Manhattan-world stereo

By Yasutaka Furukawa, Brian Curless, S. M. Seitz, Richard Szeliski. 2009 IEEE Conference on Computer Vision and Pattern Recognition, 2009

391 citations recorded by the source corpus.

Subfields: Computer Vision and Pattern Recognition, Geology

View Paper: https://www.semanticscholar.org/paper/2767d6b27d4b2c36ea35b3a47379ef6617578512

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

In Geology

Citations in the corpus

School of Electrical and Computer Engineering Oklahoma State University , Stillwater , OK 74078 , USA (methodology): “The methods in the first group are RGB-based, which rely on extracting lines, edges, and orthogonal vanishing points from RGB images [9, 5].”
Revisiting Stereo Depth Estimation From a Sequence-to-Sequence Perspective with Transformers (2020, background): “At the same time, the biological vision system imposes geometric assumptions on objects as a piece-wise smoothness prior [11], which involves mid-level cortex processing.”
Semi-global Stereo Matching with Surface Orientation Priors (2017, background): “, expected ground plane orientation in autonomous driving [28]), from semantic analysis [15, 16, 2], or from vanishing point analysis and Manhattan-world assumptions [41, 43, 17, 32, 36], which we also explore in this paper.”
Just Look at the Image: Viewpoint-Specific Surface Normal Prediction for Improved Multi-View Reconstruction (2016, background): “Furukawa et al. [15] go even further and assume a Manhattan world [8].”
A Survey of Urban Reconstruction (2013, methodology): “The Manhattan World assumption has been successfully used by several urban reconstruction approaches [FCSS09a,VAW ∗ 10], since it robustly allows to identify fundamental shapes of most buildings.”
DeepMVS: Learning Multi-view Stereopsis (2018, positive): “Aside from photometric consistency, other 3D cues such as lighting [39, 24], shadows [1], color [11], geometric structures [7], and semantic cues [14] have been incorporated into the MVS pipeline for improving the reconstruction accuracy.”
Structured Prediction of Unobserved Voxels from a Single Depth Image (2016, methodology): “We also compare to a bounding box baseline, which fits minimum volume bounding boxes to the points of each segment using Manhattan directions computed using [16].”
Convolutional Occupancy Networks (2020, methodology): “cher geometric structure in the z dimension. Furthermore, by aligning the world coordinate frame with the dominant surface orientations of the scene, we eectively incorporate a Manhattan-World prior [7,11] into our model. Volume Encoder: While planar feature representations allow for encoding at lar”
Creating Large-Scale City Models from 3D-Point Clouds: A Robust Approach with Hybrid Representation (2012, positive): “They can be obtained from various types of data such as multi-view images (Dick et al., 2004; Strecha et al., 2008; Furukawa et al., 2009; Sinha et al., 2009; Vu et al., 2009; Vanegas et al., 2010), laser scans (Banno et al., 2008; Chen and Chen, 2008; Frueh and Zakhor, 2004) or video (Pollefeys et…”
Scene Understanding by Reasoning Stability and Safety (2015, methodology): “Inspired by Furukawa's method in Furukawa et al. (2009), the Manhattan space is carved by the point cloud into three parts, as shown in Fig.”
Patches, Planes and Probabilities: A Non-Local Prior for Volumetric 3D Reconstruction (2016, methodology): “Existing works on planar multi-view stereo or Manhattan world representations treat these as hard constraints [13, 14, 32].”
Automatic Creation of Semantically Rich 3D Building Models from Laser Scanner Data (2011, background): “The geometry of as-is BIMs is starting to be generated using data from laser scanners, although image-based techniques show promise for some applications (Furukawa et al., 2009b).”
Multi-View Stereo: A Tutorial (2015, methodology): “6 shows the reconstruction results of Manhattan-world Stereo algorithm [67] on both indoor and outdoor architectural scenes, which are full of planar surfaces that meet at right angles.”
Semantic Multi-view Stereo: Jointly Estimating Objects and Voxels (2017, background): “Existing works that consider low-level geometric priors such as spatial smoothness [6, 22, 38], piecewise planarity [16, 33] or Manhattan-world constraints [12, 31] cannot complete large occluded regions, especially for objects with complex geometry.”
Beyond Point Clouds: Scene Understanding by Reasoning Geometry and Physics (2013, methodology): “Voxel generation and gravity direction We first generate voxels for each segment obtained by above point cloud segmentation by 1) detecting Manhattan axes [7], 2) constructing voxels from point cloud along Manhattan axes by octree construction method [19], and 3) detecting gravity direction.”
Reconstructing building interiors from images (2009, methodology): “For this purpose, our system uses a stereo algorithm proposed by Furukawa et al. [8], which exploits the Manhattan-world assumption: surfaces are piece-wise planar and aligned with three dominant directions.”