Learning Transformation Synchronization

Reconstructing the 3D model of a physical object typically requires us to align the depth scans obtained from different camera poses into the same coordinate system. Solutions to this global alignment problem usually proceed in two steps. The first step estimates relative transformations between pairs of scans using an off-the-shelf technique. Due to limited information presented between pairs of scans, the resulting relative transformations are generally noisy. The second step then jointly optimizes the relative transformations among all input depth scans. A natural constraint used in this step is the cycle-consistency constraint, which allows us to prune incorrect relative transformations by detecting inconsistent cycles. The performance of such approaches, however, heavily relies on the quality of the input relative transformations. Instead of merely using the relative transformations as the input to perform transformation synchronization, we propose to use a neural network to learn the weights associated with each relative transformation. Our approach alternates between transformation synchronization using weighted relative transformations and predicting new weights of the input relative transformations using a neural network. We demonstrate the usefulness of this approach across a wide range of datasets.

[1]  Frank Dellaert,et al.  Initialization techniques for 3D SLAM: A survey on rotation estimation and its use in pose graph optimization , 2015, 2015 IEEE International Conference on Robotics and Automation (ICRA).

[2]  Andrea Fusiello,et al.  Model Acquisition by Registration of Multiple Acoustic Range Views , 2002, ECCV.

[3]  Andrea Fusiello,et al.  Spectral Synchronization of Multiple Views in SE(3) , 2016, SIAM J. Imaging Sci..

[4]  Ira Kemelmacher-Shlizerman,et al.  Global Motion Estimation from Point Matches , 2012, 2012 Second International Conference on 3D Imaging, Modeling, Processing, Visualization & Transmission.

[5]  Marc Levoy,et al.  Geometrically stable sampling for the ICP algorithm , 2003, Fourth International Conference on 3-D Digital Imaging and Modeling, 2003. 3DIM 2003. Proceedings..

[6]  Martial Hebert,et al.  Fully automatic registration of multiple 3D data sets , 2003, Image Vis. Comput..

[7]  Yifan Sun,et al.  Joint Map and Symmetry Synchronization , 2018, ECCV.

[8]  William T. Freeman,et al.  The Patch Transform , 2009, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[9]  Vladlen Koltun,et al.  A Large Dataset of Object Scans , 2016, ArXiv.

[10]  H. Pottmann,et al.  Reassembling fractured objects by geometric matching , 2006, SIGGRAPH 2006.

[11]  M. Hebert,et al.  Automatic three-dimensional modeling from reality , 2002 .

[12]  Leonidas J. Guibas,et al.  An optimization approach for extracting and encoding consistent maps in a shape collection , 2012, ACM Trans. Graph..

[13]  Leonidas J. Guibas,et al.  An Optimization Approach to Improving Collections of Shape Maps , 2011, Comput. Graph. Forum.

[14]  Leonidas J. Guibas,et al.  Consistent Shape Maps via Semidefinite Programming , 2013, SGP '13.

[15]  Amit Singer,et al.  Exact and Stable Recovery of Rotations for Robust Synchronization , 2012, ArXiv.

[16]  Leonidas J. Guibas,et al.  Near-Optimal Joint Object Matching via Convex Relaxation , 2014, ICML.

[17]  Venu Madhav Govindu,et al.  Robust Relative Rotation Averaging , 2018, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[18]  Andrea Torsello,et al.  Multiview registration via graph diffusion of dual quaternions , 2011, CVPR 2011.

[19]  Arrigoni Federica,et al.  Camera Motion from Group Synchronization , 2016 .

[20]  Sujay Sanghavi,et al.  Normalized Spectral Map Synchronization , 2016, NIPS.

[21]  Andrea Fusiello,et al.  Global Registration of 3D Point Sets via LRS Decomposition , 2016, ECCV.

[22]  A. Singer,et al.  Vector diffusion maps and the connection Laplacian , 2011, Communications on pure and applied mathematics.

[23]  Marc Pollefeys,et al.  Disambiguating visual relations using loop constraints , 2010, 2010 IEEE Computer Society Conference on Computer Vision and Pattern Recognition.

[24]  Qi-Xing Huang,et al.  SMAC: Simultaneous Mapping and Clustering Using Spectral Decompositions , 2018, ICML.

[25]  Qi-Xing Huang,et al.  Translation Synchronization via Truncated Least Squares , 2017, NIPS.

[26]  Vincent Lepetit,et al.  Learning to Find Good Correspondences , 2017, 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition.

[27]  Niloy J. Mitra,et al.  Super4PCS: Fast Global Pointcloud Registration via Smart Indexing , 2019 .

[28]  Leonidas J. Guibas,et al.  Functional map networks for analyzing and exploring large shape collections , 2014, ACM Trans. Graph..

[29]  N. Mitra,et al.  4 PCS Fast Global Pointcloud Registration via Smart Indexing , 2015 .

[30]  Venu Madhav Govindu,et al.  On Averaging Multiview Relations for 3D Scan Registration , 2014, IEEE Transactions on Image Processing.

[31]  Johan Thunberg,et al.  A solution for multi-alignment by transformation synchronisation , 2014, 2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

[32]  Stephen Lin,et al.  Recurrent Transformer Networks for Semantic Correspondence , 2018, NeurIPS.

[33]  Vladlen Koltun,et al.  Deep Fundamental Matrix Estimation , 2018, ECCV.

[34]  Yong Jae Lee,et al.  FlowWeb: Joint image set alignment by weaving consistent, pixel-wise correspondences , 2015, 2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

[35]  Sang Wook Lee,et al.  Multiview registration of 3D scenes by minimizing error between coordinate frames , 2002, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[36]  Vladlen Koltun,et al.  Open3D: A Modern Library for 3D Data Processing , 2018, ArXiv.

[37]  B. Rossi,et al.  Robust Absolute Rotation Estimation via Low-Rank and Sparse Matrix Decomposition , 2014, 2014 2nd International Conference on 3D Vision.

[38]  Xiaowei Zhou,et al.  Distributed consistent data association via permutation synchronization , 2016, 2017 IEEE International Conference on Robotics and Automation (ICRA).

[39]  I. Daubechies,et al.  Iteratively reweighted least squares minimization for sparse recovery , 2008, 0807.0575.

[40]  Onur Özyesil,et al.  Robust camera location estimation by convex programming , 2014, 2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

[41]  Noah Snavely,et al.  Robust Global Translations with 1DSfM , 2014, ECCV.

[42]  Vikas Singh,et al.  Permutation Diffusion Maps (PDM) with Application to the Image Association Problem in Computer Vision , 2014, NIPS.

[43]  Geoffrey E. Hinton,et al.  ImageNet classification with deep convolutional neural networks , 2012, Commun. ACM.

[44]  M. K. Kadalbajoo,et al.  An overview on the eigenvalue computation for matrices , 2011, Neural Parallel Sci. Comput..

[45]  Tobias Höllerer,et al.  Optimizing the Viewing Graph for Structure-from-Motion , 2015, 2015 IEEE International Conference on Computer Vision (ICCV).

[46]  Xiaowei Zhou,et al.  Multi-image Matching via Fast Alternating Minimization , 2015, 2015 IEEE International Conference on Computer Vision (ICCV).

[47]  Stephen DiVerdi,et al.  Exploring collections of 3D models using fuzzy correspondences , 2012, ACM Trans. Graph..

[48]  Matthias Nießner,et al.  ScanNet: Richly-Annotated 3D Reconstructions of Indoor Scenes , 2017, 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

[49]  Vikas Singh,et al.  Solving the multi-way matching problem by permutation synchronization , 2013, NIPS.

[50]  Vladlen Koltun,et al.  Fast Global Registration , 2016, ECCV.

[51]  Vladlen Koltun,et al.  Robust reconstruction of indoor scenes , 2015, 2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

[52]  Andrea Fusiello,et al.  Robust synchronization in SO(3) and SE(3) via low-rank and sparse matrix decomposition , 2015, Comput. Vis. Image Underst..

[53]  Venu Madhav Govindu,et al.  Efficient and Robust Large-Scale Rotation Averaging , 2013, 2013 IEEE International Conference on Computer Vision.