155 results about "Rendering (computer graphics)" patented technology

The invention discloses a large-scene three-dimensional reconstruction method based on three-dimensional Gaussian sputtering, and relates to computer graphics. The method comprises the following steps: collecting a multi-view image set of a large scene; obtaining a scene sparse point cloud according to the multi-view image set; performing monocular depth estimation on the multi-view image by using a pre-trained depth prediction network to obtain monocular depth estimation priori; the method comprises the following steps of: performing global training on a scene by utilizing scene sparse point cloud and monocular depth estimation prior to obtain an initial three-dimensional Gaussian model, and performing space grid division on the initial three-dimensional Gaussian model to obtain a plurality of scene blocks with axis alignment bounding boxes; setting image view angle data of each scene block; performing deep supervised training on the Gaussian ellipsoids in the plurality of scene blocks by using a parallel GPU (Graphics Processing Unit); combining the trained scene blocks to obtain a final three-dimensional Gaussian model; in view of low geometric structure reconstruction precision caused by only depending on color information of a multi-view image in large-scene three-dimensional rendering, the method improves the reconstruction precision of large-scene rendering.

The invention discloses a global nerve drawing method and system for a multivariate mixed representation scene, and belongs to the technical field of computer graphics, and the method comprises a scene expression module which is used for loading multivariate mixed scene data; the geometric module is used for generating a multi-channel geometric buffer based on the multivariate mixed scene data; the illumination module comprises at least one neural network sub-module and is used for carrying out illumination calculation based on multi-channel geometric buffering; the scheduling and synchronizing module is used for coordinating cross-module data streams through a rendering graph structure and executing resource merging and task scheduling between the geometric module and the illumination module based on a unified request structure; and the merging output module is used for fusing neural network illumination output and traditional rendering output to generate a final frame image so as to complete global neural rendering. According to the invention, multi-element scene data is integrated through modular design, efficient collaboration of traditional and neural rendering is realized by using a unified interface and a scheduling structure, and the image quality and the system flexibility are improved.

The invention discloses a point cloud model rendering optimization method and system based on cloud edge collaboration, and belongs to the field of computer graphic processing, and the method comprises the steps: carrying out the spatial partitioning of multi-stage point cloud data, generating a point cloud hierarchical structure which can be quickly accessed, building a user portrait model, and predicting the region of interest of the user portrait model; the multi-level LOD point cloud data structure is distributed to an edge computing node, the current network state is evaluated in real time, and if the local cache of the edge node cannot meet the rendering requirement, the edge node adaptively requests for cloud incremental data; the client receives the data blocks from the edge nodes and uploads real-time interaction data to the edge nodes; and the client interaction record and the edge node cache hit rate are transmitted back to the cloud end, and the cloud end performs iterative optimization on a point cloud data distribution strategy by using a reinforcement learning model. According to the method, priority loading and high-precision rendering of the region of interest are achieved by constructing the multi-level LOD point cloud data structure, the delay and jamming phenomena are effectively avoided, and the user interaction experience is improved.

A full-automatic movie and television video production system based on digital actors belongs to the field of artificial intelligence, computer graphics and movie and television production and comprises a creative engine unit, a production unit, an intelligent shooting unit, a post-processing unit and a quality inspection and rendering unit. The method has the advantages that 1, the manufacturing efficiency is greatly improved, the traditional film and television manufacturing period is usually 6-12 months, the manufacturing period can be shortened to 4 hours by the method, and the efficiency is improved by about 1000 times; 2, the manufacturing cost is obviously reduced: the manufacturing cost of traditional movies and other movies is usually 5 million to 20 million dollars, but the method only needs the computing power cost, and the cost is less than one percent of that of the traditional movies; 3, the style consistency is guaranteed, traditional film and television production depends on personal ability, the style consistency is difficult to guarantee, and the style consistency is strictly guaranteed through an algorithm and is improved by 100%.

The invention belongs to the technical field of computer graphics and real-time 3D scene reconstruction, and relates to a real-time compression and optimization method based on a 3D Gaussian sputtering (3DGS) model, which realizes efficient storage, transmission and rendering of a three-dimensional scene through residual vector quantization (RVQ) and a dynamic pruning technology, and comprises the following steps: S1, residual quantization vector quantization; s2, enhancing a dynamic pruning strategy; and S3, index compression and coding optimization. The invention provides a real-time 3D Gaussian sputtering (3D Gaussian sputtering) model compression system combining residual vector quantization and dynamic pruning, which can realize high-quality real-time rendering and extremely high storage efficiency. Through three core technologies of a vector quantization strategy, dynamic pruning strategy enhancement and index compression and coding optimization, on the premise that the rendering quality is guaranteed, the purposes that the model storage amount is reduced by more than 45 times, and the rendering speed is increased by 3 times are achieved.

The invention relates to the technical field of computer graphics, and provides a rendering method and device for a static scene under a dynamic view angle, equipment and a medium, and the method comprises the steps: obtaining a future view angle probability cloud picture of a preset frame number according to the historical behavior data of a user and a hybrid neural network view angle prediction model; dividing a pre-rendering static scene of at least one target pre-rendering view angle based on a preset dynamic semantic weight matrix to obtain layered baking data; obtaining target optical flow data of a target adjacent view angle adjacent to the current view angle from a preset optical flow mapping table; during view angle switching, dynamically mixing a static scene rendered in real time at the current view angle and a static scene pre-rendered according to a pixel displacement relation based on target optical flow data to obtain scene mixed data; and rendering based on the scene mixed data to obtain a target static scene. The problem of real-time high-quality rendering of the static scene under the dynamic view angle can be solved.

The invention discloses a general nerve drawing method and system based on an illumination function generation model, and belongs to the technical field of computer graphics, and the method comprises the steps: collecting light source information containing multi-view observation data, and collecting scene information; constructing an illumination function generation model comprising a light source coding module and a light source decoding module for converting the light source information into neural illumination representation and performing joint inference based on the neural illumination representation and scene features of the drawing points; and generating a final drawn image conforming to the real illumination distribution based on the inference result. According to the method, generalization neural drawing across light sources and scenes can be achieved, images with off-line rendering quality can be generated at the cost close to real-time calculation, the sense of reality, stability and rendering efficiency are considered, and the method has good expansibility and wide application value.

The invention belongs to the technical field of computer graphics and virtual reality, and discloses a virtual-real fusion-based panoramic roaming construction method for a digital exhibition hall, which effectively solves the problems of easy drift and offset in a virtual environment in a traditional system through a multi-sensor fusion algorithm and a closed-loop optimization mechanism. Under the complex environment of people flow change, illumination fluctuation and the like in an exhibition hall, accurate alignment of virtual content and physical space can still be kept, it is ensured that a user obtains coherent and stable visual experience in the roaming process, and a reliable space reference is provided for virtual-real interaction; the light and shadow effect and the material performance of the virtual content are naturally fused with the physical environment through the physical rendering and multi-sensory interaction technology, and meanwhile, a multi-dimensional immersion sense is constructed in combination with the spatial sound effect and interaction feedback; when the user interacts with the virtual exhibit, sensory experience close to reality can be obtained, and the sense of substitution and attraction of the digital exhibition hall are remarkably improved.

The invention discloses an SPH fluid rendering method and system based on a Gaussian function, and belongs to the technical field of computer graphics, and the method comprises the steps: receiving SPH simulated fluid particle data, and constructing a particle connection graph based on a particle neighborhood relation; based on the particle connection graph, extracting local geometric features of each particle through a neural network and aggregating the local geometric features into global geometric features; analyzing the global geometric features into Gaussian function parameters to generate a plurality of Gaussian functions for each particle, and superposing all Gaussian functions in space to form an implicit fluid surface; performing screen space rendering on the implicit fluid surface to obtain a final fluid rendering image; and after foam particles generated by SPH simulation are also mapped into a Gaussian function, a fluid rendering image containing a foam visual effect is obtained through screen space rendering. The method can improve the smoothness and reality of the rendering result while guaranteeing the calculation efficiency, and achieves the consideration of the efficiency and quality of fluid rendering.

The invention relates to the technical field of computer graphics and virtual reality, and discloses a meta-universe social space generation method and system. The method comprises the following steps: acquiring multi-modal input data of a user and constructing an intention vector; extracting social relation characteristics based on the intention vector, and generating a multi-dimensional scene constraint condition; generating a three-dimensional virtual environment and optimizing a topological structure by adopting a generative adversarial network according to a scene constraint condition; and fusing the scene data with the user social behavior data, and outputting a multi-terminal interactive rendering result. The system comprises a data acquisition and intention modeling module, a scene constraint mapping and environment generation module and a fusion rendering and interactive output module. According to the invention, multi-user immersive social space interaction can be realized on the VR head-mounted display, the AR glasses and the mobile terminal, and the method has the advantages of low delay, high simulation degree and cross-platform synchronization.

The invention discloses a dynamic rendering method and system for virtual reality, a storage medium and electronic equipment, and relates to the technical field of virtual reality, augmented reality and computer graphics. Therefore, the frequency and time sequence change of the sight focus of the user hitting the surface of the virtual object in the virtual scene are analyzed, so that the relative motion trend and observation intention of the user and the object are pre-judged, predictive and adaptive scheduling of rendering resources is realized, and optical feedback conforming to the physical law, namely, dynamic material, is triggered at the sight focus, so that the real-time performance of the user is improved. Therefore, the interaction fineness in the process of interaction between the user and the VR system is improved.

The invention belongs to the technical field of computer graphics and cloud computing, discloses an end-cloud collaborative rendering method and system based on mixing of an object stream and a video stream, and solves the problems of low collaborative rendering efficiency of the object stream and the video stream, heavy transmission load and insufficient dynamic interaction real-time performance in a mixed stream rendering scene. By researching the influence of user behaviors on the priority of rendering tasks, the system adopts a dynamic rendering task allocation method based on user behavior requirements to adjust the priority of the rendering tasks in real time, allocates high-priority tasks to a client for processing, and submits low-priority tasks to a cloud for rendering. Therefore, the real-time performance of the key rendering content is ensured, and the calculation burden of the client is effectively reduced. In combination with a fine granularity optimization strategy of rendering pipeline splitting, the system distributes tasks with complicated calculation to the cloud for execution according to the calculation intensity of the tasks, and the tasks with relatively high real-time requirements are left at the client for completion, so that the efficiency of end-cloud collaborative rendering is improved.

The invention discloses an automatic logistics digital twinning model adaptive precision rendering method and system, and relates to the technical field of computer graphics and automatic logistics digital twinning crossing, in particular to a rendering optimization scheme capable of dynamically adjusting model rendering precision for a large-scale smart factory scene in a Unity engine. Through a dynamic decision-making mechanism based on a use scene and a service factor, in combination with a current use scene, a model with a high service priority in the scene is subjected to refined rendering, and a model with a low current weight in the scene is subjected to simple rendering, so that it is ensured that under each specific scene, the service quality of the scene is improved. And the system can intelligently and accurately guide the computing resources to the object which most needs high-fidelity rendering.

The invention provides a view rendering method and device based on a two-dimensional Gaussian splash model, and electronic equipment, and belongs to the technical field of computer graphics and machine learning, and the method comprises the steps: calculating an intersection point of light emitted from a pixel point in an original image of each viewpoint and each two-dimensional Gaussian primitive, aiming at each viewpoint, obtaining a two-dimensional Gaussian splash model; rendering a rendering view of the viewpoint according to each two-dimensional Gaussian primitive, a preset object space filter and the plurality of intersection points of the viewpoint, and determining a rendering loss value of the viewpoint according to the original image of the viewpoint, the monocular normal graph, the rendering view and each intersection point of the viewpoint, according to the number of pixels of the original image covered by each two-dimensional Gaussian primitive, the depth scaling factor and the rendering loss value of each viewpoint, obtaining a region sensing gradient of each two-dimensional Gaussian primitive, and splitting the two-dimensional Gaussian primitive according to the optimization round and the region sensing gradient, the two-dimensional Gaussian primitive can accurately render detail information of a scene, so that the quality of view rendering is improved.

The invention relates to the technical field of computer graphics, in particular to a three-dimensional geological section real-time cutting method and system and a storage medium, and the method comprises the steps: constructing a BVH tree to accelerate the cutting of a three-dimensional geological model, and calculating the projection of intersecting line segments on a cutting surface two-dimensional coordinate system to obtain a projection line segment set; constructing an undirected graph structure based on the projection line segment set; and extracting a closed profile contour based on the undirected graph structure, subdividing a triangular grid, and rendering and displaying. The implementation of the scheme can break through the traditional preprocessing dependence, and improve the profile generation efficiency and the visualization integrity.

The invention discloses a 3D Gaussian splash rendering optimization method based on a Unity engine in VR, and relates to the technical field of computer graphics. PLY point cloud files are converted into SOG high-compression formats containing multi-level LOD information, after low-transparency Gaussian points are removed, Morton coding sorting is carried out on the center coordinates of the Gaussian points, and then the 3D Gaussian splash rendering optimization method based on the Unity engine in VR is realized. The method comprises the following steps: constructing an octree spatial index containing a spatial bounding box and multiple LOD data, removing Gaussian points in a grading manner according to a camera distance, selecting an LOD hierarchy, carrying out deep barrel sorting on visible Gaussian points, executing drawing by taking a barrel as a unit, and combining near-to-far sorting, an improved mixed equation and screen space edge removal and amplification operation to obtain a target image. According to the method, through multi-dimensional optimization such as loading, space elimination and rendering, the loading efficiency of the VR equipment is improved, the rendering burden is reduced, Tile-Based hardware is adapted, and smooth operation of 3D Gaussian splashing on the VR all-in-one machine is realized.

The invention belongs to the technical field of crossing of computer graphics and physical simulation, and discloses a furniture material dynamic rendering system based on physical simulation. Comprising a furniture modeling module used for constructing initial scene data of furniture to be analyzed; the physical simulation module is used for constructing a structure physical simulation model of the to-be-analyzed furniture based on the initial scene data and executing simulation to obtain structure response data of each structure position of the to-be-analyzed furniture; the multi-field generation module is used for mapping the structure response data into a three-dimensional geometric model of furniture to be analyzed to form a surface risk field, generating a joint priority field in combination with scene visibility data, and determining a grid reconstruction weight field based on the structure response data and the surface risk field; the texture generation module is used for acquiring physical driving texture data; and the rendering and feedback module is used for carrying out material coloring strategy distribution and outputting a current frame of furniture material dynamic rendering image so as to realize optimal distribution of rendering quality.

The invention discloses a frame generation method and system adopting predictive sparse rendering, and belongs to the technical field of computer graphics, and the method comprises the steps: obtaining historical frame rendering data, current frame geometric rendering data and current frame sparse rendering data in a graphic rendering pipeline, and extracting a multi-scale feature pyramid through different encoders; receiving and fusing the multi-scale feature pyramids through a unified decoder, and generating a frame synthesis residual error, an extrapolation error score and a time sequence mixed weight through multi-stage up-sampling; performing time sequence synthesis by using the time sequence mixed weight and fusing the frame synthesis residual error to obtain a final rendering result of the current frame, and updating the final rendering result into historical frame rendering data of the next frame; and obtaining a next-frame sparse rendering mask by using the extrapolation error score and optical flow prediction, and generating next-frame sparse rendering data from the next-frame sparse rendering mask through tile-level sparse rendering. According to the method, the frame generation quality and the time stability can be remarkably improved under the condition of low rendering budget.

The invention discloses a three-dimensional realistic real-time simulation method suitable for jacquard warp-knitted fabric, and belongs to the technical field of computer graphics. Comprising the following steps: analyzing an execution signal file of an industrial jacquard knitting machine and translating the execution signal file into yarn laying numerical codes; modeling a single three-dimensional coil model on the basis of a Catmull-Rom spline curve; calculating the arrangement and tension simulation of each coil of the whole cloth according to the yarn laying numerical code; gPU hardware is used for accelerating generation of fiber-level details; simulating a shadow shielding phenomenon in the cloth; and performing physics-based rendering on the fiber-grade model on the basis of the bidirectional curve scattering distribution function. The problems that a traditional three-dimensional simulation method is long in consumed time and low in speed when large jacquard cloth is subjected to three-dimensional simulation are solved, and the problems that the trueness is low and the credibility is low when three-dimensional knitted cloth is subjected to simulation rendering are solved.

The invention relates to a virtual fabric rendering method and device, equipment, a storage medium and a program product, and relates to the technical field of computer graphics. The method comprises the steps of obtaining rendering attribute information of a virtual fabric model, wherein the rendering attribute information comprises a fabric material; the virtual fabric model at least comprises a model part corresponding to the fabric, and the fabric material is the material of the model part corresponding to the fabric; rendering configuration parameters are obtained based on the rendering attribute information fabric materials, the rendering configuration parameters are configuration parameters used when the virtual fabric model is rendered, and at least two fabric materials correspond to different rendering configuration parameters; and calling a rendering tool to create a virtual fabric model, and setting rendering configuration parameters in the rendering tool, so that the rendering tool renders the virtual fabric model based on the rendering configuration parameters to obtain at least one frame of rendered picture. The rendering process of the virtual fabric model is optimized for the fabric material, and the rendering quality of the virtual fabric is improved.

The invention discloses a digital human video rendering method and device and a medium, and relates to the crossing field of computer graphics and generative adversarial networks, and the method comprises the steps: constructing a generative adversarial model based on a generative adversarial network of a single fusion architecture; performing multi-modal preprocessing on the original audio and video data based on the digital human reference image; extracting dual-granularity speech features through a speech feature extraction module of a generative adversarial model, and fusing the dual-granularity speech features; determining a local deformation field of the digital human reference image in a UV parameterized space based on a reference key point corresponding to the digital human reference image and the fused speech features; sampling the identity texture of the digital human reference image to generate a target digital human face image; and verifying the digital human face image based on multiple scales through a discriminator of a generative adversarial model. Through dual-granularity speech feature fusion and generator multi-resolution injection, deep alignment of speech semantics and facial actions is realized.

The invention discloses a virtual reality level simulation method based on a material point method and gridding surface reconstruction, and belongs to the crossing field of computer graphics, physical simulation and virtual reality technologies. The method comprises the following steps: S1, classifying and storing position coordinates of material points according to material numbers of the material points, and traversing the coordinates to obtain a boundary of a bounding box of the material points; s2, selecting a grid size and a boundary extension distance for reconstructing the surface in a self-defined manner, and generating a grid; s3, selecting a Gaussian function to calculate the contribution of each substance point to the overall substance point density field, and obtaining parameters; s4, calculating the contribution of each substance point to the overall substance point density field, and performing accumulation to obtain the overall substance point density field; and S5, obtaining a contour surface of the overall material point density field, and exporting the contour surface as a rendering engine general three-dimensional model to obtain a simulation visual effect. According to the method, the problems that in the prior art, tens of millions of substance points cannot be processed, reconstruction parameters are difficult to select, and connection with a high-end rendering engine is not smooth can be solved, and the virtual reality level simulation visual effect is achieved.

The invention relates to the technical field of computer graphic processing, in particular to a VR display dynamic parallax optimization method based on eyeball tracking, which comprises the following steps: S1, dynamically repairing eyeball tracking data; s2, eyeball movement bimodal classification: identifying a microtremor mode and a fast eye movement mode through spectrum energy distribution; s3, parallax compensation driven by physiological features: predicting a trajectory based on tremor statistical features in a microtremor mode, predicting a drop point based on a motion vector and a head posture in a rapid eye movement mode, and generating a radial attenuation weight field combined with retina biological characteristics; and S4, dynamic parallax rendering correction: performing pixel displacement on the left and right views according to the predicted coordinates and the weight field, and repairing a texture fracture region caused by displacement based on the scene depth map. Through accurate eyeball tracking and dynamic parallax optimization, discomfort caused by eye movement tracking errors or parallax errors is reduced. The method can adapt to physiological features of different users, and the practicability of the virtual reality technology is improved.

The invention provides a cartoon digital human real-time expression driving method, electronic equipment and a storage medium, and relates to the technical field of terminals. The method comprises the following steps that: terminal equipment firstly obtains target images containing face information of a target user according to a preset frequency through a camera, and then inputs each target image into a FaceMesh network frame by frame to carry out 3D reconstruction processing on the face image of the target user so as to obtain face 3D key point information of the target user. Inputting the face 3D key point information of the target user into the cartoon role expression driving network model, and predicting to obtain a cartoon role expression base coefficient and rotation feature representation; the cartoon role expression base coefficient and the rotation feature representation are utilized, a preset BlendShape parameter model is combined, rendering is carried out according to related knowledge of computer graphics, a cartoon role expression animation image corresponding to the target image is obtained, and the cartoon role expression animation image is continuously output. And the cartoon digital human real-time facial expression driving animation corresponding to the target user is formed.

The invention belongs to the field of computer graphics, and discloses a self-adaptive LOD subdivision method and system oriented to virtual reality and based on a perceptual model. The method comprises the following steps: firstly, researching factors influencing visual perception, establishing a perception model BD-castleCSF fused with stereoscopic vision by combining binocular parallax, and dynamically dividing a rendering region; then, on the basis of a discrete LOD initial order processing grid model, introducing a mosaic technology to execute grid subdivision, calculating a downsampling factor according to a perception model, dynamically adjusting the subdivision degree, and proposing a perception adaptive subdivision LOD algorithm; experimental results show that compared with a traditional static rendering method, the LOD switching frequency of the method is reduced by 67%, the effective subdivision triggering rate is improved by 19.9%, the redundant triangular surface compression rate is improved by 11.9%, and the frame time standard deviation is reduced. Therefore, the method can dynamically and smoothly adjust the model details under different watching conditions, guarantees the consistency of visual effects, and remarkably improves the rendering efficiency. According to the method, an efficient and adaptive rendering solution is provided for the field of virtual reality.

The invention relates to the technical field of crossing of computer vision, computer graphics and digital fashion, and discloses a three-dimensional digital clothing stylization method, which comprises the following steps of: firstly, constructing a double-encoder color mapping network to realize color alignment of a clothing image and a style image; the problem that training data is difficult to obtain is solved through a self-supervision strategy; then extracting depth features of the multi-view garment image and the style image by using a pre-trained VGG network, and embedding the style features into a garment feature space by means of an edge enhancement nearest neighbor feature matching algorithm; secondly, introducing an attention-guided edge enhancement feature extraction network, extracting edge features, and performing alignment optimization; finally, a stylized three-dimensional digital clothing image which has high style fidelity and clear structure details and supports rendering at any viewing angle is generated, the problems of color distortion, fuzzy details, inconsistent styles and the like in the prior art are effectively solved, and the method is suitable for the fields of virtual fitting, digital fashion design and the like.

This invention relates to the field of computer graphics data processing technology, specifically a method for hot-plugging dynamic outer contour meshes driven by three-dimensional spatial domain and time slices. The steps include: S1, receiving air traffic control message information in real time via a transmission protocol, converting geographic coordinates into target world coordinates, and simultaneously generating management files; S2, constructing a spatial domain mesh in real time based on the four types of management files using a combination of outer contour extrusion and ear-cutting methods; S3, mapping the spatial domain ID and time slice from the message to the spatial domain mesh constructed in S2, constructing a double-key hash index of spatial domain ID and time slice, which serves as the sole basis for the hot-plug manager to determine whether to update or create a mesh, dynamically executing spatial domain mesh hot-plugging. Employing the contour extrusion-ear-cutting method and vertex cache reuse mechanism, combined with the "spatial domain ID and time slice" double-key hash index, it supports dynamic construction and hot-plugging operations of the mesh model, meeting the low-latency rendering requirements of real-time air traffic control data.

The invention relates to the technical field of computer graphics, in particular to a method for collecting material reflection data in a low-bandwidth environment, and the method comprises the steps: obtaining a target object image, and calling a pre-stored target terminal sampling network to map the target object image into a corresponding target hidden variable; transmitting the target hidden variable to a cloud server to reconstruct material data of the target object, and rendering a reference schematic diagram of the material of the target object based on the reconstructed material data of the target object; wherein the cloud server calls a decoder in a pre-stored target variational auto-encoder to reconstruct target object material data; and receiving a reference schematic diagram fed back by the cloud server, and if the reference schematic diagram meets user requirements, taking the reconstructed target object material data corresponding to the reference schematic diagram as reflection data of the target object material. Therefore, the problems that high-quality material reflection data is difficult to generate conveniently and efficiently due to the limitation of a working scene, and the picture rendering effect is reduced are solved.

The invention belongs to the technical field of computer vision and computer graphics, and discloses a digital human-oriented four-dimensional virtual fitting method and system. The method comprises the steps of obtaining a source figure image sequence and a target clothing image; a parallel double-branch structure is adopted to generate a first Gaussian mapping and a second Gaussian mapping irrelevant to view angles and postures under the standard postures; generating a Gaussian parameter offset according to the driving attitude and the visual angle through a shared nonlinear deformation module, and superposing the Gaussian parameter offset to Gaussian mapping; performing time sequence consistency correction on the two-dimensional virtual fitting sequence to obtain a correction supervision signal; and generating a four-dimensional virtual fitting result through micro rendering based on the correction supervision signal and the Gaussian mapping after superposition offset. Compared with the prior art, the method has the advantages that the multi-view consistency, the dynamic continuity, the rendering quality, the generation speed and the like are obviously superior, high-fidelity dynamic virtual fitting is realized, and advanced performance is achieved in the aspects of visual perception indexes, physical stability, cross-garment generalization ability and the like.

The invention discloses an end-to-end field optimization surface reconstruction method based on Gaussian sputtering guidance, and belongs to the field of computer vision and computer graphics. Comprising the following steps: 1) Gaussian guidance field optimization: defining an implicit symbol distance field by using a parameterized MLP network, and initializing and optimizing a surface through a microcosmic contour surface extraction method; gaussian units are bound with the extracted surface triangular patches, the positions of the patches are parameterized through the vertexes of the patches, rotation and scaling are carried out, regularization constraints based on the minimum wrapped ellipse and the distribution entropy are introduced, it is ensured that the Gaussian units are accurately attached to the surface and evenly distributed on the surface, and a result obtained through rendering guides field optimization. And 2) local multi-resolution: identifying surface voxels based on the optimized symbol distance field to carry out adaptive octree subdivision so as to realize multi-resolution improvement. And 3) adopting an inner and outer double-layer cyclic training architecture, optimizing the symbolic distance field and the grid geometric parameters on the outer layer, and optimizing a bound Gaussian unit on the inner layer to realize end-to-end joint optimization of surface geometry and Gaussian rendering.