the requirement of a user for quickly acquiring the key information of the literature is better met.

The invention relates to the technical field of signal processing, and provides a hidden weak signal feature recognition method and system, and the method comprises the steps: carrying out the time-frequency transformation of a to-be-recognized signal, and generating a time-frequency diagram; inputting the time-frequency graph into the feature recognition model, and extracting multi-scale features from the time-frequency graph through a backbone network; sending a feature map with the highest semantic hierarchy in the multi-scale features into a convolution attention module, and sequentially executing channel attention weighting and space attention weighting in the convolution attention module to obtain an enhanced feature map; fusing the enhanced feature map and other scale features in a feature fusion layer to obtain a fused feature map; and based on the fused feature map, identifying a weak signal through a detection head. Compared with the prior art, the method has the advantage that the recognition accuracy of weak signals in communication signals is greatly improved.