553 results about "Text matching" patented technology

The invention relates to an intelligent diagnosis and treatment method and system based on a knowledge graph and a large language model. The method comprises the steps of constructing a basic framework of a medical field knowledge graph based on collected medical field data, constructing multi-level knowledge graph expression, obtaining symptom description of a patient through multiple rounds of dialogues, diagnosing the patient and generating a diagnosis report, rechecking and adjusting the diagnosis report by a professional physician, and finally outputting the diagnosis report. The intelligent diagnosis and treatment system comprises a knowledge graph construction module, an intelligent question and answer system, an entity matching module, an intelligent diagnosis module, a doctor review and diagnosis report adjustment module, a diagnosis report output module and the like. And main diseases are screened, so that the coverage is wider. And three levels of knowledge maps are constructed, so that text matching with finer granularity can be carried out. The non-standardized expression of the patient is firstly converted into standardized medical terms through a large model, and drug allergy and the like are taken into consideration, so that the diagnosis accuracy is improved, and the diagnosis and treatment risk is reduced.

The invention discloses a multi-modal entity and relation extraction method and system based on cross-modal alignment and fusion, and the method comprises the steps: carrying out the processing and coding of an input text and an image, and obtaining a plurality of types of image and text features; performing feature alignment on the fine-grained and coarse-grained text features and the pixel-level image representation by taking the semantic representation of the image as an anchor point, and mapping the image and the text features to the same semantic space; performing multi-granularity feature fusion through text-guided dynamic gating aggregation, visual prefix cross-modal fusion and cross-modal image-text matching, modeling association between noun phrases and image objects in a text while increasing feature complementarity, and obtaining multi-granularity multi-modal feature representation; fusing multi-granularity multi-modal features through entity guidance attention gating, and gathering visual information related to a text entity to obtain final multi-modal fusion representation; according to the multi-modal fusion representation, task prediction of multi-modal named entity recognition and multi-modal relation extraction is carried out.

The invention discloses a power transmission line insulator defect detection method and system based on machine vision, and relates to the technical field of machine vision, and the method comprises the steps: combining a pre-trained DenseCLIP model with a YOLOv11 network, extracting an insulator image feature through a visual encoder of the DenseCLIP model, converting the insulator image feature into a visual embedded vector, and carrying out the detection of the defect of the insulator through the YOLOv11 network; constructing a text prompt according to a defect category, converting the text prompt into a language embedded vector, calculating the similarity between the vision embedded vector and the language embedded vector, generating a pixel-text matching score graph to guide YOLOv11 network learning, and constructing an insulator defect detection model; and constructing a mixed training set by using an artificial sample and a real sample, carrying out training by adopting a transfer learning staged training strategy, importing a to-be-detected insulator image into an insulator defect detection model, obtaining an enhanced feature graph and a pixel-text score graph after visual and language embedding vector fusion, and generating a final prediction result. According to the invention, the detection precision and real-time performance of insulator defects are improved, and the intelligent operation and maintenance efficiency of a power grid is remarkably improved.

The invention discloses a voice recognition model optimization method and device based on reinforcement learning, equipment and a medium, and relates to the technical field of voice recognition, and the method comprises the steps: determining a current optimization strategy of a current time step; randomly selecting one piece of enhanced speech training data and corresponding real text training data from the data training set as target training data, and inputting the target training data into the current automatic speech recognition model for prediction to obtain a plurality of prediction texts output by the current automatic speech recognition model; determining a text matching degree and a semantic similarity between each prediction text and real text training data to obtain a group of target reward values, updating an optimization strategy at the current moment based on the target reward values, and skipping to execute the step of determining the current optimization strategy of the current time step to obtain a target reward value; and outputting the corresponding optimized automatic speech recognition model until the current automatic speech recognition model meets the optimization ending condition. And reasonable optimization of the automatic speech recognition model is realized.

The invention provides a multi-modal geographic positioning method based on a knowledge graph. The method comprises the following steps: graph construction: constructing a geographic knowledge graph; image positioning: calculating the entity similarity between the to-be-queried image and each sub-image so as to select an image positioning candidate sub-image, calculating the image matching similarity between the to-be-queried image and each image positioning candidate sub-image so as to determine an image positioning target sub-image, and taking the latitude and longitude coordinates of each image positioning target sub-image as a positioning result; text positioning: converting a to-be-queried text into map representation, calculating the map similarity between a text query sub-graph and each sub-graph to select a text positioning candidate sub-graph, and calculating the semantic similarity between the to-be-queried text and each text positioning candidate sub-graph, and calculating a text matching similarity corresponding to each text positioning candidate sub-graph based on the graph similarity and the semantic similarity corresponding to each text positioning candidate sub-graph so as to determine a text positioning target sub-graph, and taking latitude and longitude coordinates of each text positioning target sub-graph as a positioning result.

The invention discloses a transformer substation defect identification method and system based on multi-mode open set associative reasoning. The method comprises the steps that transformer substation oily equipment defect images are collected and marked; constructing a multi-modal data set of image-text matching; yOLOv8 is adopted to extract visual features, and a CLIP model is adopted to process text features; designing a re-parameterized visual language path aggregation network for feature fusion; training the model based on a contrast loss function; the applications are deployed after the performance is evaluated through multiple indexes; through the multi-modal feature fusion and the re-parameterization path aggregation network, the problem that the oil leakage defect is difficult to mark due to liquid flowability and form variability is effectively solved, the false detection condition under multi-device cross interference is remarkably improved, and by combining multi-scale feature learning and an attention mechanism, the oil leakage detection accuracy is improved. And the identification accuracy of the oil stain defect is greatly improved.

Disclosed in the present invention are a method and system for building a compilable customized module. The method comprises: customizing a module on the basis of a process, defining a modeling language syntax, and generating a modeling language file corresponding to the customized module; establishing a text matching rule and a grammar matching rule, using flex to generate a source program file of a lexical analyzer from the text matching rule, and using bison to generate a source program file of a syntax analyzer from the grammar matching rule; performing lexical analysis and generating an abstract syntax tree; traversing the syntax tree, generating a translation C++ file for a modeling language on the basis of a syntax-based translation configuration file, and compiling the translation C++ file; and importing the built customized module and an executable binary file into a simulation platform, and performing flow simulations for different scenarios including process simulation, optimization and parameter tuning. The present invention can realize the rapid commissioning of a process and can seamlessly switch a customized model to various application scenarios.

The invention belongs to the technical field of medical image recognition, and particularly discloses a myocardial infarction complication recognition method based on a heart multi-modal image, and the method comprises the steps: firstly obtaining multi-modal medical image data and corresponding diagnosis report text data; constructing a myocardial infarction complication identification model; dividing the whole network into a feature extraction stage and an image text feature matching stage by adopting a separation training thought, carrying out back propagation on the network in the feature extraction stage through L1 loss, and optimizing weight parameters of the feature extraction network; performing back propagation on the network in the image text matching stage by comparing loss, and optimizing parameters of an image and a text encoder; and finally, collecting a medical image of the heart of a patient, inputting the medical image into the trained myocardial infarction complication identification model, and generating a character diagnosis report of heart disease complication. Therefore, early automatic diagnosis of myocardial infarction complications is realized.

The invention discloses a graph representation learning and retrieval method and system based on a code warehouse graph structure, and the method comprises the steps: constructing a heterogeneous code graph related to an original code, a newly-added code or a modified code, carrying out the graph representation learning of the heterogeneous code graph, generating a node embedding vector, carrying out the self-supervised learning through an intelligent model, and carrying out the retrieval of the node embedding vector. And outputting the updated node embedding vector. The method comprises the steps of detecting whether newly-added codes or modified codes exist or not in real time, only performing heterogeneous graph modification on the newly-added or modified codes, receiving a query request from a user, generating corresponding code snippets according to the request, checking whether the target code snippets meet grammar specifications or not, and outputting a query result. According to the method, the structural relationship between codes is fully modeled, a structure perception retrieval mode is realized, the method is superior to a traditional text matching scheme, and the code semantic understanding capability is improved; and large-scale code graph online query is supported, and the applicability is wider.

The invention discloses a dance costume prop retrieval method based on causal reasoning and cross-modal matching, which comprises the following steps: in a matching task, calculating the similarity between costume prop images and text matching features by a model, and judging whether the costume prop images and the text matching features are matched or not according to a threshold value; and when the task is retrieved, calculating the similarity between the query text and the image feature set, and sorting to obtain a result. The process of calculating the features and the similarity comprises the following steps: firstly, extracting multi-scale visual features of an image by using Faster R-CNN, and extracting multi-granularity semantic features of a text by using BERT; then, the two types of features are subjected to self-attention and gating fusion respectively and then are input into a Transform layer, and final visual and text representation is obtained; then, multi-head attention calculation is carried out on the two representations, self-attention calculation is carried out after sequences are combined, and matching features are obtained; and finally, dividing the dot product of the two feature vectors by the modular length product to calculate the similarity. According to the invention, the matching accuracy of dance costume props can be improved.

The invention relates to the technical field of policy matching, and discloses an intelligent matching system based on big data and a matching method thereof. The system comprises a data acquisition module, a condition analysis module and an intelligent matching module. The data acquisition module is used for acquiring policy text, enterprise qualification and historical case data and constructing a matching basic database; the condition analysis module is used for extracting policy dominant conditions through a regular expression, mining implicit conditions from failure cases by utilizing an association rule algorithm, and calculating condition weights by adopting a random forest algorithm; the intelligent matching module integrates dominant and implicit conditions to calculate a matching coefficient, dynamically divides risk levels and generates an early warning, the matching comprehensiveness and the early warning sensitivity are improved by mining the implicit conditions, dynamically adjusting the matching model and reinforcing a learning optimization mechanism, upgrading from text matching to risk prejudgment is achieved, and the method has the advantages of being high in practicability and easy to popularize. The enterprise declaration success rate can be improved, and the repeated declaration rate is reduced.

The invention discloses an alarm information processing method and device and a storage medium, and relates to the technical field of data processing.The method comprises the steps that an unstructured alarm text is converted into an accurate keyword set through word segmentation processing, the limitation of traditional simple text matching is broken through, alarm core semantic information is effectively extracted, and redundant interference is reduced; according to the method, the mapping relation between the keyword and the alarm identifier is established and stored in the database, so that an efficient retrieval structure based on the inverted index is realized, the overhead of full-text scanning is avoided, and the retrieval speed of massive alarm data is remarkably improved; according to the method, the target alarm text can be more accurately positioned based on semantic relevance of the keywords, and the rigidity of traditional fixed condition retrieval is overcome, so that the practicability of alarm information retrieval is synchronously optimized in the aspects of efficiency and precision. The technical problem of low retrieval efficiency in a traditional retrieval mode is solved, and the technical effect of improving the retrieval speed of mass alarm data is achieved.

The invention belongs to the technical field of large model application, and particularly relates to a semantic enhancement text matching method and system based on an achievement transformation large model. The semantic enhancement text matching method based on the achievement transformation large model comprises five parts, namely preparation of data and knowledge, enhancement of multi-level semantic knowledge, learning of semantic difference, intelligent weight distribution of semantic information and prediction of text similarity. The preparation of data and knowledge is used for constructing basic data and knowledge resources required by a text matching task through text preprocessing and knowledge base construction. And enhancement of multi-level semantic knowledge: enriching semantic representation of the text by injecting external knowledge at three levels of word level, phrase level and sentence pattern level. And semantic difference learning: learning fine-grained semantic features by comparing a learning framework, and improving the recognition capability of the model on the text semantic difference. Intelligent weight distribution of semantic information is carried out, and dynamic balance and optimal combination of different source information are realized by evaluating feature importance and dynamically adjusting weights. And predicting the text similarity, calculating the similarity score of the text pair by integrating the feature information of each stage, and providing an interpretable prediction result. The complete text matching method and system are designed based on the achievement transformation large model, the problems of insufficient knowledge utilization, incomplete semantic representation, fixed feature weight and the like of the large model in a text matching task are effectively solved, the text matching effect of the large model in an actual application scene is improved, and the text matching efficiency is improved. And a feasible solution is provided for achievement transformation of a large model technology.

The invention relates to an AIS data-based port ship unloading area analysis method. The method comprises the following steps of: 1, preprocessing AIS data, namely carrying out cleaning screening and attribute supplementation on original data; 2, ship staying leg extraction: defining a staying leg by using motion features and expressing the staying leg by using a Geohash grid; step 3, port destination text matching: matching irregular destinations with standard port names; 4, bidirectional destination retrieval: analyzing an accurate destination by using a ship trajectory, and determining an affiliation port of a staying grid; and 5, grid integration and route cutting: integrating grids belonging to different ports to obtain an unloading area, and cutting a whole-cycle route of the ship. According to the method, the problems of complex unloading mode and complex route staggering around a port are solved, and reliable technical support is provided for trade monitoring of bulk energy goods such as crude oil by utilizing the complementary advantages of time-space information and destination texts in AIS data.

The invention discloses a multi-modal data processing method and device based on a large model, equipment and a medium. Performing feature extraction and association mining on the multi-modal data to obtain multi-modal data features; according to a retrieval sequence of the multiple text matching methods, sequentially using the multiple text matching methods to perform text retrieval on the multi-modal data features in a pre-constructed knowledge base to obtain a first associated sub-graph; based on the semantic similarity between the multi-modal data features and multi-modal data in a pre-constructed knowledge base and a similarity threshold value, determining a second associated sub-graph; performing standard evaluation on the plurality of candidate results in the first associated sub-graph and the second associated sub-graph, and sorting the plurality of candidate results according to an evaluation result to obtain a retrieval result; inputting the retrieval result and the multi-modal data features into a large model for information analysis to obtain prompt information and a target tool; and sending the prompt information and the target tool to the terminal, and receiving a data processing result of the terminal.

The invention discloses a large language model construction method for the vertical field of forestry, which belongs to the technical field of computer models and comprises the following steps: constructing a first data set of forestry management business questions and answers; constructing a multi-modal coding model; freezing a text stream of the multi-modal coding model, forming the multi-modal coding model and an LLM decoder into a large language model, and training the large language model for generating a target text related to user input during intelligent question answering; constructing an RAG knowledge base; and performing intelligent question and answer based on RAG retrieval. Aiming at the characteristic that forestry management business images and texts have high matching, innovative technologies of multi-modal joint coding, weighted similarity calculation and LLM training by a frozen text encoder are introduced, and the quality and application value of forestry intelligent question answering and document generation are greatly improved. And the problem that the existing image and text matching precision is insufficient is solved, and more flexible and accurate retrieval and intelligent question answering can be realized in combination with an RAG technology.

The invention relates to a weak supervision instance segmentation method and device, computer equipment and a storage medium, and the method comprises the steps: obtaining a training image set and an image-level label, and generating a pseudo-semantic segmentation label; extracting multi-scale image features in the training image set, performing prompt contrast learning based on the multi-scale image features and the pseudo-semantic segmentation labels, and generating contrast loss and a pixel-text matching graph; splicing the multi-scale image features and the pixel-text matching graph to generate enhanced features, and training an instance positioning module based on the enhanced features to generate an instance center point and positioning loss; generating a semantic mapping graph through the segmentation network, and calculating self-matching loss of prompt guidance; calculating the total loss of the model based on the obtained loss, and performing iterative training on the instance segmentation model based on the total loss of the model to generate a target instance segmentation model; and generating a target instance segmentation result based on the to-be-processed image and the category text through the target instance segmentation model. According to the invention, the precision of instance segmentation is improved.

The invention provides a name disambiguation method based on multi-relation deep retrieval text matching, and relates to the technical field of name disambiguation. Cross-modal data related to enterprises and persons are obtained, data alignment and data fusion are conducted on the cross-modal data through an entity alignment algorithm, and a person-enterprise structured data set is formed; a multi-relation deep retrieval model is established based on a pre-training language model to generate a semantic vector, semantic similarity is calculated according to the semantic vector, a character embedding vector is generated according to co-occurrence frequency and time-space correlation characteristics calculated by a human-enterprise structured data set, and a disambiguation recognition model is established based on an adversarial neural network. Performing name disambiguation according to the character embedding vector, establishing a structure updating model through a graph attention network, updating a graph structure in real time, calculating semantic similarity and the confidence coefficient of the co-occurrence frequency according to the Bayesian theorem, and updating the weight of the semantic similarity and the weight of the co-occurrence frequency according to the confidence coefficient.

The invention provides a large model-based long text matching method and system, a terminal and a medium. The method comprises the following steps of: extracting first key semantic information of a first long text by utilizing a first large model; performing matching similarity calculation with the second long text based on the first key semantic information; wherein matching similarity calculation is carried out based on the first key semantic information and the second long text, and the method further comprises any one of the following steps: dividing the first key semantic information into corresponding keywords, and carrying out matching similarity calculation on the keywords and the second long text; extracting second key semantic information of the second long text by utilizing the first large model; and performing matching similarity calculation on the first key semantic information and the second key semantic information. According to the method, dimension reduction of the long text is assisted through the large model, the problem of text processing is firstly solved, key information in the original long text can be well reserved, and meanwhile the problem that a traditional method occupies high video memory is solved.

The invention discloses a power transmission and transformation project starting preparation progress recognition method and system based on an improved GCN, and the method comprises the steps: firstly collecting starting procedure texts and construction site images, extracting text features through a graph convolution network, carrying out the classification recognition, combining image-text features, calculating a similarity score through an image-text matching model, and carrying out the recognition of the starting preparation progress of a power transmission and transformation project based on the similarity score. Constructing a fusion loss function and an original loss function, and alternately updating the gradient to optimize the model; and finally, integrating text classification and image-text matching results, and jointly evaluating the start-up progress state. According to the method, the graph convolutional network is combined with the graphic and text information, the accuracy of engineering work-starting progress state recognition is effectively improved, the graphic and text features are captured by aggregating node neighborhood information, the adaptability of the model to sparse data is enhanced, the accuracy of embedded features is improved, the gradient is alternately updated by fusing the loss function and the original loss function, and the recognition accuracy is improved. And an image-text matching model is optimized, so that the model has higher expressive power and comprehensiveness, and the accuracy of progress state recognition is improved.

The invention relates to an enterprise industry chain identification method and system based on a big language model, and the method comprises the steps: obtaining target enterprise information based on a big data analysis technology; dividing a plurality of relatively independent industry chain links for the industry to be analyzed; for each divided industrial chain link, obtaining a descriptive text; based on text matching, labeling the target enterprise information with an affiliated industry chain link label to obtain enterprise industry chain link affiliation labeling information; constructing a training data set and a verification data set based on the descriptive text of each industry chain link, enterprise industry chain link affiliation labeling information and task cues determined according to task requirements and a model optimization strategy; and inputting the training data set and the verification data set into a generative large language model for model training to obtain an exclusive large language model for enterprise industry chain identification. The implementation of the method can realize accurate identification of the enterprise industry chain.

The invention relates to a cross-modal retrieval technology, in particular to a cross-modal image-text pedestrian retrieval method based on a generative model, which comprises the following steps: acquiring text description information and a corresponding original pedestrian image; calling a diffusion generation model to generate an intermediate image based on the text description; original image features and text features are extracted through an image encoder and a text encoder respectively, and semantic-fused text feature representation is constructed; calculating the similarity between the image feature representation and the text feature representation to obtain an image-text matching score; introducing and generating an intermediate image as a semantic bridge, and realizing multi-modal fusion among the image, the text and the intermediate image through a cross attention mechanism to obtain fused image and text features; and using the fusion features to train a recognition model, and finally outputting a pedestrian image retrieval result corresponding to the text description. According to the method, the image-text alignment precision can be improved, the robustness of a scene with incomplete text description is enhanced, and the recognition accuracy in a cross-modal retrieval task is improved.

The invention provides a maintenance guidance method and device for a household electrical appliance. The method comprises the following steps: constructing a unified vector database in advance based on text data, image data and audio data related to a household electrical appliance fault; acquiring field condition information of the household electrical appliance, wherein the field condition information comprises video information, audio information and / or text information; respectively carrying out image matching, audio matching and / or text matching on the field condition information based on the unified vector database, and outputting knowledge fragments as retrieval results; and inputting a retrieval result into the large language model, and outputting a maintenance guidance scheme of the household electrical appliance. According to the method, a retrieval and reasoning framework based on multi-modal feature direct fusion can be constructed, information dimension reduction loss is avoided by directly analyzing image, audio and text features of faults, and the comprehensive accuracy of fault diagnosis is improved to 95% or above; and through a dynamic knowledge retrieval and enhanced generation technology, the risk of'knowledge illusion 'possibly occurring in the professional field of the general large language model is effectively reduced.

The invention discloses a content security identification method based on integration of multiple large language models, and belongs to the technical field of data processing. The method comprises the steps of receiving a query text, carrying out word segmentation processing on the query text, generating query text segmented words, obtaining a sensitive data text matched with the query text segmented words in a sensitive database, splicing the query text and the sensitive data text based on a preset prompt word template, generating security identification prompt words of the query text, and sending the security identification prompt words to the sensitive database. And inputting the security identification cue word into the at least two large language models, obtaining output results of the large language models, integrating the output results, and determining a security identification result of the query text. According to the method, the sensitive data text is recalled through the sensitive database, the safe recognition cue words of the large language models are enhanced, and the accuracy of sensitive word recognition is improved through integrated processing of the multiple large language models.

The invention provides a networked automobile risk scene analysis method based on a multi-modal corpus and related equipment, and the method comprises the steps: receiving the multi-modal corpus containing various types of test data, and carrying out the format standardization and time synchronization processing of the multi-modal corpus; performing risk scene analysis on the basis of association information among different types of test data in the multi-modal corpus, identifying abnormal behaviors occurring in the test process, and generating corresponding structured event description information; and extracting key data fragments and image frames associated with the abnormal behaviors, carrying out image-text association on the key data fragments, the image frames and the event description information, and generating a test report based on a preset template. According to the method, a multi-modal corpus analysis mechanism is introduced, risk scene analysis and abnormal behavior recognition are executed, and image-text matching and report generation are automatically completed. The manual arrangement burden is reduced, and the intelligent degree of risk scene identification and the automation level of test report generation are improved.

The invention discloses an intelligent engineering quantity list matching method, which comprises the following steps of: firstly, performing deep analysis on project description to form a key feature dictionary and form a basis for subsequent accurate matching; then, adopting a two-stage process of coarse screening and fine arrangement: in the coarse screening stage, combining project codes pointing to professional classification with key features capable of quickly locking core elements, performing efficient filtering on a huge standard quota library, and quickly generating a candidate matching list which is controllable in scale and highly related; and in a fine ranking stage, turning to more detailed semantic and feature dual verification, comprehensively utilizing the deep semantic similarity of a complete item description text and an accurate comparison result of a key feature dictionary, and performing fusion calculation to obtain a final score so as to realize accurate ranking of candidate lists. In this way, the semantic gap caused by synonyms, inverted word order, non-standard expression and the like in traditional text matching can be effectively overcome, and the matching accuracy is greatly improved.

The invention relates to the technical field of image processing, solves the problem that in the prior art, a video collection matched with an advertisement target cannot be accurately screened and generated according to different scene types, and provides an advertisement video collection generation method and device, equipment and a storage medium. The method comprises the steps that the scene type of an advertisement to be put is acquired, and the scene type comprises an introduction scene, a product display scene, a user experience scene and a problem solving scene; according to the scene type, utilizing a preset text matching algorithm to carry out similarity evaluation on character information related to the video clip content and a preset text template, and determining a target similarity corresponding to each video clip; and according to the target similarity and a preset similarity threshold value, synthesizing the video clips of which the target similarity is greater than the similarity threshold value into a video collection. According to the method, the video collection which better conforms to the advertisement target and can accurately convey the advertisement information is generated, and the advertisement effect is improved.

A computer-implemented method is provided. The method includes: obtaining a sample text and a sample image corresponding to the sample text; labeling a true semantic tag for the sample text according to a first preset rule; obtaining a text feature representation of the sample text and a predicted semantic tag output by a text coding sub-model; obtaining an image feature representation of the sample image output by an image coding sub-model; calculating a first loss based on the true semantic tag and the predicted semantic tag; calculating a contrast loss based on the text feature representation of the sample text and the image feature representation of the sample image; adjusting parameters of the text coding sub-model based on the first loss and the contrast loss; and adjusting parameters of the image coding sub-model based on the contrast loss.

The invention discloses an unsupervised text-to-image pedestrian re-identification method and system. The method comprises the following steps: constructing an image-text matching relationship; clustering operation is carried out on the image features, and false identity tags are distributed to the pedestrian images and the corresponding text descriptions; constructing an image memory dictionary; obtaining a similarity set according to the image memory dictionary; removing pedestrian images with abnormal similarity and corresponding text descriptions according to the similarity set; respectively acquiring image features and text features of pedestrian images and text descriptions, and constructing a multi-stage ternary joint learning process; and obtaining text features to be re-recognized, and performing pedestrian image matching to obtain a pedestrian recognition result. According to the method, the intra-class compactness and inter-class separability of the samples are enhanced, and the retrieval precision and efficiency of the unsupervised text-to-image pedestrian re-recognition method are greatly improved.

The present disclosure provides a method and apparatus for generating a digital human video based on a large model, an intelligent agent, an electronic device, and a storage medium, which relate to a field of artificial intelligence technologies, in particular to technical fields such as deep learning, large models, computer vision and so on, and may be applied to scenarios such as video livestreaming, advertisement production, and e-commerce sales. The method for generating a digital human video based on a large model includes: acquiring a requirement information, where the requirement information includes an action description information for describing a specified action video segment, and the action video segment represents a specified action of a target object; processing the requirement information using a first large model to obtain a target script, where the target script includes a target speech segment text matching the action description information; and processing the target script and the action video segment using a second large model to obtain a target video for displaying a target digital human performing a speech delivery based on the target speech segment text while performing the specified action.