43 results about "Human body modeling" patented technology

The invention discloses a virtual garment customization method based on 3D modeling, and relates to the technical field of garment customization, and the method comprises the steps: collecting image video data corresponding to the body data of a user, and carrying out the preprocessing of the collected data, and obtaining the basic three-dimensional point cloud data of a human body; carrying out primary human body modeling according to the human body basic three-dimensional point cloud data, then adjusting the model shape based on key parameters input by a user, and carrying out body type matching to realize parameterized human body modeling; secondly, modeling the clothes, predicting clothes deformation by adopting machine learning, simulating cloth dynamics by adopting a physical engine, performing collaborative tracking and matching on clothes and human body dynamic characteristics according to a clothes template selected by a user, and realizing virtual try-on through an AR (Augmented Reality) technology; the customized order confirmation information of the user is obtained, the 2D garment pattern graph is generated and sent to the customized production server, it can be ensured that the user can obtain a high-quality 3D modeling result under different devices and environments, and the user experience can be enhanced through AR try-on.

The invention discloses a costume design assisting method and system based on artificial intelligence, and belongs to the field of costume intelligent design, and the method comprises the steps: extracting multi-dimensional style feature information from a pre-constructed costume image data set and a fashion trend database through employing a convolutional neural network, and constructing a style vector space model; vector similarity matching is carried out, and a costume design sketch is generated by adopting a graph neural network; based on the design sketch, generating a structured process data packet for proofing; the design sketch and the structural data are input into a three-dimensional human body modeling and simulation module, and wearing effect simulation of virtual clothes is realized based on human body motion capture and physical cloth simulation technologies; and constructing a human-computer interaction interface, and quickly responding to a modification request based on the generative adversarial network model. Structured information such as the garment pattern structure, the sewing sequence, the material specification and the process route is automatically generated through the reinforcement learning algorithm, and the garment proofing efficiency and accuracy are remarkably improved.

The invention discloses a figure modeling and fitness action teaching method and system based on non-contact action capture, and the method carries out the real-time monitoring of a human body posture through a millimeter wave radar, carries out the 3D human body posture reconstruction and analysis through combining with a human body posture frame of Google image processing, and achieves the high-precision action capture. The method can be used for human body modeling, body building action scoring and action guidance, action evaluation and real-time feedback in body building teaching can be effectively supported, the millimeter wave radar is adopted for human body posture capture, external markers or wearable equipment are not needed, constraint on user activities is avoided, and use convenience and comfort are improved. The 3D posture of the human body can be accurately reconstructed in combination with a deep learning neural network, and real-time processing and analysis are carried out. The method not only can realize accurate motion capture, but also can be combined with fitness motion scoring and guidance, is widely applied to the fields of virtual fitness, exercise rehabilitation and the like, and has great application potential.

The invention provides a physiological mechanism driving and data fusion method and system for sitting posture human body modeling, and relates to the technical field of digital twinning and biomechanics crossing. The method comprises the following steps: mapping human body macroscopic parameters into biomechanical attribute parameters; generating a basic three-dimensional human body grid based on the human body macroscopic parameters; generating a deformed three-dimensional human body grid in combination with the basic three-dimensional human body grid and the human body macroscopic parameters; solving an optimal attitude based on the vehicle geometric parameters, the task scene identifier and the biomechanical attribute parameters; finally, the deformed three-dimensional human body grid and the optimal posture are integrated, the high-fidelity sitting posture digital twinborn body is generated, end-to-end automatic modeling from macroscopic parameters to the digital twinborn body is achieved, and the core problems that the form and the posture are not matched, crowd coverage is insufficient and task adaptability is poor are solved.

The invention discloses a digital human rendering method based on Gaussian splashing and multi-scale characteristic field distillation, and belongs to the field of three-dimensional human body digital reconstruction. According to the method, feature extraction is carried out through a Vision Transformer encoder, based on an SMPL model, through a cross-modal parameter estimation module and dynamic human body modeling of three-dimensional Gaussian splashing and semantic feature rendering, three-dimensional Gaussian is projected to a two-dimensional image plane to calculate a covariance matrix and color mixing, and after a rendered color image and an initial feature field image are output, a three-dimensional image is obtained. A student feature map is obtained through a convolution acceleration module, a teacher feature map is obtained after feature extraction is carried out through a two-dimensional basic model, the constructed model is trained, and optimization training is completed through comprehensive total loss function calculation; according to the method, the problems of fuzzy semantics, detail missing, low rendering efficiency and inaccurate human body-scene separation in the existing method are effectively solved, so that more efficient, fine and robust monocular or multi-view human body three-dimensional reconstruction is realized.

The invention relates to the technical field of virtual reality, and discloses a personalized clothing virtual try-on method and system, a terminal and a storage medium. The method comprises the following steps: S1, acquiring a three-dimensional body model of a user, and constructing a high-precision three-dimensional body model by analyzing a multi-angle picture uploaded by the user and utilizing a deep learning technology; s2, according to the three-dimensional body model and personal attributes of the user, performing personalized adjustment on the virtual clothes, including color, size and material; and S3, superposing the adjusted virtual costume on the three-dimensional body model of the user by using an augmented reality technology. Through combination of multi-angle image analysis and a deep learning three-dimensional reconstruction technology, a high-precision human body modeling effect is realized. The purpose that the user can quickly generate the three-dimensional body model only through the daily photos is achieved. Compared with a modeling method depending on a depth camera or somatosensory equipment in the prior art, the user operation process is remarkably simplified, and the problems that a traditional scheme is high in threshold and limited in application scene are solved.

The invention relates to the technical field of human body three-dimensional modeling, in particular to a human body measurement parameter set-based three-dimensional body surface shape generation method, which comprises the following steps of: S1, establishing a parameterized body surface model T; s2, determining a target anthropometric parameter group # imgabs0 #; S3, determining a transformation relation F based on the parameterized body surface model T and the target anthropometric parameter group # imgabs1 #; S4, constructing and verifying an optimized physical sign parameter beta *; and S5, substituting the optimized physical sign parameter beta * into any pose parameter theta, and changing the body surface geometric space pose of the optimized physical sign parameter beta *. By means of the method, finite element modeling of the human body in any form can be achieved, a statistical regression or condition generation model from one-dimensional measurement parameters to three-dimensional human body shapes does not need to be established, and various downstream human body modeling tasks oriented to sign diversity are achieved.

The invention discloses an intelligent human body modeling method and system based on a three-dimensional dot matrix, and relates to the field of intelligent modeling, and the method comprises the steps: carrying out the structural analysis of the semantic level of an original three-dimensional point cloud, and predicting a space displacement field pointing from the surface of observed clothes to the surface of an internal real human body under the semantic guidance. By applying the displacement field, double transformation of point cloud stripping and standardization is carried out on the point cloud, so that repeated fine tuning and optimization are carried out on the body type and posture parameters of the human body model in a standardized space, and finally a high-fidelity human body three-dimensional model is generated through driving. In this way, the complex clothes interference problem is deconstructed into a staged geometric mapping and parameter optimization process, and therefore the robustness and accuracy of human body three-dimensional reconstruction under the real wearing condition are effectively improved.

The invention relates to the technical field of computer vision, and particularly provides a virtual digital human generation method and electronic equipment, and the method can comprise the steps: obtaining three-dimensional human body modeling data corresponding to each frame of image in a monocular video sequence of a target object; the three-dimensional human body modeling data comprises body posture data and identity offset data; processing the three-dimensional human body modeling data by using a pre-constructed to-be-trained virtual model to generate a virtual object image of each frame of image; the to-be-trained virtual model comprises a to-be-optimized three-plane feature body and a plurality of multi-layer neural network MLP models; optimizing the to-be-trained virtual model through the loss value between the virtual object image and each corresponding frame image, and obtaining a three-dimensional digital human model corresponding to the target object; the three-dimensional digital human model is used to generate a virtual digital human based on any drive data. According to the embodiment of the invention, the generation and rendering quality of the virtual digital human can be improved.

The present invention discloses a three-dimensional human body modeling and motion acquisition method based on multi-view synchronous shooting. A tracking video of a person shot synchronously from multiple perspectives is obtained, and two-dimensional human body key points and two-dimensional hand key points are extracted from each frame; then, the arm parameters of the human body parametric model are optimized using the two-dimensional human body key points. Afterwards, the parameters of the hand model are optimized based on the two-dimensional hand key points; according to the wrist position, the arm position is corrected to obtain the arm parameters of the human body parametric model. By arranging the hand model point cloud and the human body parametric model point cloud, an index of points and triangular facets is established to obtain a fusion model, and the parameters of the human body parametric model and the hand model are imported into the fusion model to obtain the result of human body reconstruction. The present invention can use multi-view images to perform human body modeling and motion acquisition including hands, and does not require a sophisticated multi-camera configuration and motion capture clothing.

The invention provides a human body posture data set construction method based on a large model. A large language model is combined with a fine-grained attribute database to generate motion description, then the motion description is converted into 3D human skeleton motion data through a MoMask model, and quality control is carried out. And then view control and gait speed control are performed on the motion data to increase data diversity. And then radar data simulation is carried out, including human body modeling, RCS estimation, radar signal generation and processing, and micro-Doppler spectrogram generation. Meanwhile, a VQ-VAE model is adopted to generate background noise and add the background noise, shielding is simulated, and the data trueness is enhanced. And finally, constructing a data set containing a plurality of activities. The data set generated by the method has semantic and physical diversity, a real scene can be simulated with high fidelity, the performance of a human body posture recognition model is effectively improved, and powerful support is provided for application of related technologies in multiple fields.

The present invention discloses a modeling method of neural radiation field and two-dimensional volume density segmentation based on Hashgrid, which includes the following modeling steps: step 1, firstly, collecting human body images from multiple angles through an ordinary RGB camera; step 2, using colmap to perform camera intrinsic parameter calibration and extrinsic parameter calibration on the collected images; step 3, performing pixel segmentation and image preprocessing on the collected images, and removing the background area outside the image to be reconstructed. The present invention collects images of the human body to be reconstructed from various angles, sets the overlapping area between adjacent images, removes the background, and reduces interference. Through multiple iterative training, the rendering is changed from continuous to discrete form, which can quickly perform three-dimensional human body modeling. By substituting sampling points into the calculation, the rendering effect is improved, and the volume density slice map is segmented using u-net, and the noise is removed while the segmentation is performed, the human body is accurately reconstructed, and the modeling efficiency is effectively improved.

The application discloses a 3D human body modeling method and device based on contrast learning and a medium. The method comprises the following steps: acquiring 2D image information corresponding to a user and body state information corresponding to the user; extracting original features corresponding to the 2D image information through a shared encoder; performing mask prediction and determining clothing coverage rate according to the original features; performing feature fusion according to the original features, shape prior features corresponding to the body state information and the clothing coverage rate to obtain fused features; performing adaptive feature purification on the fused features according to the clothing coverage rate to obtain pure features; and performing human body parameter prediction according to the pure features.

The present application provides a physiological mechanism driving and data fusion method and system for seated human body modeling, which relates to the cross technical field of digital twin and biomechanics. The method comprises: mapping human macroscopic parameters to biomechanical attribute parameters; generating a basic three-dimensional human body grid based on human macroscopic parameters; generating a deformed three-dimensional human body grid by combining the basic three-dimensional human body grid and the human macroscopic parameters; solving the optimal posture based on the vehicle geometric parameters, the task scene identifier and the biomechanical attribute parameters; finally integrating the deformed three-dimensional human body grid and the optimal posture to generate a high-fidelity seated digital twin, realizing end-to-end automatic modeling from macroscopic parameters to digital twin, and solving the core problems of shape and posture mismatch, insufficient population coverage and poor task adaptability.

The application relates to the technical field of medical devices and particularly discloses a pain dynamic evaluation and treatment linkage system based on three-dimensional somatosensory interaction, which comprises the following modules: a human body modeling module for constructing a three-dimensional human body dissection model; a somatosensory interaction module for receiving pain area marking instructions and pain intensity scores; an infrared sensing module for monitoring human body infrared thermal imaging data of a patient in real time; a physiological monitoring module for monitoring heart rate variability indexes and skin electric reaction indexes of the patient based on wearable devices; a region marking module for obtaining a calibrated pain area of the patient based on the pain area marking instructions and abnormal temperature zones in the human body infrared thermal imaging data; a pain prediction module for predicting a pain intensity prediction curve in a future period based on the pain intensity scores, the heart rate variability indexes, the skin electric reaction indexes and the calibrated pain area of the patient; and a treatment linkage module for generating treatment parameter control instructions based on the pain trend prediction curve and a multi-dimensional pain characteristic vector, so that precise and dynamic treatment intervention is realized.

The present invention relates to the technical field of human body modeling, and in particular to a cross-domain unified human body modeling method, apparatus, equipment and storage medium. The present invention uses a relative posture similarity space to represent the relative similarity between any two skeleton postures of the human body, and then extracts the relative similarity between the unsampled skeleton posture and the anchor prompt from the relative posture similarity space. Based on the relative similarity, the target posture is screened out from the unsampled skeleton posture, and the anchor prompt set is updated with the target posture to obtain a final set of anchor prompts. Finally, the anchor prompts and soft prompts in the final set of anchor prompts cooperate with each other to complete the reasoning task of the model. Since the soft prompts are based on the parameters generated when the model is trained for each reasoning task, when processing different reasoning tasks, it is only necessary to replace the soft prompts corresponding to the reasoning task to cooperate with the anchor prompts to complete the reasoning task without replacing the model. Therefore, the present invention can achieve unified modeling across tasks.

The invention discloses a convenient three-dimensional human body reconstruction method based on a monocular video, and the method comprises the steps: constructing a parameterized human body model low-dimensional shape space, and obtaining the approximate representation of a linear model. And extracting a multi-view key frame from the input human body rotation video. And extracting a feature vector of each view key frame by using an improved ResNet-50 network fused with an attention mechanism. And defining the multi-view feature vectors as graph nodes, and constructing a graph structure. And performing cross-view depth feature fusion through the graph convolutional network to obtain global features. And mapping the global features to a parameterized human body model low-dimensional shape space. And reconstructing a complete three-dimensional human body grid by using linear model approximate representation. The method has the advantages that multi-view feature extraction and a graph structure fusion mechanism are combined, the end-to-end process from consumption-level video input to high-precision three-dimensional human body reconstruction is achieved, certain generalization ability is achieved in a real scene, and a low-cost and high-precision three-dimensional human body modeling solution can be provided for virtual fitting, digital twinning and other applications.

The present invention relates to the field of computer vision technology, and provides a human body depth estimation method, device, electronic device, and storage medium. The method includes: obtaining human body images of a person to be estimated from multiple camera perspectives, and accordingly performing three-dimensional human body modeling to obtain three-dimensional human body models corresponding to the human body images from multiple camera perspectives; performing projection based on the three-dimensional human body models, and compressing and outputting the depth maps of the person to be estimated obtained by the projection from multiple camera perspectives, which solves the problem that the depth cannot be restored in the current weakly textured regions, resulting in holes in the depth map, and can effectively solve the depth information of the weakly textured regions, thereby ensuring the accuracy and integrity of human body depth estimation. By compressing and outputting the depth maps, not only the number of output files is reduced, the transmission speed is increased, the transmission space occupancy is reduced, but also the integrity and orderliness of the information are improved, realizing the effective simplification of the output results and facilitating access.

The invention discloses a 3D human body modeling method and device based on comparative learning and a medium. The method comprises the steps that 2D image information corresponding to a user and posture information corresponding to the user are acquired; extracting an original feature corresponding to the 2D image information through a shared encoder; according to the original features, mask prediction is carried out, and the clothing coverage rate is determined; performing feature fusion according to the original feature, a shape prior feature corresponding to the posture information and the clothing coverage rate to obtain a fused feature; according to the clothing coverage rate, performing adaptive feature purification on the fused feature to obtain a pure feature; and performing human body parameter prediction according to the pure features.

The invention discloses a complex human body modeling analysis method based on human factor engineering, and belongs to the technical field of human body modeling. According to the method, firstly, a joint DOF model is constructed, a virtual human body model is divided into a plurality of limb segments for modeling according to the joint DOF model and the membership relation between different body parts and joints, the level of the whole limb segment is regarded as a tree structure, a single limb segment is regarded as a tree node, and then a human body limb segment hierarchical model is formed; performing parametric modeling on the virtual human body based on the human body limb section hierarchical model to obtain a virtual human body DOF model; after a virtual human body DOF model is established, a virtual human body posture is analyzed, a D-H method is used for kinematics analysis, an inter-joint transformation matrix based on human kinematics is obtained, a human body reachable domain range is determined based on the inter-joint transformation matrix, and then complex human body modeling for assembling or maintaining in a narrow space is achieved.

The invention relates to the technical field of three-dimensional human body modeling, and particularly discloses a modeling and rendering method and device based on layered three-dimensional Gaussian representation and a medium, and the method comprises the steps: obtaining an image sequence of a target, and initializing a three-dimensional Gaussian point in a standard space based on an SMPL human body template vertex; taking the initialized three-dimensional Gaussian point as the first layer of the innermost side, constructing a hierarchical Gaussian structure from coarse to fine, and establishing conditional dependence among the layers through a father-child node connection mode; setting a hierarchical motion condition, and predicting rigid body and non-rigid body deformation transformation parameters of each Gaussian point by using a rigid body transformation network and a non-rigid body transformation network; and mapping the layered Gaussian points to an observation space in a unified manner according to predicted transformation parameters, and performing image synthesis by using a three-dimensional Gaussian rendering technology. Compared with the prior art, the method can naturally adapt to and express deformation characteristics of different levels in human motion, and realizes high-quality reconstruction and real-time rendering of a complex dynamic scene.

The invention discloses an intelligent fitting recommendation system based on 3D human body modeling and clothing parameter adaptation, and relates to the technical field of intelligent fitting recommendation. The system generates a high-precision dynamic body surface grid through a multi-modal data acquisition module; a human body-costume double-layer decoupling digital model construction module is utilized to realize efficient and realistic costume dynamic deformation simulation; through a dynamic pressure simulation and evaluation module, in combination with an improved Hertz contact theory and a fabric mechanical model, carrying out accurate simulation and multi-dimensional quantitative evaluation on dynamic contact pressure of clothes and a human body; and finally, the scene intelligent recommendation module provides personalized clothing recommendation according to the scene selected by the user and the dynamic evaluation result. According to the method, the problem that traditional virtual fitting lacks dynamic comfort evaluation is solved, the user experience is remarkably improved, the return rate of merchants is reduced, and the method has important market application value.

The invention discloses a shopping mall virtual fitting room implementation method and system based on the AR technology, and belongs to the technical field of e-commerce and computer vision, and the method comprises the steps: carrying out the real-time high-precision three-dimensional human body modeling based on a common camera; performing intelligent cloth physical simulation fused with a machine learning algorithm; multi-terminal cross-platform consistency rendering is supported; and carrying out personalized clothing recommendation in combination with user body type data. The virtual try-on effect of the user can be truly presented, multi-device terminal access is supported, the shopping experience is remarkably improved, and the refunding and changing rate is reduced.

The invention relates to the technical field of costume design, and discloses an intelligent interactive costume design system and method based on a data model. The system comprises an interaction module, a human body modeling module, a demand intelligent analysis module, an intelligent design generation module, a fabric matching module, a virtual fitting module and a scheme optimization module. The interaction module obtains PLY format human body data of a user and costume design requirements and receives modification feedback of the user, and the human body modeling module generates a 3D human body model, human body size parameters and posture feature tags. The demand intelligent analysis module converts a costume design demand into a structured demand document containing a style tag and a version constraint parameter in combination with a posture feature tag, the intelligent design generation module outputs an optimal costume design scheme through a VAE generation model, and the fabric matching module matches candidate fabrics based on the optimal costume design scheme. The virtual fitting module realizes dynamic fitting simulation, and the scheme optimization module iteratively optimizes the optimal costume design scheme to meet user requirements. The personalized costume design efficiency of the user is improved.

The embodiment of the invention relates to the technical field of artificial intelligence, and discloses a digital person business card generation method, electronic equipment, a medium and a program product. The method comprises the following steps: acquiring an input image of a target user in digital person business card generation, and preprocessing the input image to obtain image feature data; performing three-dimensional space coordinate alignment and mapping on the image feature data to obtain image three-dimensional data; reconstructing the image three-dimensional data into a three-dimensional Gaussian point cloud; establishing a binding relationship between the three-dimensional Gaussian point cloud and a skeleton hierarchical structure in the three-dimensional human body modeling tool; performing image rendering according to the binding relation to obtain a digital human model; generating corresponding digital human voice data according to the reply data in the digital human interaction process; and the digital human model and the digital human voice data are combined and displayed to obtain the digital human business card, so that a single-picture realistic digital human technology can be realized, and the fidelity of a digital human image is improved and the digital human image is deeply bound with a user image on the basis of reducing the data volume required for making the digital human.

The present invention relates to the technical field of human body modeling, and provides a method, a system, a device and a storage medium for constructing a human body model. The method includes: when the user to be modeled enters the first preset position, invoking the first image sensor array to collect the first image acquisition result; performing clustering analysis on the user to be modeled to obtain the clustering result of the modeling area; traversing the clustering result of the modeling area and invoking the first image acquisition result to obtain the outer frame modeling result; reminding the user to be modeled to enter the second preset position, invoking the second image sensor array to collect the second image acquisition result; traversing the clustering result of the modeling area and invoking the second image acquisition result to obtain the inner skeleton modeling result; constructing a human body static simulation model according to the outer frame modeling result and the inner skeleton modeling result; uploading the basic parameters of the user, and initializing the human body static simulation model to generate a human body dynamic simulation model. The technical effect of improving the fitness between the human body model and the requirements of the metaverse human body model is achieved.

The invention relates to the technical field of medical instruments, and particularly discloses a pain dynamic assessment and treatment linkage system based on three-dimensional somatosensory interaction, and the system comprises a human body modeling module which constructs a three-dimensional human body anatomical model; the somatosensory interaction module receives a pain area marking instruction and a pain intensity score; the infrared sensing module monitors human body infrared thermal imaging data of a patient in real time; the physiological monitoring module monitors a heart rate variability index and a skin electric response index of the patient based on the wearable device; the region marking module obtains a calibrated pain region of the patient based on the pain region marking instruction and an abnormal temperature region in the human body infrared thermal imaging data; the pain prediction module calibrates a pain area based on a pain intensity score, a heart rate variability index and a skin electric response index of the patient, and predicts a pain intensity prediction curve in a future period; the treatment linkage module generates a treatment parameter control instruction based on the pain trend prediction curve and the multi-dimensional pain feature vector; and accurate and dynamic treatment intervention is realized.

The embodiment of the invention provides a human body and object contact prediction method and device, electronic equipment and a storage medium. The method comprises the steps of obtaining a target image, wherein the target image comprises an interaction scene of a human body and at least one object; performing contact prediction processing on the target image to obtain a contact vertex list; performing target detection processing on the target image to obtain detection information; carrying out human body modeling processing on the target image to obtain a three-dimensional vertex and a corresponding two-dimensional projection coordinate; semantic fusion processing is conducted on the contact vertex list, the detection information and the two-dimensional projection coordinates, a semantic contact result is obtained, and the semantic contact result is used for representing the contact relation between all the parts of the human body and different objects; the semantic contact result is visualized, a final contact result is obtained, and the final contact result is used for visually representing the contact relation between the human body and the object. According to the method, the accuracy of the human body and object contact prediction result can be improved.

The invention discloses a three-dimensional human body modeling method and system based on binocular vision and a laser radar, solves the problem of poor reconstruction effect of a three-dimensional human body in a dynamic motion scene, and belongs to the technical field of computer vision, intelligent perception and three-dimensional modeling. A binocular camera and a laser radar are used for synchronously collecting a human body, and a binocular image and a laser radar point cloud are obtained; processing and fusing the obtained binocular image and the laser radar point cloud to generate a human body initial three-dimensional model; tracking and updating the parameterized human body model in real time, and outputting a dynamic three-dimensional human body model; in the real-time tracking and updating process, a parameterized human body model is constructed according to the human body initial three-dimensional model, the human body volume, the image contour and the point cloud distance are used as constraint terms, the human body elastic deformation is used as a physical regularization term to construct an optimized energy function, and the deformation of the parameterized human body model in motion is optimized by using the optimized energy function. According to the method, the geometric fidelity and the detail reduction degree of the three-dimensional model are effectively improved.