279 results about "Search graph" patented technology

The invention discloses an autoregressive visual target tracking algorithm based on token fusion, and the algorithm specifically comprises the following steps: 1, obtaining a public visual target tracking data set, and converting the data set into a central coordinate format; 2, preprocessing an initial template of a given video sequence, an online template at the current moment and a search image to obtain visual embedding of the initial template, a dynamic template and the search image; the method comprises the following steps: preprocessing an angular coordinate sequence of a target frame in previous historical N frames of a current search image of an input video to obtain historical track word embedding of the current search image; 3, constructing an autoregressive visual target tracking model based on token fusion, and designing a loss function; step 4, obtaining a final autoregressive visual target tracking model based on token fusion; the method solves the problem of insufficient model feature extraction capability caused by insufficient utilization of spatio-temporal information and redundant background information of a search area in an existing method.

The invention provides an RGBT target tracking method based on domain adaptation and spatio-temporal information fusion. The problem that an existing target tracking scheme is poor in tracking effect under complex scenes such as target appearance change and shielding is mainly solved. Comprising the following steps: 1) constructing an RGBT data sample set; 2) constructing a dual-branch RGBT target tracking model based on a pre-trained RGB tracker and training the dual-branch RGBT target tracking model; 3) designing an online module updating strategy for dynamically updating the initial template image; 4) taking the initial template image and the search image as the input of the trained model, and taking the dynamic template image as the additional input of the model; and 5) loading the optimal parameter, predicting the motion state of the target in a subsequent frame by using the trained model, and obtaining and outputting a target tracking result. According to the method, the calculation cost is greatly reduced, and the time information of the target can be fully utilized, so that the performance of the feature extraction network is effectively improved, and the accuracy and robustness of target tracking are improved.

The invention relates to a single-target tracking method based on global perception and lightweight self-attention, and the method is characterized in that the method comprises the steps: respectively inputting an obtained template image and a search image into a weight sharing feature extraction module, and extracting a template feature map and a search feature map; performing attention enhancement of channel and space dimensions on the extracted feature map through a joint global perception module, and outputting an enhanced template feature map and an enhanced search feature map; an improved Transform module is adopted to carry out local feature enhancement and lightweight self-attention calculation on the enhanced feature map, and reconstructed template features and search features are obtained; and inputting the reconstructed template features and search features into the trained target model predictor to realize target tracking and positioning. Compared with the prior art, the method has the advantages of realizing balance between target tracking precision and real-time performance and the like.

Examples described herein relate to techniques for facilitating selection of stickers for inclusion in messages within the context of an interaction system. According to some examples, message content is detected and a set of candidate stickers is identified based on the message content. A search icon is dynamically replaced with a representation of respective ones of the set of candidate stickers. At a first point in time, the search icon represents a first candidate sticker of the set of candidate stickers. At a second point in time, the search icon represents a second candidate sticker of the set of candidate stickers.

The invention provides a target tracking method and system based on global and local two-way extraction and asynchronous enhancement, and the method comprises the steps: carrying out the initialization of a template image and a search image, and carrying out the sequence division through block embedding; extracting global features and local features of the template image and the search image at the same time, and performing feature fusion on the global features and the local features of the template image and the search image; the output of the global and local joint module is input into an asynchronous enhancement module for asynchronous interaction and feature enhancement; and inputting the finally enhanced output feature to a prediction head to obtain a tracking result. According to the method, global and local features are extracted at the same time through the feature combination module, and the features after interaction are enhanced by using asynchronous enhancement operation in the feature fusion stage, so that the robustness and the feature expression ability of the model are improved.

The invention relates to a hyperspectral image classification method based on a meta-pseudo-label multi-scale image convolution network, and the method comprises the steps: building a meta-pseudo-label multi-scale feature search image convolution network (MPL-MFSGCN), dividing the HSI into superpixel images of different scales through employing an SLIC superpixel segmentation technology, and converting the superpixel images into image structures; secondly, constructing a multi-scale feature search graph convolutional network MFSGCN architecture, optimizing a superpixel feature weight combination by using a neural network search algorithm, and enhancing the discrimination of feature expression; a semi-supervised teacher-student model sharing the same graph convolutional network is adopted to generate meta-pseudo tags, and a student network is finely adjusted to improve the robustness of the network; the key features can be automatically captured under the small sample condition, the classification precision higher than that of an advanced method is obtained compared with a common HSI data set, and the classification accuracy is effectively improved.

The invention discloses a multi-modal target tracking method guided by sparse space-time markers, which performs joint feature extraction, modal fusion and template search matching through a network model, and introduces sparse space-time markers at multiple levels to embed target appearance change information to complete accurate target tracking. In the first stage, search area images and template images of two modalities are cut according to a target frame of an initial frame, then the template search images are coded and projected to a sequence feature space, templates and search sequence features of the two modalities are spliced together, and the spliced templates and search sequence features are sent to a stacked sequence feature coding layer; in the second stage, sparse space-time marks from historical frames are introduced into a preset coding layer, and cross attention is carried out on the sparse space-time marks and search areas of the modes of the sparse space-time marks so as to embed target change information; search mark features of two modals extracted through multilayer coding are spliced together and sent to a tracking prediction head to obtain a response point coordinate with the highest score and a value of a regression frame, and finally a tracking result coordinate and a bounding box of an actual target are obtained through relative transformation.

The invention relates to a visual target tracking method and device based on efficient sequence generation and electronic equipment, and the method comprises the steps: carrying out the processing of an obtained search image, a template image and a randomly initialized tracking mark, and obtaining search visual embedding and template visual embedding; the method comprises the following steps: performing position coding after search, template visual embedding and tracking mark connection, then coding a position coding result by adopting an encoder based on a visual transformer framework, and processing a tracking mark coding feature by adopting a tracking prediction head to obtain a target tracking prediction result; if the target tracking prediction result meets a preset condition, taking the target tracking prediction result as a target tracking result; otherwise, decoding the coding structure by adopting a transformer decoder; and processing by adopting a tracking prediction head according to a decoding result to obtain a target tracking result. By adopting the method, the forward reasoning speed is increased, and the tracking speed is increased.

The invention discloses a knowledge graph multi-hop retrieval method fusing semantic analysis and graph embedding, provides an innovative fusion strategy, and aims to improve the efficiency and accuracy of knowledge graph multi-hop questions and answers by integrating semantic analysis and graph embedding technologies, and effective multi-hop questions and answers should be a knowledge graph multi-hop retrieval method based on question semantic features. In the process of searching an entity closest to the meaning of a question in a graph embedding space in a relational semantic path, relational semantics is used for guiding path exploration, the number of jumps is dynamically determined according to semantic content of a question, probability evaluation between the question and the entity in the graph embedding space is introduced, so that the correlation and reliability of candidate answers are enhanced, and the probability evaluation efficiency is improved. The specially designed loss function gives consideration to the exploration efficiency of the semantic path and the accuracy of the answer, thereby promoting the model to continuously optimize the reasoning ability in the training process.

The invention discloses a visual target tracking technology based on a local fine-grained target. The visual target tracking technology comprises a visual target tracking system, an establishment method of the visual target tracking system and a visual target tracking method. The visual target tracking system excavates deep features between a template image containing a target to be tracked and a search image, and performs target tracking through classification and regression. The template image is input into a hierarchical self-adaptive mask mechanism module for self-adaptive local fine-grained mask processing, and a template mask image is obtained; the template image feature extraction network and the search image feature extraction network extract deep features of a template image and a search image respectively, deep fusion and relation modeling are performed on the features through the Transform encoder and the Transform decoder, and then prediction of a target bounding box in the search image is performed through classification and regression. The system can track a fine-grained target under the condition that the structure priori is not clear.

The invention provides a target tracking method and system based on a Mama visual hybrid module, and the method comprises the steps: inputting a search image and a template image into the Mama visual hybrid module, and extracting and superposing global context information and local detail information; inputting the output of the Mama visual mixing module into a feature fusion module for fusion; performing feature enhancement on the output of the feature fusion module by using a feature enhancement module to obtain enhanced features; pre-training the optimized tracking model by using a large-scale data set, and updating parameters of the tracking model; and inputting the search image and the template image into the parameter-updated tracking model to obtain a tracking result. According to the method, a feature aggregation network is designed, historical state information of a search area is integrated by using an image state space module, the historical state information and interacted features are spliced along a channel direction, and effective aggregation of the features is realized through convolution, activation, normalization and jump connection to filter irrelevant information.

The invention discloses a data search processing method and a host. A host processor obtains a query vector and a centroid of each to-be-searched graph; the to-be-searched graph is constructed for the query scene, so that the number of candidate vectors of the corresponding to-be-searched graph is determined according to the query vector and the centroid; a query request is sent to the hardware accelerator, the query request comprises query vectors and the number of candidate vectors, the hardware accelerator distributes the query request to different search units, and candidate vector search is conducted on the corresponding graphs to be searched in the external storage through the different search units.

A search data processing method includes: obtaining a media search graph including a query node, a media node, and an association node; obtaining first training sample pairs, including a positive node pair and a negative node pair, from the media search graph, sampling the media search graph based on meta-paths respectively corresponding to the query and the media data, to obtain sampling sub-graphs; inputting the sampling sub-graphs to an initial graph neural network to obtain respective initial semantic features and form a semantic feature pair corresponding to the first training sample pair; and training the initial graph neural network based on a difference between semantic feature pairs of the positive node pair and the negative node pair, to obtain a target graph neural network that is configured to determine a target semantic feature corresponding to a query node or a media node.

The invention provides a construction method of a generative heterogeneous graph model, an agent distribution method and a related device, and relates to the technical field of artificial intelligence such as generative models, generative search, graph models, agents and object distribution. The method comprises the following steps: extracting a first interaction action chain containing agent distribution behaviors exceeding a preset satisfaction degree from historical search behaviors and search behavior feedback of a first sample user; extracting a second interaction action chain containing agent distribution behaviors exceeding the preset satisfaction from historical recommended behaviors and recommendation behavior feedback of a second sample user; and on the basis of the first interaction action chain and the second interaction action chain, by taking the user, the behavior object and the agent as different nodes and taking the interaction action as an edge for connecting the corresponding nodes, constructing a generative heterogeneous graph model. The method comprises the following steps: acquiring user related information of a target user; inputting the user related information into a generative heterogeneous graph model obtained by pre-training to obtain a target agent output by the generative heterogeneous graph model; and distributing the target agent to the target user.

The embodiment of the invention provides a human resource matching system based on big data, and the system comprises a construction module which is used for carrying out the entity recognition of a target text block through a rule base, and obtaining an initial entity; performing entity recognition on the target text block by using a first machine learning model, and adjusting the recognition process of the first machine learning model by using the initial entity to obtain a target entity; constructing a graph database according to the target entity; the target text block comprises a text block obtained by partitioning the job application text and a text block obtained by partitioning the recruitment text. And the search module is used for generating search graph data according to the user input operation and the graph data corresponding to the user. And the matching module is used for calculating the similarity between the search graph data and the graph data in the graph database, and selecting matched graph data from the graph database according to the similarity. In this way, fusion of entity recognition based on rules and in combination with context information is achieved, and the accuracy of entity recognition of a machine learning model can be improved.

Disclosed are a method and an electronic device for image search. The method includes: obtaining a search string provided to search for images; determining whether the search string matches one of the classification tags, and while the search string matches one of the classification tags, to-be-searched images corresponding to the one of the classification tags are presented in a user interface; while the search string does not match the classification tags, obtaining a second comparison vector according to the search string based on a multi-modal artificial intelligence (AI) model, determining a correlation degree between the second comparison vector and the first comparison vector corresponding to the to-be-searched images to generate a search result; and, identifying whether the search string has a second location information, so that the search results generated based on the search results and the second location information are presented in the user interface.

The present application discloses a method, device and medium for hot word enhancement for speech recognition. The method includes: obtaining an audio file of the hot word through an encoder, extracting features based on the audio file, sending the extracted features to a CTC decoder, and obtaining a streaming recognition result through the CTC decoder; inputting the streaming recognition result into a language model for shallow fusion, so as to bias the recognition result according to a prefix tree to obtain a search graph; performing shallow fusion through WFST to obtain the best path according to the search graph, sending the best path to an attention decoder, and obtaining accurate results through an attention decoder to complete the enhancement of the hot word. The present application improves the recognition accuracy of out-of-domain (OOD) audio through a hot word enhancement method that combines shallow fusion of WFST, deep bias based on a prefix tree and a language model. The recognition accuracy of hot words is significantly improved.

The invention provides a target tracking method and system based on multi-kernel convolution and grouping residual self-attention, and the method comprises the steps: constructing a feature fusion network which comprises a multi-kernel convolution module and a grouping residual self-attention module under a twin network framework, and extracting the features of a template image and a search image, the template features are used as static tokens to be connected with search features in series, fusion is carried out through a self-attention mechanism, and a fusion feature sequence is formed; calculating a maximum similarity score by using a multi-head self-attention encoder to generate a center score graph, and planarizing a search region in the fusion feature sequence into a 2D feature graph; and inputting the data into the head network for processing to obtain a tracking result. According to the method, the advantages of deep convolution and a self-attention mechanism are absorbed, the global modeling capability is achieved, meanwhile, the capability of fully learning local details is achieved, feature fusion is conducted through an asynchronous structure, and the robustness of the model is enhanced.

The invention discloses a visual target tracking method based on natural language and target state information. The method comprises the following steps: (1) constructing a training sample set; (2) constructing a visual target tracking model based on a natural language and target state information; step (3), adjusting parameters of an image-text encoder and loading a pre-training weight to obtain a feature after the text and the first template are fused, a feature of the second template and a feature of the search image; (4) fusing the position information of the target in the sample set and the bounding box information of the target into the features of the second template; step (5), obtaining features after joint modeling; step (6), acquiring a token containing target position information after query; (7) obtaining a predicted target bounding box regression result; and (8) obtaining a final tracking result. According to the invention, the tracking accuracy of the visual tracker based on the natural language is effectively improved.

Method and system for identifying design templates that match an input query are disclosed. The method includes: receiving a design search query; performing a template search based on the design search query, the template search returning a first template design, the first template design including a target image; processing the design search query to generate an image search query; performing an image search based on the image search query, the image search returning a candidate image; and generating a new design. The new design is based on the first template design and includes the candidate image instead of the target image.

The invention discloses a network fault prediction method, and relates to the technical field of fault prediction. The method comprises the following steps: constructing a cross-level semantic topology model; screening a key link by using a node importance algorithm and extracting a structural fingerprint vector; calculating the deviation degree between the current link and the structure fingerprint vector, and judging that the structure is abnormal if the deviation degree is smaller than a threshold value; constructing an anomaly propagation rule base according to historical faults and operation and maintenance data; when a plurality of structure anomalies appear at the same time, the conflict resolution algorithm determines an anomaly dominant factor; potential signals are collected according to the abnormal dominant factors, weights are given according to a rule base, and weighting is carried out to obtain comprehensive potential signals; based on this, a fault propagation path is constructed, and a path weight is calculated by using a Pearch ranking algorithm; a fault chain is rehearsed through a depth-first search graph traversal algorithm; and verifying the rehearsal chain by using a preset test statistic circumstantial evidence to obtain a fault prediction result. According to the invention, on the basis of the fault prediction result, cross-level network fault prediction in network propagation is realized.

Embodiments described herein provide systems and methods for text generation using a natural language processing model. Natural language input is received and encoded by an encoder. A decoder generates a plurality of next-node probabilities for a set of candidate nodes that are next to at least one of K previously decoded paths of nodes on a directed search graph. Scores are computed for each of the candidate nodes and done in parallel among the K previously decoded paths. The highest scoring K nodes are selected for addition to the directed search graph. The search continues until a search budget is exhausted. Generated sequences are output. These sequences may represent a plurality of sentences or questions, depending on the task.

The invention discloses a generative machine vision matching method, and belongs to the technical field of vision matching, and the method comprises the steps: calculating the matching degree of a top-layer search image based on the node direction of a pixel node of a top-layer template image and the gradient direction of the pixel node of the top-layer search image, and reserving a result of which the matching degree is greater than a user input threshold value; on the basis of the matching result of the top-layer search image, calculating the matching result of the next-layer search image layer by layer downwards until the calculation of the matching results of all sampling layers of the search image is completed; and based on matching results of all sampling layers of the search image, drawing rectangles in combination with the width and the height of the template, and determining a final matching result of visual matching according to the overlapping rate of the rectangles. According to the method and the device, a user can be supported to draw a template through a mouse or generate templates in other shapes in a self-defined manner for visual matching under the condition that a real template picture does not need to be provided.

The invention provides a target tracker training method and a target tracking method, and the training method comprises the steps: training a neural network model based on a plurality of template and search image pairs with respective labels, each template and search image pair comprising a target template frame and a search frame, the labels are used for representing whether each template is matched with a target template frame and a search frame in a search image pair or not and representing a real target position, and the neural network model comprises an encoder, a decoder and a prototype memory network; and determining the trained neural network model as a target tracker used for outputting a target tracking result corresponding to a template and a search frame in the search image pair, wherein the target tracking result comprises a confidence coefficient used for representing a matching result and a target position prediction result under the matching condition. According to the method, accurate target state reasoning can be carried out by utilizing positioning uncertainty, and reliable target appearance matching and target tracking results can still be ensured under the condition that the appearance of the target changes drastically.

The invention provides a target tracking method and system based on a double-domain attention and wavelet enhancement module, and the method comprises the steps: carrying out the preprocessing of a template image and a search image, and inputting a parameter adjustment feature extraction network, respectively extracting features of the template image and the search image in a spatial domain and a channel domain through double-domain attention of a feature extraction network; splicing the template image sequence features and the search image sequence features along channel dimensions, adjusting a self-attention module and a wavelet enhancement module of a feature fusion network by using parameters after splicing, and performing interactive fusion and feature enhancement; and sending the fusion feature into a prediction head to obtain a tracking result. According to the method, the spatial domain features and the channel domain features of the input image sequence are extracted through the double-domain attention, so that the model not only can pay attention to the spatial position of a target, but also can model the importance of channel semantics, double-domain complementary modeling is realized, a target region is positioned more accurately, and target features and noise channels are effectively distinguished.

Disclosed is a network and method for creating and searching a graph in a massively parallel manner. The network includes a plurality of graph storage instances that collectively store a graph formed from a plurality of entities and comprising a vertex for each entity, edges connecting pairs of vertices, and adjacency relations. Each graph storage instance stores a subgraph of the graph and comprises a partition with a non-overlapping vertex set of all the vertices in the respective subgraph, an edge set of all the edges between vertices in the respective vertex set, and the adjacency relations for the edge set and to any edges outside the partition, to which the edge set connects. Each graph storage instance also includes an executor for executing one or more of the actions over the respective subgraph, and an action monitor for supplying actions to the respective executor.

1. The name of the design product: content search module of content display dynamic graphical user interface of electronic device. 2. The use of the design product: an electronic device. 3. The design points of the design product: the claimed part of the graphical user interface. 4. The picture or photo that best shows the design points: the front view. 5. The use of the graphical user interface: for displaying content; the use of the claimed part is to search for content. 6. The human-computer interaction mode of the graphical user interface: in the content search mode, in the order from the front view to the change state view 3, as the user taps, clicks or touches a certain area in the oval search bar at the top of the interface, the content window below the search bar gradually shrinks and blurs, and at the same time the search bar itself moves to the center of the interface and gradually enlarges. 7. Other circumstances that need to be explained: the part depicted by the solid line in the view is the claimed part of the design.

1. Name of the product in this design: Intelligent Question-Answering Search Graphical User Interface for Electronic Devices. 2. Intended use of this design: for use in an electronic device. 3. The key design feature of this product is its graphical user interface. 4. The picture or photo that best illustrates the key design points: Design 1 front view. 5. Design 1 is designated as the basic design. 6. Uses of the graphical user interface: for AI teaching assistant intelligent question answering and search interaction. 7. Description of the changing states of the graphical user interface: In the main view of Design 1, click on any covered area below "Referencing XX materials as references" in the middle of the interface to enter the changing state diagram of Design 1 and view the video content; the interaction method of Design 2 is the same as that of Design 1, so the description is omitted. 8. Other situations requiring explanation: Other explanation: The covered area is the content screen; X represents text content.

The large-scale image search method based on deep hash comprises the following steps: acquiring a query image uploaded by a user and preprocessing the query image to obtain a to-be-searched image; the to-be-searched image is input into a pre-trained image Hash coding model, the image Hash coding model comprises a feature extraction network and a Hash mapping network, and the feature extraction network is constructed based on Vision Transform and is used for carrying out feature extraction on the to-be-searched image to obtain image high-dimensional features; the Hash mapping network is used for performing Hash coding on the image high-dimensional features to obtain search Hash codes; calculating the similarity between the search hash code and the in-library hash code of the inventory image in the image library based on the Hamming distance; and screening the inventory images based on the similarity to obtain a target image and feeding back the target image to the user. The invention provides a deep hash-based large-scale image search method, an image recommendation method, a recommendation system and computer equipment, which have better search performance.