Recent developments on direct relative orientation
By Henrik Stewénius, C. Engels, David Nistér. 2006
505 citations recorded by the source corpus.
Subfields: Computational Mechanics, Computer Graphics and Computer-Aided Design, Control and Systems Engineering
View Paper: https://www.semanticscholar.org/paper/514fa8d4981cc2b2aecfc02e0e3a8f4be717bcd7
How researchers see this paper
The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.
In Computational Mechanics
- Vladlen Koltun
- Francesc Moreno-Noguer
- Lourdes de Agapito
- Jaesik Park
- Qian-Yi Zhou
- Johannes L. Schönberger
- Pascal Fua
- Jan-Michael Frahm
In Computer Graphics and Computer-Aided Design
- Vladlen Koltun
- Johannes L. Schönberger
- Jan-Michael Frahm
- Jerry Quispe
- Francesc Moreno-Noguer
- Rui Yu
- Enrique Dunn
- Chris Russell
In Control and Systems Engineering
- Davide Scaramuzza
- Francesc Moreno-Noguer
- Friedrich Fraundorfer
- Vincent Lepetit
- Pascal Fua
- Roland Yves Siegwart
- Jerry Quispe
- Rui Yu
Citations in the corpus
- Real-time monocular visual odometry for on-road vehicles with 1-point RANSAC (2009, background): “For unconstrained motion of a calibrated camera this would be 5 point correspondences [19], [20].”
- Parallel Tracking and Mapping for Small AR Workspaces (2007, methodology): “These methods build an initial map f rom five-point stereo [27] and then track a camera using local bun dle adjustment over theN most recent camera poses (where N is selected to maintain real-time performance), achieving exceptiona l accuracy over long distances.”
- Colored Point Cloud Registration Revisited (2017, methodology): “The initial alignment is estimated by building correspondences between ORB features in the color images [30], pruning with the 5-point RANSAC algorithm [36], and computing a transformation that aligns the corresponding depth pixels [9].”
- Video Pop-up: Monocular 3D Reconstruction of Dynamic Scenes (2014, methodology): “First we initialise from the two frames using fivepoint algorithm [34], and then keep adding new frame into the process, triangulating new points and doing incremental bundle adjustment until all the frames are reconstructed.”
- Reconstructing the World* in Six Days *(As Captured by the Yahoo 100 Million Image Dataset) (2015, methodology): “Coupled with this, we use a 5-point essential matrix estimator [27], with estimates for the intrinsic camera parameters initialized using JPEG EXIF data whenever possible (assuming a 40◦ horizontal field-of-view otherwise).”
- Accurate Non-Iterative O(n) Solution to the PnP Problem (2007, background): “We demonstrated it in the context of the PnP problem but it is potentially applicable to problems ranging from the estimation of the Essential matrix from a large number of points for Structure-from-Motion applications [27] to shape recovery of deformable surfaces.”
- Visual Odometry : Part II: Matching, Robustness, Optimization, and Applications (2012, methodology): “Several five-point minimal solvers were proposed later in [38]–[40], but an efficient implementation, based on [39], was found only in 2003 by Nister [41] and later revised in [42].”
- Fast and robust numerical solutions to minimal problems for cameras with radial distortion (2008, background): “…reliable approach (so far) is to study a particular class of equations ( e.g . relative orientation for calibrated cameras [17], optimal three view triangulation [19], etc.) and use knowledge of what the structure of the Gr¨obner basis should be to design a special purpose Gr¨obner basis solver.”
- CZECH TECHNICAL UNIVERSITY IN PRAGUE (2005, methodology): “Various minimal problems have been recently studied extensively in computer vision [43, 32, 40, 37, 41, 42, 38, 33, 8, 31, 30].”