Multi-View Stereo Reconstruction and Scene Flow Estimation with a Global Image-Based Matching Score

By Jean-Philippe Pons, Renaud Keriven, Olivier D. Faugeras. International Journal of Computer Vision, 2007

331 citations recorded by the source corpus.

Subfields: Computer Vision and Pattern Recognition

View Paper: https://www.semanticscholar.org/paper/db47a9a47b83f2294ca89c90d643829b4a909cc6

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computer Vision and Pattern Recognition

Citations in the corpus

Accurate, Dense, and Robust Multi-View Stereopsis (2007, background): “They can be roughly classified into four classes according to the underlying object models: Voxel-based approaches [3], [4], [5], [6], [7], [8], [9] require knowing a bounding box that contains the scene, and their accuracy is limited by the resolution of the voxel grid.”
Towards Internet-scale multi-view stereo (2010, methodology): “Many of the best performing MVS algorithms instead reconstruct a global 3D model directly from the input images [9, 13, 20, 24].”
Multi-view Scene Flow Estimation: A View Centered Variational Approach (2010, methodology): “5.”
Efficient Multi-View Reconstruction of Large-Scale Scenes using Interest Points, Delaunay Triangulation and Graph Cuts (2007, methodology): “Large-scale cluttered scenes for which no reliable initial guess of geometry is available also disqualify the deformable model framework [11, 13, 12, 21, 28, 32].”
Evaluation of Stereo Matching Costs on Images with Radiometric Differences (2009, unknown): “, NCC or MI) can be used for handling radiometric differences [6].”
FlowNet3D: Learning Scene Flow in 3D Point Clouds (2018, background): “Since this seminal work, many others have tried to jointly estimate structure and motion from stereoscopic images [12, 18, 34, 26, 5, 33, 28, 29, 1, 30, 16], mostly in a variational setting with regularizations for smoothness of motion and structure [12, 1, 26], or with assumption of the rigidity of”
Structure from motion photogrammetry in physical geography (2016, background): “As the final step of MVS is computationally burdensome when a global reconstruction is specified (Pons et al., 2007), there is the option to split a project into overlapping clusters of images to reduce random-access memory (RAM) requirements.”
Accurate, Dense, and Robust Multiview Stereopsis (2010, background): “They can be roughly classified into four classes according to the underlying object models: Voxel-based approaches [3], [4], [5], [6], [7], [8], [9] require knowing a bounding box that contains the scene, and their accuracy is limited by the resolution of the voxel grid.”
Joint 3d Estimation of Vehicles and Scene Flow (2015, background): “…(Horn and Schunck, 1980), the problem is often formulated in a coarse-to-ﬁne variational setting (Basha et al., 2013, Huguet and Devernay, 2007, Pons et al., 2007, Valgaerts et al., 2010, Wedel et al., 2011, Vogel et al., 2011) and local regularizers are leveraged to encourage smoothness in…”
Multiview Stereo via Volumetric Graph-Cuts and Occlusion Robust Photo-Consistency (2007, methodology): “This class consists of techniques using implici t representations such as voxel occupancy grids [16], or level-sets of 3D scalar fields [7], [20], and explicit representations such as polygonal meshes [8], [11].”
Scene Flow from Depth and Color Images (2011, background): “In the case of multiple image streams, both structure and motion can be estimated in a combined way [2, 13].”
Object scene flow for autonomous vehicles (2015, background): “[33, 34], the problem is traditionally formulated in a variational setting [1, 18, 23, 32, 36, 38] where optimization proceeds in a coarse-to-fine manner and local regularizers are leveraged to encourage smoothness in depth and motion.”
Efficient Dense Scene Flow from Sparse or Dense Stereo Data (2008, unknown): “Other dense scene flow algorithms have been presented in multiple camera set-ups [13,17].”
Kinecting the dots: Particle based scene flow from depth sensors (2011, methodology): “As with several other scene flow approaches [9, 13], the Scene Particle algorithm employs a coarse to fine strategy, to reduce the effects of local maxima in the probability distribution.”
Continuous Global Optimization in Multiview 3D Reconstruction (2007, methodology): “A generalization of this approach has been developed in (Pons et al., 2007), which allows to replace the classical pointwise photoconsistency estimation with a global matching score on the entire image domain.”
3D scene flow estimation with a rigid motion prior (2011, methodology): “To date, most scene flow approaches have relied on relatively simple priors favoring smooth surfaces and motion fields [3, 12, 22].”
High Accuracy and Visibility-Consistent Dense Multiview Stereo (2012, unknown): “Finally, the variational refinement uses an energy inspired from our previous work [46], but in a lightweight and scalable Lagrangian framework.”
Photometric Bundle Adjustment for Dense Multi-view 3D Modeling (2014, methodology): “For example [17] uses the Level Set framework using a global image score, [10] uses a convex formulation minimizing a photometric error defined over a discretized grid.”
SphereFlow: 6 DoF Scene Flow from RGB-D Pairs (2014, background): “[20] introduce a variational framework for multiple-view motion-stereo estimation that works directly in world space, evolving a 3D vector field to register the input images captured at different times.”