According to the characteristics of bore data,a model of 3D geologic body with generalized tri-prism as the primitive modeling element is constructed while the modeling process and key algorithms of modeling are prese...According to the characteristics of bore data,a model of 3D geologic body with generalized tri-prism as the primitive modeling element is constructed while the modeling process and key algorithms of modeling are presented here in detail.Using this method,the original bore data go through Delaunay triangulation to generate irregular triangular network on the surface,and then links stratum segments on the adjoining bores in session to form tri-prisms which would be pinched out.Finally stratified 3D geologic body model is built by an iterated search which searches for consecutive layer of the same property.The result shows that this method can effectively simulate stratified stratum modeling.展开更多
From the mathematical principles, the generalized potential theory can be employed to create constitutive model of geomaterial directly. The similar Cam-clay model, which is created based on the generalized potential ...From the mathematical principles, the generalized potential theory can be employed to create constitutive model of geomaterial directly. The similar Cam-clay model, which is created based on the generalized potential theory, has less assumptions,clearer mathematical basis, and better computational accuracy. Theoretically, it is more scientific than the traditional Cam-clay models. The particle flow code PFC3 D was used to make numerical tests to verify the rationality and practicality of the similar Cam-clay model. The verification process was as follows: 1) creating the soil sample for numerical test in PFC3 D, and then simulating the conventional triaxial compression test, isotropic compression test, and isotropic unloading test by PFC3D; 2)determining the parameters of the similar Cam-clay model from the results of above tests; 3) predicting the sample's behavior in triaxial tests under different stress paths by the similar Cam-clay model, and comparing the predicting results with predictions by the Cam-clay model and the modified Cam-clay model. The analysis results show that the similar Cam-clay model has relatively high prediction accuracy, as well as good practical value.展开更多
Robust stereo disparity estimation plays a critical role in minimally invasive surgery,where dynamic soft tissues,specular reflections,and data scarcity pose major challenges to traditional end-to-end deep learning an...Robust stereo disparity estimation plays a critical role in minimally invasive surgery,where dynamic soft tissues,specular reflections,and data scarcity pose major challenges to traditional end-to-end deep learning and deformable model-based methods.In this paper,we propose a novel disparity estimation framework that leverages a pretrained StyleGAN generator to represent the disparity manifold of Minimally Invasive Surgery(MIS)scenes and reformulates the stereo matching task as a latent-space optimization problem.Specifically,given a stereo pair,we search for the optimal latent vector in the intermediate latent space of StyleGAN,such that the photometric reconstruction loss between the stereo images is minimized while regularizing the latent code to remain within the generator’s high-confidence region.Unlike existing encoder-based embedding methods,our approach directly exploits the geometry of the learned latent space and enforces both photometric consistency and manifold prior during inference,without the need for additional training or supervision.Extensive experiments on stereo-endoscopic videos demonstrate that our method achieves high-fidelity and robust disparity estimation across varying lighting,occlusion,and tissue dynamics,outperforming Thin Plate Spline(TPS)-based and linear representation baselines.This work bridges generative modeling and 3D perception by enabling efficient,training-free disparity recovery from pre-trained generative models with reduced inference latency.展开更多
Subsurface rocks,as complex porous media,exhibit multiscale pore structures and intricate physical properties.Digital rock physics technology has become increasingly influential in the study of subsurface rock propert...Subsurface rocks,as complex porous media,exhibit multiscale pore structures and intricate physical properties.Digital rock physics technology has become increasingly influential in the study of subsurface rock properties.Given the multiscale characteristics of rock pore structures,direct three-dimensional imaging at sub-micrometer and nanometer scales is typically infeasible.This study introduces a method for reconstructing porous media using multidimensional data,which combines one-dimensional pore structure parameters with two-dimensional images to reconstruct three-dimensional models.The pore network model(PNM)is stochastically reconstructed using one-dimensional parameters,and a generative adversarial network(GAN)is utilized to equip the PNM with pore morphologies derived from two-dimensional images.The digital rocks generated by this method possess excellent controllability.Using Berea sandstone and Grosmont carbonate samples,we performed digital rock reconstructions based on PNM extracted by the maximum ball algorithm and compared them with stochastically reconstructed PNM.Pore structure parameters,permeability,and formation factors were calculated.The results show that the generated samples exhibit good consistency with real samples in terms of pore morphology,pore structure,and physical properties.Furthermore,our method effectively supplements the micropores not captured in CT images,demonstrating its potential in multiscale carbonate samples.Thus,the proposed reconstruction method is promising for advancing porous media property research.展开更多
Automatic target recognition （ATR） is an important issue for military applications, the topic of the ATR system belongs to the field of pattern recognition and classification. In the paper, we present an approach fo...Automatic target recognition （ATR） is an important issue for military applications, the topic of the ATR system belongs to the field of pattern recognition and classification. In the paper, we present an approach for building an ATR system with improved artificial neural network to recog- nize and classify the typical targets in the battle field. The invariant features of Hu invariant moments and roundness were selected to be the inputs of the neural network because they have the invari- ances of rotation, translation and scaling. The pictures of the targets are generated by the 3-D mod- els to improve the recognition rate because it is necessary to provide enough pictures for training the artificial neural network. The simulations prove that the approach can be implement ed in the ATR system and it has a high recognition rate and can be applied in real time.展开更多
Background Cumulus clouds are important elements in creating virtual outdoor scenes.Modeling cumulus clouds that have a specific shape is difficult owing to the fluid nature of the cloud.Image-based modeling is an eff...Background Cumulus clouds are important elements in creating virtual outdoor scenes.Modeling cumulus clouds that have a specific shape is difficult owing to the fluid nature of the cloud.Image-based modeling is an efficient method to solve this problem.Because of the complexity of cloud shapes,the task of modeling the cloud from a single image remains in the development phase.Methods In this study,a deep learning-based method was developed to address the problem of modeling 3D cumulus clouds from a single image.The method employs a three-dimensional autoencoder network that combines the variational autoencoder and the generative adversarial network.First,a 3D cloud shape is mapped into a unique hidden space using the proposed autoencoder.Then,the parameters of the decoder are fixed.A shape reconstruction network is proposed for use instead of the encoder part,and it is trained with rendered images.To train the presented models,we constructed a 3D cumulus dataset that included 2003D cumulus models.These cumulus clouds were rendered under different lighting parameters.Results The qualitative experiments showed that the proposed autoencoder method can learn more structural details of 3D cumulus shapes than existing approaches.Furthermore,some modeling experiments on rendering images demonstrated the effectiveness of the reconstruction model.Conclusion The proposed autoencoder network learns the latent space of 3D cumulus cloud shapes.The presented reconstruction architecture models a cloud from a single image.Experiments demonstrated the effectiveness of the two models.展开更多
Denoising diffusion models have demonstrated tremendous success in modeling data distributions and synthesizing high-quality samples.In the 2D image domain,they have become the state-of-the-art and are capable of gene...Denoising diffusion models have demonstrated tremendous success in modeling data distributions and synthesizing high-quality samples.In the 2D image domain,they have become the state-of-the-art and are capable of generating photo-realistic images with high controllability.More recently,researchers have begun to explore how to utilize diffusion models to generate 3D data,as doing so has more potential in real-world applications.This requires careful design choices in two key ways:identifying a suitable 3D representation and determining how to apply the diffusion process.In this survey,we provide the first comprehensive review of diffusion models for manipulating 3D content,including 3D generation,reconstruction,and 3D-aware image synthesis.We classify existing methods into three major categories:2D space diffusion with pretrained models,2D space diffusion without pretrained models,and 3D space diffusion.We also summarize popular datasets used for 3D generation with diffusion models.Along with this survey,we maintain a repository https://github.com/cwchenwang/awesome-3d-diffusion to track the latest relevant papers and codebases.Finally,we pose current challenges for diffusion models for 3D generation,and suggest future research directions.展开更多
Recent advances in generative artiflcial intelligence(AI)technologies have been signiflcantly driven by models such as generative adversarial networks(GANs),variational autoencoders(VAEs),and denoising diffusion proba...Recent advances in generative artiflcial intelligence(AI)technologies have been signiflcantly driven by models such as generative adversarial networks(GANs),variational autoencoders(VAEs),and denoising diffusion probabilistic models(DDPMs).Although architects recognize the potential of generative AI in design,personal barriers often restrict their access to the latest technological developments,thereby causing the application of generative AI in architectural design to lag behind.Therefore,it is essential to comprehend the principles and advancements of generative AI models and analyze their relevance in architecture applications.This paper flrst provides an overview of generative AI technologies,with a focus on probabilistic diffusion models(DDPMs),3D generative models,and foundation models,highlighting their recent developments and main application scenarios.Then,the paper explains how the abovementioned models could be utilized in architecture.We subdivide the architectural design process into six steps and review related research projects in each step from 2020 to the present.Lastly,this paper discusses potential future directions for applying generative AI in the architectural design steps.This research can help architects quickly understand the development and latest progress of generative AI and contribute to the further development of intelligent architecture.展开更多
Inspired by the rapid progress of generative AI techniques,there have been huge advances made for the 3D(three-dimensional)reconstruction community,which promoted the traditional 3D reconstruction framework from deep ...Inspired by the rapid progress of generative AI techniques,there have been huge advances made for the 3D(three-dimensional)reconstruction community,which promoted the traditional 3D reconstruction framework from deep implicit 3D reconstruction to generative 3D reconstruction,achieving more robust and expansive 3D reconstruction results with the help of generative AI models.Meanwhile,there is still a lack of corresponding review articles to provide a comprehensive analysis of recent advances from the perspective of 3D reconstruction.In response,this paper gives a comprehensive review for the generative 3D reconstruction approaches,especially on the recent advances made from the computer graphics and vision communities.Firstly,this paper mainly divides the recent generative 3D reconstruction approaches into four categories,including generative structure-from-motion/multiview-sterero(SfM/MVS),generative adversarial networks(GAN)based 3D reconstruction,diffusion-based 3D reconstruction,and cross-modal 3D reconstruction,which cover most generative-model aided 3D reconstruction work with a comprehensive review and analysis.Thereafter,some representative applications inspired by the generative 3D reconstruction including dynamic human avatars,3D interactive editing,and autonomous driving are also reviewed.Besides,some major datasets widely used for the generative 3D reconstruction approaches are included.Finally,this paper makes a discussion of the potential future work in further improving the quality of generative 3D reconstruction,towards better and more intelligent 3D reconstruction and generation.展开更多
Heat and thermal problems are major obstacles to achieving high power density in compact permanent magnet(PM)topologies.Consequently,a comprehensive,accurate,and rapid temperature rise estimation method is required fo...Heat and thermal problems are major obstacles to achieving high power density in compact permanent magnet(PM)topologies.Consequently,a comprehensive,accurate,and rapid temperature rise estimation method is required for novel electric machines to ensure safe and reliable operations.A unique three-dimensional(3D)lumped parameter thermal network(LPTN)is presented for accurate thermal modeling of a newly developed outer-rotor hybrid-PM flux switching generator(OR-HPMFSG)for direct-drive applications.First,the losses of the OR-HPMFSG are calculated using 3D finite element analysis(FEA).Subsequently,all machine components considering the thermal contact resistance,anisotropic thermal conductivity of materials,and various heat flow paths are comprehensively modeled based on the thermal resistances.In the proposed 3-D LPTN,internal nodes are considered to predict the average temperature as well as the hot spots of all active and passive components.Experimental measurements are performed on a prototype OR-HPMFSG to validate the efficiency of the 3-D LPTN.A comparison of the results at various operating points between the developed 3-D LPTN,experimental test,and FEA indicates that the 3-D LPTN quickly approximates the hotspot and mean temperature of all components under both transient and steady states with high accuracy.展开更多
The quantitative analysis shows that no theoretical model for 3-d magnetoelastic bodies, in literatures to date, can commonly simulate two kinds of distinct experimental phenomena on magnetoelastic interaction of ferr...The quantitative analysis shows that no theoretical model for 3-d magnetoelastic bodies, in literatures to date, can commonly simulate two kinds of distinct experimental phenomena on magnetoelastic interaction of ferromagnetic structures. This makes it difficult to effectively discribe the magnetoelastic mechanical behavior of structures with complex geometry, such as shells. Therefore, it is a key step for simulating magnetoelastic mechanical characteristics of structures with complex geometry to establish a 3-d model which also can commonly characterize the two distinct experimental phenomena. A theoretical model for three dimension magnetizable elastic bodies, which is commonly suitable for the two kinds of experimental phenomena on magnetoelastic interaction of ferromagnetic plates, is presented by the variational principle for the total energy functional of the coupling system of the 3-d ferromagnetic bodies. It is found that for the case of linear isotropic magnetic materials, the magnetic forces obtained by this model include not only the body magnetic force which is the same as that got from the magnetic dipole model, but also a distribution of the magnetic traction on the surface of the magnetizable body. And the value of the traction is equal to the jumping one of the Faraday electromagnetic stress on the two sides of the surface, which does not appear in any model, such as magnetic dipole model and axiomatic model.展开更多
We introduce Hair-GAN,an architecture of generative adversarial networks,to recover the 3D hair structure from a single image.The goal of our networks is to build a parametric transformation from 2D hair maps to 3D ha...We introduce Hair-GAN,an architecture of generative adversarial networks,to recover the 3D hair structure from a single image.The goal of our networks is to build a parametric transformation from 2D hair maps to 3D hair structure.The 3D hair structure is represented as a 3D volumetric field which encodes both the occupancy and the orientation information of the hair strands.Given a single hair image,we first align it with a bust model and extract a set of 2D maps encoding the hair orientation information in 2D,along with the bust depth map to feed into our Hair-GAN.With our generator network,we compute the 3D volumetric field as the structure guidance for the final hair synthesis.The modeling results not only resemble the hair in the input image but also possesses many vivid details in other views.The efficacy of our method is demonstrated by using a variety of hairstyles and comparing with the prior art.展开更多
Deep learning has been successfully used for tasks in the 2D image domain.Research on 3D computer vision and deep geometry learning has also attracted attention.Considerable achievements have been made regarding featu...Deep learning has been successfully used for tasks in the 2D image domain.Research on 3D computer vision and deep geometry learning has also attracted attention.Considerable achievements have been made regarding feature extraction and discrimination of 3D shapes.Following recent advances in deep generative models such as generative adversarial networks,effective generation of 3D shapes has become an active research topic.Unlike 2D images with a regular grid structure,3D shapes have various representations,such as voxels,point clouds,meshes,and implicit functions.For deep learning of 3D shapes,shape representation has to be taken into account as there is no unified representation that can cover all tasks well.Factors such as the representativeness of geometry and topology often largely affect the quality of the generated 3D shapes.In this survey,we comprehensively review works on deep-learning-based 3D shape generation by classifying and discussing them in terms of the underlying shape representation and the architecture of the shape generator.The advantages and disadvantages of each class are further analyzed.We also consider the 3D shape datasets commonly used for shape generation.Finally,we present several potential research directions that hopefully can inspire future works on this topic.展开更多
Various techniques have been developed and introduced to address the pressing need to create three-dimensional(3D)content for advanced applications such as virtual reality and augmented reality.However,the intricate n...Various techniques have been developed and introduced to address the pressing need to create three-dimensional(3D)content for advanced applications such as virtual reality and augmented reality.However,the intricate nature of 3D shapes poses a greater challenge to their representation and generation than standard two-dimensional(2D)image data.Different types of representations have been proposed in the literature,including meshes,voxels and implicit functions.Implicit representations have attracted considerable interest from researchers due to the emergence of the radiance field representation,which allows the simultaneous reconstruction of both geometry and appearance.Subsequent work has successfully linked traditional signed distance fields to implicit representations,and more recently the triplane has offered the possibility of generating radiance fields using 2D content generators.Many articles have been published focusing on these particular areas of research.This paper provides a comprehensive analysis of recent studies on implicit representation-based 3D shape generation,classifying these studies based on the representation and generation architecture employed.The attributes of each representation are examined in detail.Potential avenues for future research in this area are also suggested.展开更多
文摘According to the characteristics of bore data,a model of 3D geologic body with generalized tri-prism as the primitive modeling element is constructed while the modeling process and key algorithms of modeling are presented here in detail.Using this method,the original bore data go through Delaunay triangulation to generate irregular triangular network on the surface,and then links stratum segments on the adjoining bores in session to form tri-prisms which would be pinched out.Finally stratified 3D geologic body model is built by an iterated search which searches for consecutive layer of the same property.The result shows that this method can effectively simulate stratified stratum modeling.
基金Projects(51378131,51378403)supported by the National Natural Science Foundation of ChinaProject(2012210020203)supported by the Fundamental Research Funds for the Central Universities,China
文摘From the mathematical principles, the generalized potential theory can be employed to create constitutive model of geomaterial directly. The similar Cam-clay model, which is created based on the generalized potential theory, has less assumptions,clearer mathematical basis, and better computational accuracy. Theoretically, it is more scientific than the traditional Cam-clay models. The particle flow code PFC3 D was used to make numerical tests to verify the rationality and practicality of the similar Cam-clay model. The verification process was as follows: 1) creating the soil sample for numerical test in PFC3 D, and then simulating the conventional triaxial compression test, isotropic compression test, and isotropic unloading test by PFC3D; 2)determining the parameters of the similar Cam-clay model from the results of above tests; 3) predicting the sample's behavior in triaxial tests under different stress paths by the similar Cam-clay model, and comparing the predicting results with predictions by the Cam-clay model and the modified Cam-clay model. The analysis results show that the similar Cam-clay model has relatively high prediction accuracy, as well as good practical value.
基金Support by Sichuan Science and Technology Program[2023YFSY0026,2023YFH0004]Guangzhou Huashang University[2024HSZD01].
文摘Robust stereo disparity estimation plays a critical role in minimally invasive surgery,where dynamic soft tissues,specular reflections,and data scarcity pose major challenges to traditional end-to-end deep learning and deformable model-based methods.In this paper,we propose a novel disparity estimation framework that leverages a pretrained StyleGAN generator to represent the disparity manifold of Minimally Invasive Surgery(MIS)scenes and reformulates the stereo matching task as a latent-space optimization problem.Specifically,given a stereo pair,we search for the optimal latent vector in the intermediate latent space of StyleGAN,such that the photometric reconstruction loss between the stereo images is minimized while regularizing the latent code to remain within the generator’s high-confidence region.Unlike existing encoder-based embedding methods,our approach directly exploits the geometry of the learned latent space and enforces both photometric consistency and manifold prior during inference,without the need for additional training or supervision.Extensive experiments on stereo-endoscopic videos demonstrate that our method achieves high-fidelity and robust disparity estimation across varying lighting,occlusion,and tissue dynamics,outperforming Thin Plate Spline(TPS)-based and linear representation baselines.This work bridges generative modeling and 3D perception by enabling efficient,training-free disparity recovery from pre-trained generative models with reduced inference latency.
基金supported by the Shandong Provincial Natural Science Foundation(ZR2024MD116)National Natural Science Foundation of China(Grant Nos.42174143,42004098)Technology Innovation Leading Program of Shaanxi(No.2024 ZC-YYDP-27).
文摘Subsurface rocks,as complex porous media,exhibit multiscale pore structures and intricate physical properties.Digital rock physics technology has become increasingly influential in the study of subsurface rock properties.Given the multiscale characteristics of rock pore structures,direct three-dimensional imaging at sub-micrometer and nanometer scales is typically infeasible.This study introduces a method for reconstructing porous media using multidimensional data,which combines one-dimensional pore structure parameters with two-dimensional images to reconstruct three-dimensional models.The pore network model(PNM)is stochastically reconstructed using one-dimensional parameters,and a generative adversarial network(GAN)is utilized to equip the PNM with pore morphologies derived from two-dimensional images.The digital rocks generated by this method possess excellent controllability.Using Berea sandstone and Grosmont carbonate samples,we performed digital rock reconstructions based on PNM extracted by the maximum ball algorithm and compared them with stochastically reconstructed PNM.Pore structure parameters,permeability,and formation factors were calculated.The results show that the generated samples exhibit good consistency with real samples in terms of pore morphology,pore structure,and physical properties.Furthermore,our method effectively supplements the micropores not captured in CT images,demonstrating its potential in multiscale carbonate samples.Thus,the proposed reconstruction method is promising for advancing porous media property research.
基金Supported by the Ministerial Level Advanced Research Foundation(9140A01010411BQ01)the National Twelfth Five-Year Project(40405050303)
文摘Automatic target recognition （ATR） is an important issue for military applications, the topic of the ATR system belongs to the field of pattern recognition and classification. In the paper, we present an approach for building an ATR system with improved artificial neural network to recog- nize and classify the typical targets in the battle field. The invariant features of Hu invariant moments and roundness were selected to be the inputs of the neural network because they have the invari- ances of rotation, translation and scaling. The pictures of the targets are generated by the 3-D mod- els to improve the recognition rate because it is necessary to provide enough pictures for training the artificial neural network. The simulations prove that the approach can be implement ed in the ATR system and it has a high recognition rate and can be applied in real time.
基金the National Key R&D Program of China(2017YFB1002702).
文摘Background Cumulus clouds are important elements in creating virtual outdoor scenes.Modeling cumulus clouds that have a specific shape is difficult owing to the fluid nature of the cloud.Image-based modeling is an efficient method to solve this problem.Because of the complexity of cloud shapes,the task of modeling the cloud from a single image remains in the development phase.Methods In this study,a deep learning-based method was developed to address the problem of modeling 3D cumulus clouds from a single image.The method employs a three-dimensional autoencoder network that combines the variational autoencoder and the generative adversarial network.First,a 3D cloud shape is mapped into a unique hidden space using the proposed autoencoder.Then,the parameters of the decoder are fixed.A shape reconstruction network is proposed for use instead of the encoder part,and it is trained with rendered images.To train the presented models,we constructed a 3D cumulus dataset that included 2003D cumulus models.These cumulus clouds were rendered under different lighting parameters.Results The qualitative experiments showed that the proposed autoencoder method can learn more structural details of 3D cumulus shapes than existing approaches.Furthermore,some modeling experiments on rendering images demonstrated the effectiveness of the reconstruction model.Conclusion The proposed autoencoder network learns the latent space of 3D cumulus cloud shapes.The presented reconstruction architecture models a cloud from a single image.Experiments demonstrated the effectiveness of the two models.
文摘Denoising diffusion models have demonstrated tremendous success in modeling data distributions and synthesizing high-quality samples.In the 2D image domain,they have become the state-of-the-art and are capable of generating photo-realistic images with high controllability.More recently,researchers have begun to explore how to utilize diffusion models to generate 3D data,as doing so has more potential in real-world applications.This requires careful design choices in two key ways:identifying a suitable 3D representation and determining how to apply the diffusion process.In this survey,we provide the first comprehensive review of diffusion models for manipulating 3D content,including 3D generation,reconstruction,and 3D-aware image synthesis.We classify existing methods into three major categories:2D space diffusion with pretrained models,2D space diffusion without pretrained models,and 3D space diffusion.We also summarize popular datasets used for 3D generation with diffusion models.Along with this survey,we maintain a repository https://github.com/cwchenwang/awesome-3d-diffusion to track the latest relevant papers and codebases.Finally,we pose current challenges for diffusion models for 3D generation,and suggest future research directions.
基金supported by the Innovative Research Group Project of the National Natural Science Foundation of China(Grant No.202401-202712)。
文摘Recent advances in generative artiflcial intelligence(AI)technologies have been signiflcantly driven by models such as generative adversarial networks(GANs),variational autoencoders(VAEs),and denoising diffusion probabilistic models(DDPMs).Although architects recognize the potential of generative AI in design,personal barriers often restrict their access to the latest technological developments,thereby causing the application of generative AI in architectural design to lag behind.Therefore,it is essential to comprehend the principles and advancements of generative AI models and analyze their relevance in architecture applications.This paper flrst provides an overview of generative AI technologies,with a focus on probabilistic diffusion models(DDPMs),3D generative models,and foundation models,highlighting their recent developments and main application scenarios.Then,the paper explains how the abovementioned models could be utilized in architecture.We subdivide the architectural design process into six steps and review related research projects in each step from 2020 to the present.Lastly,this paper discusses potential future directions for applying generative AI in the architectural design steps.This research can help architects quickly understand the development and latest progress of generative AI and contribute to the further development of intelligent architecture.
基金supported by the National Key Research and Development Program of China under Grant No.2024YFB2808804supported by the National Natural Science Foundation of China under Grant No.U22B2034+1 种基金the Fundamental Research Funds for the Central Universities of China under Grant No.2253200009supported by the National Natural Science Foundation of China under Grant No.62402262.
文摘Inspired by the rapid progress of generative AI techniques,there have been huge advances made for the 3D(three-dimensional)reconstruction community,which promoted the traditional 3D reconstruction framework from deep implicit 3D reconstruction to generative 3D reconstruction,achieving more robust and expansive 3D reconstruction results with the help of generative AI models.Meanwhile,there is still a lack of corresponding review articles to provide a comprehensive analysis of recent advances from the perspective of 3D reconstruction.In response,this paper gives a comprehensive review for the generative 3D reconstruction approaches,especially on the recent advances made from the computer graphics and vision communities.Firstly,this paper mainly divides the recent generative 3D reconstruction approaches into four categories,including generative structure-from-motion/multiview-sterero(SfM/MVS),generative adversarial networks(GAN)based 3D reconstruction,diffusion-based 3D reconstruction,and cross-modal 3D reconstruction,which cover most generative-model aided 3D reconstruction work with a comprehensive review and analysis.Thereafter,some representative applications inspired by the generative 3D reconstruction including dynamic human avatars,3D interactive editing,and autonomous driving are also reviewed.Besides,some major datasets widely used for the generative 3D reconstruction approaches are included.Finally,this paper makes a discussion of the potential future work in further improving the quality of generative 3D reconstruction,towards better and more intelligent 3D reconstruction and generation.
文摘Heat and thermal problems are major obstacles to achieving high power density in compact permanent magnet(PM)topologies.Consequently,a comprehensive,accurate,and rapid temperature rise estimation method is required for novel electric machines to ensure safe and reliable operations.A unique three-dimensional(3D)lumped parameter thermal network(LPTN)is presented for accurate thermal modeling of a newly developed outer-rotor hybrid-PM flux switching generator(OR-HPMFSG)for direct-drive applications.First,the losses of the OR-HPMFSG are calculated using 3D finite element analysis(FEA).Subsequently,all machine components considering the thermal contact resistance,anisotropic thermal conductivity of materials,and various heat flow paths are comprehensively modeled based on the thermal resistances.In the proposed 3-D LPTN,internal nodes are considered to predict the average temperature as well as the hot spots of all active and passive components.Experimental measurements are performed on a prototype OR-HPMFSG to validate the efficiency of the 3-D LPTN.A comparison of the results at various operating points between the developed 3-D LPTN,experimental test,and FEA indicates that the 3-D LPTN quickly approximates the hotspot and mean temperature of all components under both transient and steady states with high accuracy.
基金Project supported by the National Natural Science Foundation of China (Grant No. 19572031)the National Science Fundation for Outstanding Young Scientiests in Chinaa united foundation of the State Education Committee of China and National Natural
文摘The quantitative analysis shows that no theoretical model for 3-d magnetoelastic bodies, in literatures to date, can commonly simulate two kinds of distinct experimental phenomena on magnetoelastic interaction of ferromagnetic structures. This makes it difficult to effectively discribe the magnetoelastic mechanical behavior of structures with complex geometry, such as shells. Therefore, it is a key step for simulating magnetoelastic mechanical characteristics of structures with complex geometry to establish a 3-d model which also can commonly characterize the two distinct experimental phenomena. A theoretical model for three dimension magnetizable elastic bodies, which is commonly suitable for the two kinds of experimental phenomena on magnetoelastic interaction of ferromagnetic plates, is presented by the variational principle for the total energy functional of the coupling system of the 3-d ferromagnetic bodies. It is found that for the case of linear isotropic magnetic materials, the magnetic forces obtained by this model include not only the body magnetic force which is the same as that got from the magnetic dipole model, but also a distribution of the magnetic traction on the surface of the magnetizable body. And the value of the traction is equal to the jumping one of the Faraday electromagnetic stress on the two sides of the surface, which does not appear in any model, such as magnetic dipole model and axiomatic model.
基金This work is partially supported by the National Key Research&Development Program of China(2018YFE0100900)the China Young 1000 Talent Programthe Fundamental Research Funds for the Central Universities.
文摘We introduce Hair-GAN,an architecture of generative adversarial networks,to recover the 3D hair structure from a single image.The goal of our networks is to build a parametric transformation from 2D hair maps to 3D hair structure.The 3D hair structure is represented as a 3D volumetric field which encodes both the occupancy and the orientation information of the hair strands.Given a single hair image,we first align it with a bust model and extract a set of 2D maps encoding the hair orientation information in 2D,along with the bust depth map to feed into our Hair-GAN.With our generator network,we compute the 3D volumetric field as the structure guidance for the final hair synthesis.The modeling results not only resemble the hair in the input image but also possesses many vivid details in other views.The efficacy of our method is demonstrated by using a variety of hairstyles and comparing with the prior art.
基金supported by the National Natural Science Foundation of China(Grant No.61902210)RCUK grant CAMERA(Grant Nos.EP/M023281/1,EP/T022523/1).
文摘Deep learning has been successfully used for tasks in the 2D image domain.Research on 3D computer vision and deep geometry learning has also attracted attention.Considerable achievements have been made regarding feature extraction and discrimination of 3D shapes.Following recent advances in deep generative models such as generative adversarial networks,effective generation of 3D shapes has become an active research topic.Unlike 2D images with a regular grid structure,3D shapes have various representations,such as voxels,point clouds,meshes,and implicit functions.For deep learning of 3D shapes,shape representation has to be taken into account as there is no unified representation that can cover all tasks well.Factors such as the representativeness of geometry and topology often largely affect the quality of the generated 3D shapes.In this survey,we comprehensively review works on deep-learning-based 3D shape generation by classifying and discussing them in terms of the underlying shape representation and the architecture of the shape generator.The advantages and disadvantages of each class are further analyzed.We also consider the 3D shape datasets commonly used for shape generation.Finally,we present several potential research directions that hopefully can inspire future works on this topic.
基金supported by National Natural Science Foundation of China(No.62322210)Beijing Municipal Natural Science Foundation for Distinguished Young Scholars(No.JQ21013)Beijing Municipal Science and Technology Commission(No.Z231100005923031).
文摘Various techniques have been developed and introduced to address the pressing need to create three-dimensional(3D)content for advanced applications such as virtual reality and augmented reality.However,the intricate nature of 3D shapes poses a greater challenge to their representation and generation than standard two-dimensional(2D)image data.Different types of representations have been proposed in the literature,including meshes,voxels and implicit functions.Implicit representations have attracted considerable interest from researchers due to the emergence of the radiance field representation,which allows the simultaneous reconstruction of both geometry and appearance.Subsequent work has successfully linked traditional signed distance fields to implicit representations,and more recently the triplane has offered the possibility of generating radiance fields using 2D content generators.Many articles have been published focusing on these particular areas of research.This paper provides a comprehensive analysis of recent studies on implicit representation-based 3D shape generation,classifying these studies based on the representation and generation architecture employed.The attributes of each representation are examined in detail.Potential avenues for future research in this area are also suggested.
