120 results about "Object structure" patented technology

The invention discloses an indoor real scene three-dimensional reconstruction method based on improved 3D Gaussian sputtering, belongs to the technical field of three-dimensional reconstruction, and aims to complement initial point clouds in a self-adaptive manner through a designed feed-forward point cloud densification network, generate a high-density 3D Gaussian radiation field, fill scene missing details and avoid the problem of precision attenuation caused by lack of point cloud constraints. In the cost body optimization stage, a depth regularization loss module is designed by integrating global structure features, depth information is integrated into the optimization process, algorithm reasoning is accelerated, and object structure fuzziness is relieved; and finally, in combination with adaptive density control and Gaussian pruning, radiation field parameters and density distribution are finely adjusted, indoor scene reconstruction with double improvement of visual fidelity and structural integrity is realized, interactive application becomes possible, a real-time solution is provided for indoor scene reconstruction, and application of a three-dimensional reconstruction technology is widened.

There are provided systems and methods for configuration-driven efficient transformation of formats and object structures for data specifications in computing services. A service provider, such as an electronic transaction processor for digital transactions, may utilize different computing services that implement rules and artificial intelligence models for decision-making of data including data in production computing environment. Different services may process data in different data formats and structures. However, transformation of data between different services, such as a gateway service that may receive data processing requests and / or data objects and downstream services that may process such requests and objects, may take significant time and resources. A configuration-driven data transformation platform may intelligently create code for and select from transformers that may be used for data transformations. When selected, the transformers may transform data between services faster and more efficiently by being specifically selected based on past performances and code configurations.

The invention provides a three-dimensional modeling method and device based on visual fusion, a terminal and a storage medium, and relates to the technical field of computer vision and three-dimensional modeling. The method comprises the steps that multi-angle shooting is carried out on a target object through a multi-binocular-camera system after internal and external parameter calibration, a disparity map, shot by each binocular camera, of the target object is obtained, the multi-binocular-camera system comprises multiple binocular cameras, and each binocular camera comprises a left camera and a right camera; fusing each disparity map by using a spatial transformation fusion algorithm to obtain a complete disparity map of the target object; and calculating point cloud data by using the complete disparity map of the target object, and constructing a visual three-dimensional model of the target object based on the point cloud data. According to the method, not only can the integrity and accuracy of the visual three-dimensional model in a complex scene and an object structure be greatly improved, but also the modeling efficiency can be improved.

Retrieval of structured tabular data in a retrieval-augmented generation (RAG) platform is described. An example method includes receiving a tabular file including structured data, extracting short and long metadata from the file, and generating a dataset agent object using the extracted metadata. A human-readable description is associated with the structured data, and the dataset agent object and metadata are stored in a metadata index. A natural language query is received through a query agent object rendered within a canvas including interconnected agent objects. A dataset agent object is identified based at least in part on metadata filtering using the natural language query and short metadata. A structured query conforming to structured query language syntax is generated using a large language model (LLM) and executed on the structured data. A response including the result set and dataset agent identifier is generated and transmitted to a downstream agent object for further processing.

A computer-implemented method and apparatus for managing physical and digital products' customization using container format are disclosed. In some embodiments, the method comprises receiving a container file with serialized objects, deserializing these objects to form a hierarchical object structure, and interacting with a database to obtain real-time configuration data. Custom attributes of the products are defined by binding attribute / value pairs to the hierarchical structure. A visualization of the customized product is generated based on these pairs, and manufacturing instructions are determined considering the attribute / value pairs and manufacturing constraints. The apparatus comprises a memory and a processor to execute instructions for deserializing the container file, retrieving configuration data, organizing objects into a hierarchical tree, negotiating product customizations, and outputting manufacturing data. This system enables efficient and accurate production of customized products by managing data, rendering graphics, and incorporating manufacturing constraints.

The invention relates to the technical field of medical treatment, in particular to ultrasonic equipment and an image processing method. In the embodiment of the invention, an ultrasonic image of a detection part is generated according to an ultrasonic echo signal which is received by an ultrasonic probe and is reflected by the detection part; feature extraction is carried out on the ultrasonic images through encoders in multiple layers of segmentation units contained in the segmentation network; processing the multiple groups of target object structure features output by the encoder in the multi-layer segmentation unit through the decoder in the multi-layer segmentation unit to obtain a segmentation result of the target object structure in the ultrasonic image; and performing partition processing on the ultrasonic image according to a segmentation result, and performing fuzzy enhancement processing on a partition critical region to obtain a processed ultrasonic image.

The invention relates to the technical field of UI design, can be applied to business scenes of financial science and technology, medical health and the like, and discloses a font rendering optimization method, device, equipment and medium. The font rendering optimization method comprises the following steps: identifying position coordinates of a target character in a document object structure and determining a context environment type of the target character and a view area range where the target character is located; generating a font file request based on the target character and the context environment type, obtaining an adaptive font data packet, inputting a rendering model matched with the context environment type, establishing a binding relationship with the view area range, monitoring a display state change event of the view area range, and triggering a font resource updating operation according to event parameters. And obtaining incremental font data and configuring the incremental font data to font rendering parameters of the rendering model. According to the method, the character context environment and the view area state change are combined, the font data corresponding to the target character are requested and loaded on the client according to needs, and refined distribution and real-time updating of font resources are achieved.

Embodiments of the disclosure provide a method, apparatus, electronic device and storage medium for processing image, and the method includes: obtaining an image to be processed; determining an object structural feature within the image to be processed corresponding to a target object and determining a style texture feature corresponding to a reference style image to be applied; and determining a target style image corresponding to the image to be processed based on the object structural feature and the style texture feature.

The invention discloses an adaptive video restoration method based on a digital video technology, and relates to the technical field of image communication. The method comprises the following steps: preprocessing an original video containing hard subtitles and automatically removing the hard subtitles to obtain a video to be repaired; screening a reference frame and a to-be-repaired frame by comparing the original video with the to-be-repaired video; determining a to-be-repaired area according to the pixel gray scale difference; extracting matched feature points, calculating a reference motion vector, estimating possible positions and motion differences of points in the to-be-repaired region in combination with object structure segmentation and region prediction, and generating a corrected motion vector; performing sub-pixel-level interpolation by using the vector, and constructing a corrected reference image; and finally, fusing the corrected image and the to-be-repaired area to generate a repaired video frame. The method can effectively improve the restoration quality of the subtitle shielded area.

The invention discloses a multi-angle photographing unmanned aerial vehicle surveying and mapping method, and relates to the technical field of unmanned aerial vehicle surveying and mapping, through forming a photographing angle score set Vgl, illumination suitability quantitative evaluation in a full attitude domain is realized, and an attitude combination with the highest score in the score set Vgl is selected for image acquisition, so that the imaging quality is improved, and the imaging efficiency is improved. A data basis of optimal illumination conditions is provided for modeling, after image acquisition is completed, a point cloud set Pcd is constructed through an image set Img, an illumination imperfection index Lck is calculated, and the illumination shielding influence in the model is quantitatively evaluated; if the illumination incompleteness index Lck exceeds the illumination shielding threshold value LckThr, updating the attitude angle candidate space and driving an iterative optimization process; the problems of large image shadow interference, exposure imbalance and three-dimensional model defects caused by factors such as natural illumination dynamic change, uncontrollable attitude and surface feature structure shielding in an existing surveying and mapping unmanned aerial vehicle method are effectively solved.

The invention discloses a multi-stage sparse target detection algorithm and device based on multi-modal feature fusion and electronic equipment, and aims to significantly improve the three-dimensional detection precision of a sparse target through multi-stage target mask-based heat map prediction and target-to-region feature interaction. And respectively using the obtained target features to aggregate geometric information related to a target structure in a point cloud, and querying semantic information in image features in corresponding two-dimensional target features, so that features of a predicted target are enhanced by using multi-modal information, the target features help a model to enhance information of a potential sparse target, and the prediction accuracy is improved. And thus, the detection precision is improved.

The invention provides a multi-module fusion self-supervision denoising method for an industrial CT image, and belongs to the field of digital image processing and industrial nondestructive testing. In order to solve the problems of detail loss, poor denoising effect and the like in an industrial CT image denoising process, the invention provides a self-supervised denoising model which takes U-net as a trunk network and fuses an edge enhancement module, a convolution block attention module and a convolution attention fusion module. Meanwhile, by means of a mixed loss function composed of structural similarity index loss, mean square error loss and ResNet perception loss, the noise level of the industrial CT image can be remarkably reduced, meanwhile, the integrity of the object structure and the contour edge is accurately kept, and the image quality is effectively improved.

Disclosed is an improved approach to manage hierarchical metadata for a database system. A hierarchical metadata structure pertaining to a hierarchical object structure of a multitenant database architecture for a plurality of tenants may be maintained where the multitenant database architecture comprising a container database (CDB) that includes a pluggable database (PDB). A request for an access to a metadata object in the hierarchical metadata structure for one or more database objects in the container database may be received. In response to the request for the access to the metadata, access to at least a portion of the hierarchical metadata structure is provisioned.

The invention relates to the field of image processing, and provides a three-dimensional scene style generation method and system based on a weighted style neural radiation field, and the method comprises the steps: enabling a weighted fusion network to achieve the effective capture of potential codes of styles and the fusion weighting of multi-scale features through potential style coding and pyramid weighted fusion feature decoding, and achieving the fusion of the multi-scale features. The texture style hidden variable neural radiation field self-adaptively optimizes the style migration effect, enhances the performance of the generated image in multiple aspects of object structure feature maintenance, color rendition, style fusion naturalness and multi-view coherence, and improves the texture style hidden variable neural radiation field through a network structure combination processing mode of a basic neural radiation field network, an implicit coding network and a combined output network. Three-dimensional scene geometric information and style features of style images are organically fused, then image patch pair sampling and similarity measurement loss functions are utilized, scene texture details and multi-view consistency are reserved, and the three-dimensional scene style generation quality and detail precision are improved.

The application discloses a hypertext page rendering method and device, electronic equipment and storage medium, the method comprises the following steps: in response to the cross-platform application starting instruction, generating the hypertext page rendering request of the target application, and obtaining the web page code information of the target application corresponding to the hypertext page rendering request from the management server, the web page code information is analyzed, and the second hypertext object structure corresponding to the preset rendering frame is generated based on the first hypertext object structure in the code analysis result, the rendering is carried out based on the second hypertext object structure, and the hypertext page is obtained. The method can improve the consistency and smoothness of the hypertext page rendering on multiple platforms, so that the user can obtain the same experience when opening the hypertext page and opening the native page, and the user has no perception.

The invention relates to the field of intelligent question and answer calculation, in particular to a visual impairment group vertical field multi-modal intelligent question and answer method based on a neural network, which comprises the following steps: S1, knowledge base preprocessing: standardizing a knowledge base, including performing information extraction, semantic understanding and multi-modal dialogue on texts in the knowledge base; s2, performing structured processing on the multi-mode vertical domain knowledge base; s3, generating detailed description information from the knowledge of the picture modality through an image-text big model; s4, forming a document by using the decomposed subject-predicate-object structure and the fixed complementary structure to replace the original characters, and identifying the document by using an ID (Identity); s5, collecting related attributes of the entities according to the query ID, and generating a candidate attribute list; and S6, according to the score, determining an attribute relationship closest to the question, and taking out an attribute value of the closest attribute relationship to complete answering. The method has the advantages that more accurate and personalized multimode vertical domain knowledge introduction question and answer service can be provided for visually impaired people to obtain information.

The invention provides a structured data processing method, and relates to the technical field of data processing, and the method comprises the following steps: obtaining task structured data of a target task, obtaining object structured data, and if the object type of a target object belongs to the task type of the target task, obtaining the object structured data of the target task; if the attribute data of the target object meets the sub-task condition corresponding to the intermediate task and the geographic area where the object is located belongs to the task application area of the target task, taking the target task as the intermediate task, and if the attribute data of the target object meets the sub-task condition corresponding to the intermediate task, taking the sub-score corresponding to the sub-task condition as the attribute score of the intermediate task; and taking the sum value of all attribute scores of the intermediate task as a task matching score of the target object and the intermediate task, and if the task matching score of the target object and the intermediate task is greater than a matching success threshold value corresponding to the intermediate task, judging that the target object and the intermediate task are successfully matched. Therefore, the accuracy of judging whether the target object and the target task are successfully matched or not is improved.

Embodiments are disclosed for a process of detecting and processing repeating structure groups of objects in a document using a digital design system. The method may include obtaining, by a page segmentation model, object information for a plurality of objects in a document. The disclosed systems and methods further comprise determining, using the object information, a plurality of object clusters based on distances between the plurality of objects. A repeating structure group of objects can then be identified using the plurality of object clusters. The disclosed systems and methods further comprise providing information indicating the repeating structure group of objects in the document

Computer system (110), comprising: a communication interface (113) configured to communicate with at least one camera, comprising a first camera (270) with a first camera field of view (272); a control circuit (111) configured, when a stack (250, 750) with multiple objects is located in the first camera field of view (272), to: receive camera data generated by the at least one camera, wherein the camera data describes a stack structure for the stack (250, 750), the stack structure being formed from at least one object structure for a first object of the multiple objects;Identify, based on camera data generated by the at least one camera, a target feature (251B, 251C, 751B) of the object structure or a target feature (251B, 251C, 751B) located on the object structure, wherein the target feature (251B, 251C, 751B) is at least one of the following: a corner (251B) of the object structure, an edge (251C) of the object structure, a visual feature (751B) located on a surface (251A, 751A) of the object structure, or an outline of the surface (251A, 751A) of the object structure; Determine a two-dimensional, 2D, region (520, 620, 720) that is coplanar with the target feature (251B, 251C, 751B) and whose boundary is defined by the target feature (251B, 251C, 751B); Determining a three-dimensional, 3D, region (530, 630, 730) defined by connecting a location of the first camera (270) and the boundary of the 2D region (520, 620, 720), wherein the 3D region (530, 630, 730) is part of the first camera's field of view (272);Determine, based on the camera data and the 3D region (530, 630, 730), a size of an occlusion region (570, 670, 770), where the occlusion region (570, 670, 770) is a region of the stack structure located between the target feature (251B, 251C, 751B) and the at least one camera (270) and within the 3D region (530, 630, 730); Determine a value of an object detection confidence parameter based on the size of the occlusion region (570, 670, 770); and performing an operation to control robot interaction with the stack structure, wherein the operation is performed based on the value of the object recognition confidence parameter; wherein the first camera (270) with which the communication interface (113) is configured to communicate is a 3D camera configured to generate, as part of the camera data, several 3D data points that specify corresponding depth values ​​for locations on one or more faces of the stack structure.

The invention provides an object pose estimation method. The object pose estimation method comprises the steps of obtaining object structure features of a target object, hand pose features of a robot hand and contact area features of the hand in contact with the target object; fusing the object structure feature and the contact area feature to obtain a first fusion feature; fusing the hand posture feature with the contact area feature to obtain a second fused feature; fusing the first fusion feature and the second fusion feature to obtain a global fusion feature; and performing pose estimation based on the global fusion features to obtain a target pose of the target object. According to the embodiment of the invention, the accuracy of object pose estimation can be improved.

The present invention discloses a category-level 6D object pose estimation method based on a point cloud map attention network, including: S1, preprocessing the input RGB-D image data to extract the observed point cloud of the object under the depth camera; S2, using the point cloud map attention network to extract multi-scale local-to-global object structure features from the observed point cloud; S3, reconstructing the 3D point cloud model of the object by using the shape prior adaptation mechanism and the category shape prior point cloud and regressing the normalized NOCS coordinates of the object; S4, calculating the similarity transformation between the reconstructed NOCS coordinates and the observed point cloud through the Umeyama algorithm to obtain the pose and size information of the object. Experiments conducted on the NOCS-REAL dataset prove that the solution proposed by the present invention is superior to the prior art and achieves better results.

The invention relates to a software development kit generation method, system and device and a readable storage medium. The method comprises the steps of obtaining service metadata of a target application program interface, wherein the service metadata is determined according to technical metadata and service metadata related to service description information; according to the service metadata, determining an interface description object structure and a parameter object structure defined by a software development kit interface; the software development kit is used for calling the target application program interface; based on the interface description object structure and the parameter object structure, packaging the service metadata to obtain interface description object information and parameter object information; and constructing a software development kit through the interface description object information and the parameter object information. By adopting the method, the coupling degree of the SDK and the interface can be improved.

The present application relates to the field of image processing, and proposes a kind of three-dimensional scene style generation method and system based on weighted style neural radiance field, including weighted fusion network, through latent style encoding and pyramid weighted fusion feature decoding, realize the effective capture of latent encoding of style and the fusion weighting of multi-scale feature, adaptively optimize style transfer effect, enhance the multi-aspect performance of generated image in object structure feature preservation, color restoration, style fusion naturalness and multi-view coherence, texture style hidden variable neural radiance field is processed by the network structure combination of basic neural radiance field network, implicit encoding network and merging output network, organically fuses the style features of three-dimensional scene geometry information and style image, then uses picture patch to sample and similarity measure loss function, retain scene texture details and multi-view consistency, the present application improves the quality and detail precision of three-dimensional scene style generation.

The invention provides a deserialization method and device, electronic equipment and a medium, and relates to the technical field of computers. The method comprises the steps that under the condition that the structural hierarchy of a first attribute nested in a type in a data object is inconsistent with the structural hierarchy of a corresponding field in a JSON character string, in a declaration of the data object, a first decorator and a second decorator are added for the first attribute, the parameter of the first decorator is the literal quantity of the type of the first attribute, and the parameter of the second decorator is the literal quantity of the type of the first attribute; the parameter of the second decorator is a predefined analytic function; and in the deserialization process, analyzing the JS object corresponding to the first attribute according to an analysis function indicated by the second decorator, and returning an analysis result according to a specified structure. According to the method, complex and self-defined JSON analysis logic can be realized, a JSON structure does not need to be modified, and the development and joint debugging cost is saved; the method does not need to declare a data object with complex hierarchy in order to strictly match a JSON structure, so that the structure of the object is clearer and easy to maintain.

The application relates to the technical field of computer vision, and provides a high-stability virtual object position display method and device. In a mapping stage, the object structure of a virtual scene is acquired, and an exception handling boundary is established; in the exception handling boundary, a virtual object in the virtual scene is regarded as a general engine object, coordinate data corresponding to the general engine object is acquired based on the object structure, and a path graph is drawn according to the coordinate data; the running state of a data processing process in the exception handling boundary is monitored, if running exception is detected in any processing period, the exception data corresponding to the running exception is discarded; in a virtual scene running stage, the position information of the virtual object is acquired in the exception handling boundary, the position information is matched with the path graph, the relative position of the virtual object in the path graph is determined, and the relative position of the virtual object in the path graph is displayed. Runtime exceptions in memory reading or drawing processes are effectively intercepted and isolated.

The invention provides an open vocabulary image semantic segmentation method and system, and provides a text guide feature refining module and a structure perception guide enhancement module. Through the method provided by the invention, the semantic correlation region in the visual feature can be selectively focused, and the interference of irrelevant background noise is inhibited, so that the distinguishing capability of categories with similar texture features is effectively improved. Meanwhile, rich space information contained in the CLIP intermediate visual features can be fully mined to guide the up-sampling process of the model, the recovery capability of object structure details is remarkably improved, and the boundary of the object is described more accurately.

The invention discloses a dynamic component rendering method and system based on an object structure. The method comprises the steps that a parent component configuration object is received, and a child component configuration object is searched for according to child component configuration children in the parent component configuration object; the structure type of the sub-component configuration object is confirmed, different processing strategies are adopted for the sub-component configuration objects of different structure types, and a rendering instance is created; and rendering by adopting different rendering methods according to different rendering instances. According to the technical scheme, different rendering strategies can be adopted for logic rendering according to the sub-component configuration objects of different structure types, and dynamic generation and rendering of the components are achieved.

This application provides a Gaussian image rendering method and apparatus. Starting with multi-view image rendering based on an initial Gaussian loss function, color and depth images from different viewpoints are gradually acquired. Multiple loss functions are then used to specifically optimize the properties of the initial Gaussian loss function, resulting in a first Gaussian loss function. Based on the first Gaussian loss function, rendering continues from more viewpoints, and a third loss function, formed by comparing with the real image, is used to further optimize the Gaussian loss function, ultimately yielding a second Gaussian loss function for rendering images from new viewpoints. This entire process ensures that the rendered fifth color image achieves good results in terms of accuracy, realism, and detail rendering, more realistically restoring the visual appearance of the target scene from new viewpoints and reducing problems such as blurring, distortion, and unreasonable object structures in the image.

The invention relates to the technical field of electric energy meter data management, in particular to a data coding method, device and equipment of an electric energy meter facing object and a medium. The method comprises the following steps: S1, respectively defining an object attribute structure and a general object structure; s2, obtaining an object identifier of a target object from the received message or instruction, and searching a corresponding universal object structure in a preset object set according to the object identifier; s3, according to the general object structure corresponding to the target object, verifying the access authority of the target object attribute and checking the validity of the data source of the target object attribute; and S4, in response to the readable target object attribute and the valid data source, dynamically binding and calling a corresponding coding function through a virtual method table in the object attribute structure corresponding to the target attribute, and coding the data source according to a predetermined coding rule, so as to improve expansibility, code conciseness and maintenance convenience.

The application discloses a kind of methods for realizing Java Applet automation operation, comprising the following steps: step one: write a UDP Server, define interface element object structure, structure contains the handle that element holds, on the location of interface, element type data;Step two: write Java Plugin by Java Accessibility API, this Java Plugin is mainly loaded by JVM when starting;Step three: write Native Java Bridge, for JVM, operating system, and automation program communication, through HOOK technology mounts global mouse event, when mouse moves to specified Java Applet, call Plugin, and Java interface element object instance is reflected;Step four: construct UDP Client in Native Java Bridge, when Java interface element object is reflected, it is serialized into the same data type agreed with UDP Server, and is sent to UDP Server by UDP Client;The application realizes the automation operation of system developed with Java Applet technology, builds overall automatic platform, so that each system can be automatically connected, instead of traditional manual operation.