2004
DOI: 10.1023/b:visi.0000025798.50602.3a
|View full text |Cite
|
Sign up to set email alerts
|

Visual Modeling with a Hand-Held Camera

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
3
1
1

Citation Types

0
491
0
5

Year Published

2006
2006
2015
2015

Publication Types

Select...
4
3

Relationship

0
7

Authors

Journals

citations
Cited by 757 publications
(496 citation statements)
references
References 63 publications
0
491
0
5
Order By: Relevance
“…One of the methods to create an initial mesh model is to perform a Delaunay triangulation on the 2D feature points from one of the images and project it into 3D space [17,7]. This method works well for dense stereo reconstruction but for feature-based surface reconstruction, it often triangulates points from different planar surfaces (as shown in Figure 1(b)), leading to artifacts when the object is viewed from different angles.…”
Section: Edge Point Detectionmentioning
confidence: 99%
See 1 more Smart Citation
“…One of the methods to create an initial mesh model is to perform a Delaunay triangulation on the 2D feature points from one of the images and project it into 3D space [17,7]. This method works well for dense stereo reconstruction but for feature-based surface reconstruction, it often triangulates points from different planar surfaces (as shown in Figure 1(b)), leading to artifacts when the object is viewed from different angles.…”
Section: Edge Point Detectionmentioning
confidence: 99%
“…The first steps usually involve Structure from Motion (SfM) [5] and camera auto-calibration [6], delivering camera pose information as well as a sparse 3D point reconstruction based on image feature points. Pollefeys et al [7] then applied dense stereo matching techniques on the images which results in a per-pixel density reconstruction of the scene. Lhuillier and Quan [8,9] adopted a different approach by propagating points on the images to obtain a "quasi-dense" reconstruction.…”
Section: Introductionmentioning
confidence: 99%
“…for reconstruction of architectural scenes from a single moving camera. In [1] camera pose is estimated using self calibration on a sparse set of points matched between local image pairs. Following camera estimation and image rectification, dense feature matching is performed to compute a dense disparity map which may be triangulated and textured using image intensity information.…”
Section: Previous Workmentioning
confidence: 99%
“…Structure from motion has been used in single camera architectural and terrain modeling with great success [1]. These domains tend to have many distinguishable features that simplify the pose estimation and dense reconstruction phases.…”
Section: Introductionmentioning
confidence: 99%
See 1 more Smart Citation