102 results about "Geometric representation" patented technology

A system and method for processing genomic data using dynamic latent manifolds that transforms multi-modal genomic datasets into geometric representations within a curved manifold space. The system receives genomic datasets including DNA sequences, genetic variants, and expression data, then extracts biological features and assesses importance using trained neural networks. Manifold curvature values are computed based on biological significance, and genomic data is embedded as geometric structures where semantic relationships are represented through distance and curvature properties. The system generates compression pressure fields that influence processing decisions and computes optimal geodesic paths through the manifold to minimize cognitive action functionals. Adaptive compression rates are determined for different genomic regions based on geometric properties and biological importance. The manifold structure evolves through use, strengthening frequently accessed pathways while applying thermodynamic decay to unused concepts. The system supports hierarchical organization across biological scales, reversible navigation, and federated learning capabilities that enable privacy-preserving collaboration.

The invention discloses a virtual space multi-person interaction synchronous control method oriented to an end-cloud collaborative architecture. The invention relates to a computer graphics and three-dimensional modeling technology, and discloses a cross-format lightweight and geometric consistency maintenance method based on a three-dimensional model. Through format-independent geometric representation and a self-adaptive lightweight strategy, efficient compression and precision maintenance of three-dimensional model cross-format conversion are realized. The method specifically comprises the following steps: performing format analysis and geometric feature extraction on an input model, and establishing a unified internal representation; adaptively selecting a multi-level LOD lightweight strategy based on the complexity of the model; the accuracy of key information is ensured through geometric feature keeping and topology consistency detection; the geometric consistency is dynamically maintained by combining error monitoring and an iterative correction mechanism; and generating a target format lightweight model and carrying out quality verification. According to the method, adaptive precision control, multi-level consistency maintenance and format irrelevant processing are combined, the model size and conversion errors are effectively reduced, and cross-platform compatibility and geometric fidelity are improved. The method can be widely applied to the fields of industrial design, game development, virtual reality and the like.

The invention provides a data processing method in an electric power construction supervision project, and the method comprises the steps: carrying out the scanning of an engineering construction site through a data collection device, obtaining a three-dimensional point cloud data set containing geometric information, and enabling the three-dimensional point cloud data set to comprise the complete geometric representation of a site space; according to the three-dimensional point cloud data set, a data processing algorithm is adopted to remove noise information, processed point cloud data is obtained, and the processed point cloud data is used for subsequent modeling analysis; aiming at the aligned point cloud model, geometric parameters of key parts are extracted, actually measured geometric feature data are obtained, and the actually measured geometric feature data comprise actual size information of site construction; and generating a quality evaluation report according to the deviation distribution data, and sorting the abnormal regions according to deviation degrees to obtain a quality adjustment priority sequence.

Variants of the method can include determining a design, generating a geometric representation of the design, generating a topological representation of the design, determining a context value set, predicting a physical field using a trained physics model, optionally verifying the physical field, optionally generating a corrected physical field, and optionally training the physics model based on the corrected physical field.

The invention relates to the technical field of computer vision, in particular to a complex scene object recognition system based on multi-scale feature fusion, which comprises a feature extraction module, a Riemannian geometric feature pyramid module, a candidate region generation module and an instance recognition and positioning module, and is characterized in that the feature extraction module extracts multi-scale spatial features; the Riemannian geometric feature pyramid module maps spatial features to a Riemannian manifold space, calculates curvature distribution of feature manifolds, dynamically adjusts feature extraction levels and fusion strategies, and realizes geometric consistent fusion of different scale features through geodesic line feature fusion; the candidate region generation module generates an object candidate region based on the fusion features; and the instance identification and positioning module completes final object classification and positioning, introduces a differential geometry theory to construct a Riemannian geometric feature pyramid, realizes geometric representation and adaptive fusion of features, solves the problems of scale change, partial occlusion, background interference and the like in a complex scene, and effectively improves the accuracy and robustness of object identification.

A training method, apparatus and system for a feature extraction network of a 3D mesh model are provided. The method includes dividing a training 3D mesh model into a plurality of patches which do not overlap with each other, dividing the plurality of patches into first-type patches and second-type patches, and using mask embedding as a feature encoding of each second-type patch; inputting geometric representation information and positional representation information of each first-type patch into a feature extraction network; determining predicted geometric representation information of each face based on a feature encoding of each first-type patch output from the feature extraction network, the mask embedding, and positional representation information of each second-type patch, and adjusting parameters of the feature extraction network based on differences between the predicted geometric representation information and geometric representation information of the each face.

The invention discloses a point cloud registration method based on unsupervised multi-scale feature fusion and geometric representation. Two innovative modules, namely a multi-scale neighborhood feature fusion module and a triangle-based representation learning module, are provided. The multi-scale neighborhood feature fusion module is used for extracting local neighborhood information from point clouds of different scales, balancing fine details and a global structure on multiple scales by combining multi-scale neighborhood construction, an attention mechanism and a feature fusion strategy, and overcoming the challenge of single-scale feature extraction in processing local shape similarity. And the triangle-based representation learning module deeply fuses the multi-scale neighborhood features and geometric constraints based on triangular geometric descriptors to generate feature representation with higher discrimination, so that the accuracy and stability of registration are remarkably improved. Finally, the multi-scale neighborhood feature fusion module and the triangle-based representation learning module enable the model provided by the invention to show excellent registration performance and have excellent applicability.

The invention provides a three-dimensional semantic scene completion method and system based on semantic-physical joint representation. The method comprises the following steps: extracting image features according to multi-view image input, and mapping the image features to a three-dimensional voxel space; voxel features containing local semantic information and three-view features containing global semantic information are extracted through a double-branch encoder, key anchor points with identifiable semantic features are screened out according to semantic consistency of the voxel features and the three-view features, and efficient initialization of Gaussian representation in a key area is guided; and then introducing a semantic spherical harmonic function, projecting Gaussian semantic representation to an orthogonal spherical harmonic space, enhancing the modeling capability of local context geometric structure information, and promoting the semantic-geometric representation consistency of the context by using key voxel distribution alignment, thereby realizing three-dimensional semantic scene completion with both precision and efficiency.

There is provided a system and method for determining a three-dimensional model of an object using a projector and a light sensor arranged in a stereo configuration. The method including: receiving pixel intensity values captured by the light sensor for a plurality of images of a scene containing the object, wherein the images of the scene each capture different projection patterns emitted by the projector onto the scene; determining, using the pixel intensity values from a set of the plurality of images, correspondence between a projector plane and a camera plane to form a three-dimensional model of the object, the correspondence determined using a trained artificial neural network model that uses a combination of volume rendering and geometric representation; and outputting the three-dimensional model of the object.

The invention provides a visualization method of finite element result data in a dam three-dimensional model, and relates to the technical field of finite element analysis, and the method comprises the steps: constructing a finite element spatial discrete model based on geometric information in the finite element result data; based on a sampling position in the dam three-dimensional model, determining a target space unit or a target space region corresponding to the sampling position in the finite element spatial discrete model, and obtaining first dam body displacement data corresponding to the target space unit or the target space region from the finite element result data; converting the first dam body displacement data into a first visual parameter according to a preset mapping relationship, and associating the first visual parameter to a geometric representation of the dam three-dimensional model to form a visual attribute for rendering; and carrying out rendering display on the dam three-dimensional model based on the visualization attribute so as to output a displacement visualization result of the surface of the dam three-dimensional model. According to the method, the intuition, interactivity and updating efficiency of visual display can be improved.

A system and method for implementing experiential manifold cognition that extends persistent cognitive machines beyond discrete thought caching to continuous geometric representation of experience. The system maintains an experiential manifold comprising a differentiable manifold with Riemannian metric tensor encoding semantic relationships, compression pressure field governing memory consolidation, and potential field encoding goals and attention. Input data is projected onto the manifold through adaptive geometric diffusion preserving semantic structure. The system executes geometric transformations including metric evolution, geodesic computation, and curvature estimation. During non-interactive periods, autonomous evolution occurs through trajectory recombination and selective pruning. A user interface enables visualization and direct manipulation of manifold geometry, translating navigation into geodesic traversal and edits into metric modifications. The system maintains persistence across sessions and enables controlled federation between multiple manifolds through consent-bounded synchronization. Applications include persistent narrative worlds, collaborative cognitive spaces, and experiential intelligence systems that learn through geometric evolution.

The invention discloses an electromagnetic field simulation adaptive grid generation method taking electromagnetic field distribution characteristics as core guidance and a system thereof. Combining the geometric model data of each radio frequency device with the connection relation information to form a geometric representation containing topological connection information; setting a current sampling point on a non-contact surface of the surface of the metal radio frequency device and other radio frequency devices based on the geometric representation containing the topological connection information, and obtaining a current value of the current sampling point; s parameters of two rounds and current values of current sampling points are obtained through the first two rounds of mesh generation; if the convergence condition is met, the grid generated in the first round is used for full-wave-band simulation; and if the convergence condition is not met, carrying out a local encryption process until the convergence condition is met, and carrying out full-wave band simulation by using the grid generated in the previous round. According to the method, the limitation that traditional grid generation is only adaptive to geometric shapes and neglects electromagnetic characteristics is broken through, and collaborative optimization of grid quality and calculation efficiency in a high-frequency electromagnetic field simulation scene is realized.

A system and method for implementing persistent cognitive computation through geometric representation of thought in a dynamic latent manifold. The system encodes inputs into a curved space characterized by time-evolving metric tensors, compression pressure fields derived from Ricci curvature, and goal potential fields that shape attention flow. Cognition occurs through geodesic traversal of this manifold, with attention following paths that minimize cognitive action while balancing semantic density and goal relevance. A Cognitive Dynamics Engine maintains manifold geometry, computing optimal trajectories and managing thought bundle operations including consolidation, expansion, and higher-order abstraction. During idle periods, autonomous dreaming processes reorganize the manifold through perturbation, recombination, and topological surgery. This architecture enables persistent memory through geometric encoding, where frequently accessed concepts develop high-curvature regions and cognitive shortcuts emerge from usage patterns, transforming artificial intelligence from stateless computation to structured motion through shaped memory space.

A system and method for implementing persistent cognitive computation through geometric representation augmented with holonomy-based experiential memory. The system encodes inputs into a curved latent manifold and maintains bounded sets of holonomy descriptors at each location, enabling two-component cognitive states comprising position and experiential context. Cognition occurs through holonomy-sensitive traversal where paths depend jointly on geometric structure and accumulated path-dependent constraints. Holonomy generators are created during traversal from prediction errors and constraint encounters, composed into consolidated descriptors, and undergo lifecycle management including reinforcement, decay, and irreversible export to residual constraint regions. This architecture escapes location-only representations by distinguishing cognitive states that occupy identical semantic positions but arise through different experiential histories. The system supports counterfactual reasoning through holonomy switching at fixed locations and preserves semantic memory as compressed transport deformation rather than stored trajectories, enabling scalable experiential learning where repeated patterns strengthen constraints while capacity remains bounded.

A computer-implemented method trains an artificial intelligence model based on neural networks to predict physical features of a brake caliper from a geometric representation, such as a mesh. The method provides for receiving geometric representations of brake calipers, each comprising local regions such as vertices and faces, and local information relating to the geometry. The artificial intelligence model is trained using input data comprising the received local information and output data relating to physical features of the brake caliper, including fields of stresses due to the pressure of brake fluid and the friction force, and the integral of the elastic deformation energy. The system comprises an input interface to access training data and a processor module configured to train the artificial intelligence model. The trained artificial intelligence model predicts physical features of a brake caliper from the geometric representation, providing output data such as stress fields and deformation energy.

This invention relates to the field of computer graphics processing and 3D modeling technology, specifically to an automated intelligent generation system for 3D modeling of shoe exteriors. The system includes: data acquisition and streaming, where an image acquisition terminal generates and sends four-channel image data using a high-precision sensor array; feature reconstruction and representation, where a topological semantic reconstruction server extracts semantic features and constructs an initial geometric representation using an implicit function based on pixel alignment and a decoder to predict the symbolic distance function; topology optimization and segmentation, where topological streamline optimization and component semantic segmentation are performed based on the curvature field to generate a structured 3D model with independent component structures; and interaction and local updates, where an interactive terminal responds to user editing operations on independent component structures, sending local topology update commands, and the server performs parametric recalculation of the local mesh based on the commands. This invention solves the problem of the separation between geometric structure and semantic information, achieving real-time and refined model updates.

Variants of the method can include determining a design, generating a geometric representation of the design, generating a topological representation of the design, determining a context value set, predicting a physical field using a trained physics model, optionally verifying the physical field, optionally generating a corrected physical field, and optionally training the physics model based on the corrected physical field.

The application discloses a generative AI industrial design topology closed-loop optimization method and system, and relates to the technical field of computer-aided design, comprising: obtaining an initial three-dimensional geometric representation of an industrial product and design constraints, and constructing a geometric parameter offset field; in iterative deformation, calculating discrete curvature, deformation energy, time sequence mutation and topology anomaly index, triggering and positioning a topology sensitive area; in the area, performing edge collapse, edge split, local re-meshing, Loop subdivision, constraint Laplace smoothing or SDF local reconstruction, and verifying and repairing the results by Euler characteristic number, genus, boundary ring, connected component and topology anomaly; calling a physically consistent proxy model to predict performance and evaluate confidence, triggering finite element sampling when confidence or safety margin is insufficient, error risk and topology sensitive area overlap, updating the proxy model by using finite element true value increment, and finally completing layered optimization and complete finite element verification.

The embodiment of the invention mainly relates to the field of simulation, in particular to a model simplification method and device, electronic equipment, a medium and a program product. The method comprises the steps of extracting attribute data of a model; according to the attribute data of the model, all components in the model are divided into at least two categories, and the first category is used for representing at least part of invisible components of the model; deleting or hiding the components of the first category in all the components; and simplifying the remaining parts to obtain a simplified geometric expression set.

The invention relates to the technical field of integrated learning, in particular to an intelligent livestock disease prevention system which comprises a local geometric representation module, a performance model preliminary screening module, a decision logic verification module and an individual risk early warning module. According to the method, a plurality of base learners are evaluated and screened by mining a local geometrical relationship among physiological data and constructing a weight vector, a candidate model set adapted to current data features is formed, secondary preferential selection is performed according to the feature contribution degree concentration degree of each model during prediction, and a model with the most decision logic certainty is selected to execute a prediction task. The dynamic construction and preferential mechanism enhances the adaptability of the model to data fluctuation, avoids the limitation of generalization ability of a single model, and significantly improves the accuracy of disease risk identification and the stability of system operation.

The application discloses a robot control method, a computing device and a readable storage medium. The method comprises the following steps: acquiring visual information of a first region; determining a target object to be grabbed based on image position features of each object in the visual information of the first region, and controlling a robot to grab the target object; acquiring visual information of a second region; determining a target position in the second region for placing the target object based on the visual information of the second region and a three-dimensional geometric representation of the target object; and controlling the robot to place the grabbed target object to the target position. In this way, the grabbing and placing can be adapted to a dynamic environment, and the success rates of the grabbing task and the placing task are improved.

The application provides a kind of chuck suction pose generation method and system, comprising: obtaining the RGB image of the object to be grasped and generating its geometric representation;Scale correction is carried out on the geometric representation by using camera calibration, reference datum or sparse depth data and other information to generate executable target geometric representation;Suction pose is generated based on target geometric representation and robot execution is controlled;According to the sensor feedback such as negative pressure or flow rate in the suction process, success or failure label is automatically generated;Label and corresponding visual data form training sample, and closed-loop iterative training is carried out on pose generation model.The application solves the problem of poor robustness of traditional method in processing poor visual objects, and realizes low-cost, automatic, safe and continuous optimization of model by using physical feedback, and enhances the adaptability of the system.

The invention discloses a robot path navigation method based on multi-modal perception, relates to the technical field of intelligent robots, and can solve the problems of insufficient utilization of visual semantic information, limited perception capability of a three-dimensional geometric structure and difficulty in migration from simulation to reality in the prior art. The method comprises the steps of collecting multi-source original data and generating three-dimensional Gaussian splashing visual representation and voxelization geometric representation; constructing a high-throughput simulation training framework, and integrating the three-dimensional Gaussian splash model as a plug-and-play renderer to a vectorization physical simulator; extracting visual semantic features through a pre-trained visual encoder, and extracting three-dimensional geometric features through a grouped convolutional neural network; fusing the visual semantic features, the three-dimensional geometric features and the ontology perception features through a recurrent neural network to generate potential representation; and training the navigation strategy network based on the potential representation and migrating deployment. According to the invention, efficient semantic perception navigation and steady simulation-to-reality migration of the robot in a complex environment are realized.

The application discloses a head reconstruction method based on generation and inverse rendering collaborative optimization, comprising the following steps: step 1, generating a corresponding binary mask; step 2, fitting a three-dimensional geometric grid of a face matched with an input person based on detected two-dimensional face key points to obtain a unified head coordinate system; step 3, projecting pixels of an input image to a UV space of the three-dimensional geometric grid of the face to form a defective texture with original illumination and self-occlusion; performing illumination regularization and an inverse rendering network to output a multi-channel PBR material map; step 4, generating a group of videos rotating around the head and camera internal and external parameters thereof, and cutting the videos into a multi-view hair image sequence and corresponding virtual camera parameters; step 5, performing multi-view hair reconstruction to obtain a hair geometry representation or equivalent geometry representation representing a three-dimensional structure of hair; and step 6, performing spatial alignment and combination to output a head PBR three-dimensional asset.

The invention provides a dexterous hand autonomous grabbing method guided by geometric implicit representation of an object, which comprises the following steps of: S1, acquiring geometric features of the surface of the object: sequentially processing and complementing point clouds of the object, estimating and coding the geometric shape of the object, and acquiring the geometric representation of the surface of the object adaptive to grabbing operation; s2, estimating the six-degree-of-freedom pose of the object; s3, according to the object surface geometric features extracted in the step S1 and the six-degree-of-freedom pose estimation value obtained in the step S2, grabbing parameters are generated; s4, building a simulation model of a real arm based on a physical simulation platform, arranging a grabbing scene, configuring camera parameters, and building a grabbing data set in a simulation environment; and S5, generating a loss function. The object grabbing success rate and adaptability of the dexterous hand in the unstructured environment are improved, and stable and generalized grabbing of the five-finger dexterous hand is achieved.

The present application belongs to the technical field of flow field, and particularly relates to a general simulation method of complex particle two-phase flow. The method comprises the following steps: S1: constructing a geometric representation model to represent the geometric information of the complex particle and dividing a grid; S2: on the basis of the grid divided in S1, adopting a characteristic line splitting finite element method to solve the flow field to obtain the flow field velocity at pseudo n+1 time and the flow field pressure at n+1 time; S3: calculating the force and torque of the complex particle; and S4: calculating the attitude of the complex particle through a six-degree-of-freedom model. The present application has the following effects: high calculation efficiency, reduced calculation complexity and cost, and improved simulation efficiency. Strong comprehensiveness, which can accurately describe the translation and rotation motion of the complex particle and provide comprehensive dynamic analysis. Good stability, which realizes stable solution of the coupling of fluid and particle motion and avoids numerical oscillation and instability. Multi-scale simulation capability, which can accurately simulate at different scales and meet the demand of multi-scale numerical simulation.

The application discloses a kind of based on diffusion model and DMTet representation's spatial structure node intelligent generation method and device, the method includes: (1) build the initial model of space structure node, and establish mixed node sample set;(2) to mixed node sample set is enhanced and preprocessed, and the node geometry is converted into the unified three-dimensional training representation based on DMTet;(3) train diffusion model to learn node geometry distribution;(4) generate new node three-dimensional geometry representation by reverse diffusion iteration;(5) the generated result is mapped back to DMTet representation, carries out geometry reconstruction, obtains engineering available model.The application is through the technical route of " mixed sample set + diffusion model + DMTet continuous representation + tetrahedral mesh extraction", while guaranteeing the continuity of node structure and the rationality of mechanics, improves node form diversity, training stability, geometry continuity and engineering entity reconstruction efficiency, and is suitable for the intelligent design of complex spatial structure node.

The invention discloses a contour vector diagram generation method and apparatus, a storage medium, a device and a product. The method comprises the steps of obtaining a depth map covering a whole scene; constructing an agent model based on the depth map, wherein the agent model is a geometric representation of the scene surface; generating a cutting plane sequence with equal height spacing according to the proxy model; performing intersection calculation according to the cutting plane sequence and the proxy model, and extracting an intersection line as a contour line; and outputting a vector diagram file according to the contour line. According to the contour vector diagram generation method provided by the invention, the depth map is adopted to construct the proxy model and the adaptive cutting calculation is carried out, so that the contour vector diagram is efficiently generated in a large-scale scene, meanwhile, the balance between the calculation precision and the resource consumption is ensured, and unnecessary calculation waste is avoided.

The embodiment of the invention relates to the technical field of clothing detection, in particular to a clothing wrinkle real-time detection method and system based on geometric features, and aims to enable a tester to wear target clothing to execute predetermined actions. The method comprises the following steps: acquiring RGB image information in each action state and three-dimensional point cloud data of a garment surface, analyzing the RGB image information and the three-dimensional point cloud data to obtain image semantic features and geometric representation features, and performing cross-modal attention fusion on the image semantic features and the geometric representation features to generate wrinkle information of a target garment.

The present disclosure provides a system (100) and a method (200) for generating routes for flexible components that connect principal components in a predefined environment. The method (200) comprises converting (202) a boundary representation of the principal components into a geometric representation and defining (204) an enclosed space around the principal components based on user-defined parameters. The method (200) comprises discretizing (206) the enclosed space into a plurality of nodes, assigning (208) one or more parameters to a node to construct a spatial structure within the enclosed space, and identifying (210) neighboring nodes among the plurality of nodes connected to the enclosed space within the spatial structure in order to generate one or more connectors to the one or more neighboring nodes.Furthermore, the procedure (200) includes determining (212) a shortest path by identifying start and end points for each connector and generating (214) routes for the flexible component that connects the main components.