Filters








52 Hits in 7.3 sec

Learnable Triangulation for Deep Learning-based 3D Reconstruction of Objects of Arbitrary Topology from Single RGB Images [article]

Tarek Ben Charrada, Hedi Tabia, Aladine Chetouani, Hamid Laga
2021 arXiv   pre-print
We propose a novel deep reinforcement learning-based approach for 3D object reconstruction from monocular images. Prior works that use mesh representations are template based.  ...  In this paper, we propose a novel end-to-end method that reconstructs 3D objects of arbitrary topology from a monocular image.  ...  Method In this paper, we focus on the high visual quality reconstruction of 3D objects of arbitrary topology from a single RGB image.  ... 
arXiv:2109.11844v1 fatcat:cmz742pllbf7hpy2qvys2wakfy

A Review of Techniques for 3D Reconstruction of Indoor Environments

Zhizhong Kang, Juntao Yang, Zhou Yang, Sai Cheng
2020 ISPRS International Journal of Geo-Information  
Moreover, based on the hierarchical pyramid structures and the learnable parameters of deep-learning architectures, multi-task collaborative schemes to share parameters and to jointly optimize each other  ...  using redundant and complementary information from different perspectives show their potential for the 3D reconstruction of indoor environments.  ...  Acknowledgments: Sincere thanks are given for the comments and contributions of the anonymous reviewers and the members of the editorial team.  ... 
doi:10.3390/ijgi9050330 fatcat:nahmlugoa5egnji3wrcigfjrre

GAMesh: Guided and Augmented Meshing for Deep Point Networks [article]

Nitin Agarwal, M Gopi
2020 arXiv   pre-print
We show that such a separation of geometry from topology can have several advantages especially in single-view shape prediction, fair evaluation of point networks and reconstructing surfaces for networks  ...  We further show that by training point networks with GAMesh, we can directly optimize the vertex positions to generate adaptive meshes with arbitrary topologies.  ...  Single View Reconstruction Given a single RGB image of an object, our task is to reconstruct a 3D mesh with correct topology and accurate geometry.  ... 
arXiv:2010.09774v1 fatcat:7lh2zawbsjb4zp7eznn3q5iggy

Image-based 3D Object Reconstruction: State-of-the-Art and Trends in the Deep Learning Era

Xianfeng Han, Hamid Laga, Mohammed Bennamoun
2019 IEEE Transactions on Pattern Analysis and Machine Intelligence  
We focus on the works which use deep learning techniques to estimate the 3D shape of generic objects either from a single or multiple RGB images.  ...  Since 2015, image-based 3D reconstruction using convolutional neural networks (CNN) has attracted increasing interest and demonstrated an impressive performance.  ...  Intermediating Many of the deep learning-based 3D reconstruction algorithms predict the 3D geometry of an object from RGB images directly.  ... 
doi:10.1109/tpami.2019.2954885 pmid:31751229 fatcat:hc76yes6avdy5byyy7flovj5wa

Learning joint reconstruction of hands and manipulated objects [article]

Yana Hasson, Gül Varol, Dimitrios Tzionas, Igor Kalevatykh, Michael J. Black, Ivan Laptev, Cordelia Schmid
2019 arXiv   pre-print
Our approach improves grasp quality metrics over baselines, using RGB images as input.  ...  In this work, we regularize the joint reconstruction of hands and objects with manipulation constraints.  ...  We thank Tsvetelina Alexiadis, Jorge Marquez and Senya Polikovsky from MPI for help with scan acquisition, Joachim Tesch for the hand-object rendering, Mathieu Aubry and Thibault Groueix for advices on  ... 
arXiv:1904.05767v1 fatcat:mzkmscajc5ab5byjdowqywu4zm

Learning Joint Reconstruction of Hands and Manipulated Objects

Yana Hasson, Gul Varol, Dimitrios Tzionas, Igor Kalevatykh, Michael J. Black, Ivan Laptev, Cordelia Schmid
2019 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)  
Our approach improves grasp quality metrics over baselines, using RGB images as input.  ...  In this work, we regularize the joint reconstruction of hands and objects with manipulation constraints.  ...  We thank Tsvetelina Alexiadis, Jorge Marquez and Senya Polikovsky from MPI for help with scan acquisition, Joachim Tesch for the hand-object rendering, Mathieu Aubry and Thibault Groueix for advices on  ... 
doi:10.1109/cvpr.2019.01208 dblp:conf/cvpr/HassonVTKBLS19 fatcat:qoqkxbnpd5gx5n4swel34ytup4

Revealing Scenes by Inverting Structure From Motion Reconstructions

Francesco Pittaluga, Sanjeev J. Koppal, Sing Bing Kang, Sudipta N. Sinha
2019 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)  
a) SfM point cloud (top view) (b) Projected 3D points (c) Synthesized Image (d) Original Image Figure 1: SYNTHESIZING IMAGERY FROM A SFM POINT CLOUD: From left to right: (a) Top view of a SfM reconstruction  ...  We present a privacy attack that reconstructs color images of the scene from the point cloud.  ...  This is disappointing news from a privacy perspective but could be useful in other settings for generating photorealistic images from 3D reconstructions.  ... 
doi:10.1109/cvpr.2019.00023 dblp:conf/cvpr/PittalugaKKS19 fatcat:nwuhwlylsfggjpbterhk467usa

Data-Driven 3D Reconstruction of Dressed Humans From Sparse Views [article]

Pierre Zins, Yuanlu Xu, Edmond Boyer, Stefanie Wuhrer, Tony Tung
2021 arXiv   pre-print
an attention-based fusion layer that learns to aggregate visual information from several viewpoints; and third a mechanism that encodes local 3D patterns under the multi-view context.  ...  Specifically, we introduce three key components: first a spatially consistent reconstruction that allows for arbitrary placement of the person in the input views using a perspective camera model; second  ...  Acknowledgements We thank Laurence Boissieux and Julien Pansiot from the Kinovis platform at Inria Grenoble and our volunteer subjects for help with the 3D data acquisition.  ... 
arXiv:2104.08013v4 fatcat:vcvpmekukbehndo7iy7muv2yoa

A Review on Deep Learning Techniques for 3D Sensed Data Classification

David Griffiths, Jan Boehm
2019 Remote Sensing  
Over the past decade deep learning has driven progress in 2D image understanding.  ...  Finally, we give a detailed discussion about the future of deep learning for 3D sensed data, using literature to justify the areas where future research would be most valuable.  ...  Conflicts of Interest: The authors declare no conflict of interest.  ... 
doi:10.3390/rs11121499 fatcat:vg7bme6guzddleocgsm7soqaxq

Learning Structural Graph Layouts and 3D Shapes for Long Span Bridges 3D Reconstruction [article]

Fangqiao Hu, Jin Zhao, Yong Huang, Hui Li
2020 arXiv   pre-print
A learning-based 3D reconstruction method for long-span bridges is proposed in this paper. 3D reconstruction generates a 3D computer model of a real object or scene from images, it involves many stages  ...  Considering the prior human knowledge that these structures are in conformity to regular spatial layouts in terms of components, a learning-based topology-aware 3D reconstruction method which can obtain  ...  The objective of this paper is to learn a 3D model for bridges from UAV images.  ... 
arXiv:1907.03387v2 fatcat:4nd5pdnxp5e2pjbu7b5wsxq26e

Topologically Consistent Multi-View Face Inference Using Volumetric Sampling [article]

Tianye Li and Shichen Liu and Timo Bolkart and Jiayi Liu and Hao Li and Yajie Zhao
2021 arXiv   pre-print
Most learning-based methods use an underlying 3D morphable model (3DMM) to ensure robustness, but this limits the output accuracy for extreme facial expressions.  ...  High-fidelity face digitization solutions often combine multi-view stereo (MVS) techniques for 3D reconstruction and a non-rigid registration step to establish dense correspondence across identities and  ...  3D prior and reconstruct 3D faces from images.  ... 
arXiv:2110.02948v1 fatcat:d6be6b66jneotjjco4ybkj2lym

Differentiable Rendering: A Survey [article]

Hiroharu Kato, Deniz Beker, Mihai Morariu, Takahiro Ando, Toru Matsuoka, Wadim Kehl, Adrien Gaidon
2020 arXiv   pre-print
Differentiable rendering is a novel field which allows the gradients of 3D objects to be calculated and propagated through images.  ...  Deep neural networks (DNNs) have shown remarkable performance improvements on vision-related tasks such as object detection or image segmentation.  ...  We would like to thank the PyTorch3D developers for their insightful comments and suggestions.  ... 
arXiv:2006.12057v2 fatcat:6zj6besdcnebrb4qww4u4jusji

A survey on Deep Learning Advances on Different 3D Data Representations [article]

Eman Ahmed, Alexandre Saint, Abd El Rahman Shabayek, Kseniya Cherenkova, Rig Das, Gleb Gusev, Djamila Aouada, Bjorn Ottersten
2019 arXiv   pre-print
Recently, with the availability of both large 3D datasets and computational power, it is today possible to consider applying deep learning to learn specific tasks on 3D data such as segmentation, recognition  ...  3D data is a valuable asset the computer vision filed as it provides rich information about the full geometry of sensed objects and scenes.  ...  Projection-based representations are simple yet effective for learning 3D objects using 3D DL methods.  ... 
arXiv:1808.01462v2 fatcat:iuoay2sddjdqjbgm2nai6pa7gq

Deep 3D human pose estimation: A review

Jinbao Wang, Shujie Tan, Xiantong Zhen, Shuo Xu, Feng Zheng, Zhenyu He, Ling Shao
2021 Computer Vision and Image Understanding  
In this paper, we provide a thorough review of existing deep learning based works for 3D pose estimation, summarize the advantages and disadvantages of these methods and provide an in-depth understanding  ...  Three-dimensional (3D) human pose estimation involves estimating the articulated 3D joint locations of a human body from an image or video.  ...  Acknowledgments This work is supported by the National Natural Science Foundation of China under Grant No. 61972188.  ... 
doi:10.1016/j.cviu.2021.103225 fatcat:hvlgjuxd2zfgji6k4y4g65cs7y

Point2Mesh: A Self-Prior for Deformable Meshes [article]

Rana Hanocka, Gal Metzer, Raja Giryes, Daniel Cohen-Or
2020 arXiv   pre-print
The self-prior encapsulates reoccurring geometric repetitions from a single shape within the weights of a deep neural network.  ...  In this paper, we introduce Point2Mesh, a technique for reconstructing a surface mesh from an input point cloud.  ...  We are grateful for the 3D scans provided by Tom Pierce and Pierce Design.  ... 
arXiv:2005.11084v1 fatcat:flkgnpz7czgmhb4orbeuvcg2wa
« Previous Showing results 1 — 15 out of 52 results