Thumbnail
Access Restriction
Open

Author Wöhler, Christian ♦ Krüger, Lars
Source CiteSeerX
Content type Text
File Format PDF
Language English
Subject Domain (in DDC) Computer science, information & general works ♦ Data processing & computer science
Subject Keyword Volumetric Image ♦ Enable Direct Reconstruction ♦ Contour Segmentation Algorithm ♦ Greedy Algorithm ♦ Model-specific Energy Term ♦ Three-dimensional Space ♦ Segment Contour ♦ Weak Model ♦ Ziplock Ribbon Snake ♦ On-line Quality Inspection Application ♦ Separate Active Contour ♦ Object Contour ♦ Multi-view Reconstruction ♦ Flexible Object ♦ Fast Algorithm ♦ Ground Truth Amount ♦ Real-world Scene ♦ Disparity Pixel ♦ Quality Inspection ♦ Average Distance ♦ Multiview Active Contour ♦ Active Contour ♦ Model Information ♦ Hard Constraint ♦ Parametric Active Contour ♦ Symmetric Object
Description In this paper, 3D parametric active contours are used to segment contours of flexible objects. The contour is a 3D curve or a "weak " model supplying information about the object contour. Views from two or more viewpoints are incorporated and enable direct reconstruction in three-dimensional space, without utilising volumetric images or separate active contours on each image. This curve is optimised using a greedy algorithm, leading to a fast algorithm, suitable for use in on-line quality inspection applications. 3D ziplock ribbon snakes are used to model tubes and other approximately rotationally symmetric objects. Model information is incorporated by hard constraints and model-specific energy terms. The proposed 3D contour segmentation algorithm is applied one synthetic and two real-world scenes. The average distance between the reconstructed contour and the ground truth amounts to 1 mm, which is equivalent to approximately 1 disparity pixel for all three examples. Keywords: 3D Reconstruction, Active Contour, Multi-View Reconstruction
Educational Role Student ♦ Teacher
Age Range above 22 year
Educational Use Research
Education Level UG and PG ♦ Career/Technical Study
Learning Resource Type Article
Publisher Date 2004-01-01
Publisher Institution In Proc. of International Conference on Computer Vision and Graphics