More Like this

1. Sparse-view Pose Estimation and Reconstruction via Analysis by Generative Synthesis

   Zhao, Qitao; Tulsiani, Shubham (December 2024, Conference and Workshop on Neural Information Processing Systems)

   Inferring the 3D structure underlying a set of multi-view images typically requires solving two co-dependent tasks -- accurate 3D reconstruction requires precise camera poses, and predicting camera poses relies on (implicitly or explicitly) modeling the underlying 3D. The classical framework of analysis by synthesis casts this inference as a joint optimization seeking to explain the observed pixels, and recent instantiations learn expressive 3D representations (e.g., Neural Fields) with gradient-descent-based pose refinement of initial pose estimates. However, given a sparse set of observed views, the observations may not provide sufficient direct evidence to obtain complete and accurate 3D. Moreover, large errors in pose estimation may not be easily corrected and can further degrade the inferred 3D. To allow robust 3D reconstruction and pose estimation in this challenging setup, we propose SparseAGS, a method that adapts this analysis-by-synthesis approach by: a) including novel-view-synthesis-based generative priors in conjunction with photometric objectives to improve the quality of the inferred 3D, and b) explicitly reasoning about outliers and using a discrete search with a continuous optimization-based strategy to correct them. We validate our framework across real-world and synthetic datasets in combination with several off-the-shelf pose estimation systems as initialization. We find that it significantly improves the base systems' pose accuracy while yielding high-quality 3D reconstructions that outperform the results from current multi-view reconstruction baselines. 

   more » « less
2. Level-S ² fM: Structure from Motion on Neural Level Set of Implicit Surfaces

   https://doi.org/10.1109/CVPR52729.2023.01650  

   Xiao, Yuxi; Xue, Nan; Wu, Tianfu; Xia, Gui-Song (June 2023, IEEE)

   This paper presents a neural incremental Structure-from-Motion (SfM) approach, Level-S2fM, which estimates the camera poses and scene geometry from a set of uncalibrated images by learning coordinate MLPs for the implicit surfaces and the radiance fields from the established key-point correspondences. Our novel formulation poses some new challenges due to inevitable two-view and few-view configurations in the incremental SfM pipeline, which complicates the optimization of coordinate MLPs for volumetric neural rendering with unknown camera poses. Nevertheless, we demonstrate that the strong inductive basis conveying in the 2D correspondences is promising to tackle those challenges by exploiting the relationship between the ray sampling schemes. Based on this, we revisit the pipeline of incremental SfM and renew the key components, including two-view geometry initialization, the camera poses registration, the 3D points triangulation, and Bundle Adjustment, with a fresh perspective based on neural implicit surfaces. By unifying the scene geometry in small MLP networks through coordinate MLPs, our Level-S2fM treats the zero-level set of the implicit surface as an informative top-down regularization to manage the reconstructed 3D points, reject the outliers in correspondences via querying SDF, and refine the estimated geometries by NBA (Neural BA). Not only does our Level-S2fM lead to promising results on camera pose estimation and scene geometry reconstruction, but it also shows a promising way for neural implicit rendering without knowing camera extrinsic beforehand. 

   more » « less
3. 360ViewPET: View Based Pose EsTimation for Ultra-Sparse 360-Degree Cameras

   https://doi.org/10.1109/ISM52913.2021.00008  

   Zhou, Qian; Chen, Bo; Yang, Zhe; Guo, Hongpeng; Nahrstedt, Klara (December 2021, IEEE International Symposium on Multimedia)

   Immersive virtual tours based on 360-degree cameras, showing famous outdoor scenery, are becoming more and more desirable due to travel costs, pandemics and other constraints. To feel immersive, a user must receive the view accurately corresponding to her position and orientation in the virtual space when she moves inside, and this requires cameras’ orientations to be known. Outdoor tour contexts have numerous, ultra-sparse cameras deployed across a wide area, making camera pose estimation challenging. As a result, pose estimation techniques like SLAM, which require mobile or dense cameras, are not applicable. In this paper we present a novel strategy called 360ViewPET, which automatically estimates the relative poses of two stationary, ultra-sparse (15 meters apart) 360-degree cameras using one equirectangular image taken by each camera. Our experiments show that it achieves accurate pose estimation, with a mean error as low as 0.9 degree 

   more » « less
4. Trifocal Relative Pose From Lines at Points

   https://doi.org/10.1109/TPAMI.2022.3226165  

   Fabbri, Ricardo; Duff, Timothy; Fan, Hongyi; Regan, Margaret H.; de Pinho, David da; Tsigaridas, Elias; Wampler, Charles W.; Hauenstein, Jonathan D.; Giblin, Peter J.; Kimia, Benjamin; et al (December 2022, IEEE Transactions on Pattern Analysis and Machine Intelligence)

   Abstract—We present a method for solving two minimal problems for relative camera pose estimation from three views, which are based on three view correspondences of (i) three points and one line and the novel case of (ii) three points and two lines through two of the points. These problems are too difficult to be efficiently solved by the state of the art Gro ̈bner basis methods. Our method is based on a new efficient homotopy continuation (HC) solver framework MINUS, which dramatically speeds up previous HC solving by specializing HC methods to generic cases of our problems. We characterize their number of solutions and show with simulated experiments that our solvers are numerically robust and stable under image noise, a key contribution given the borderline intractable degree of nonlinearity of trinocular constraints. We show in real experiments that (i) SIFT feature location and orientation provide good enough point-and-line correspondences for three-view reconstruction and (ii) that we can solve difficult cases with too few or too noisy tentative matches, where the state of the art structure from motion initialization fails. 

   more » « less
5. FlowCam: Training generalizable 3D radiance fields without camera poses via pixel-aligned scene flow

   Smith, Cameron; Du, Yilun; Tewari, Ayush; Sitzmann, Vincent (December 2023, Neural Information Processing Systems)

   Reconstruction of 3D neural fields from posed images has emerged as a promising method for self-supervised representation learning. The key challenge preventing the deployment of these 3D scene learners on large-scale video data is their dependence on precise camera poses from structure-from-motion, which is prohibitively expensive to run at scale. We propose a method that jointly reconstructs camera poses and 3D neural scene representations online and in a single forward pass. We estimate poses by first lifting frame-to-frame optical flow to 3D scene flow via differentiable rendering, preserving locality and shift-equivariance of the image processing backbone. SE(3) camera pose estimation is then performed via a weighted least-squares fit to the scene flow field. This formulation enables us to jointly supervise pose estimation and a generalizable neural scene representation via re-rendering the input video, and thus, train end-to-end and fully self-supervised on real-world video datasets. We demonstrate that our method performs robustly on diverse, real-world video, notably on sequences traditionally challenging to optimization-based pose estimation techniques. 

   more » « less