Please use this identifier to cite or link to this item:
Title: Three-dimensional contour edge detection algorithm
Authors: Wang, Yizhou
Ong, S.H. 
Kassim, Ashraf A. 
Foong, Kelvin W.C. 
Issue Date: 2000
Citation: Wang, Yizhou,Ong, S.H.,Kassim, Ashraf A.,Foong, Kelvin W.C. (2000). Three-dimensional contour edge detection algorithm. Proceedings of SPIE - The International Society for Optical Engineering 3979 : I/-. ScholarBank@NUS Repository.
Abstract: This paper presents a novel algorithm for automatically extracting 3D contour edges, which are points of maximum surface curvature in a surface range image. The 3D image data are represented as a surface polygon mesh. The algorithm transforms the range data, obtained by scanning a dental plaster cast, into a 2D gray scale image by linearly converting the z-value of each vertex to a gray value. The Canny operator is applied to the median-filtered image to obtain the edge pixels and their orientations. A vertex in the 3D object corresponding to the detected edge pixel and its neighbors in the direction of the edge gradient are further analyzed with respect to their n-curvatures to extract the real 3D contour edges. This algorithm provides a fast method of reducing and sorting the unwieldy data inherent in the surface mesh representation. It employs powerful 2D algorithms to extract features from the transformed 3D models and refers to the 3D model for further analysis of selected data. This approach substantially reduces the computational burden without losing accuracy. It is also easily extended to detect 3D landmarks and other geometrical features, thus making it applicable to a wide range of applications.
Source Title: Proceedings of SPIE - The International Society for Optical Engineering
ISSN: 0277786X
Appears in Collections:Staff Publications

Show full item record
Files in This Item:
There are no files associated with this item.

Page view(s)

checked on Aug 18, 2019

Google ScholarTM


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.