Abstract
A parallel procedure for locating and representing 2D contours is discussed. At first, circular templates are used to determine the edge pixels and the dominant points with high local curvatures. Then arcs are acquired from the edge pixels and the dominant points. The arcs are used to examine the validity of the dominant points and to represent the contours of surfaces. The method is mainly based on template-matching procedures and a modified Hough transform. © 1992.
| Original language | English |
|---|---|
| Pages (from-to) | 629-637 |
| Number of pages | 9 |
| Journal | Pattern Recognition Letters |
| Volume | 13 |
| Issue number | 9 |
| Publication status | Published - Sept 1992 |
Keywords
- Computer vision
- conic curves
- dominant points
- edge detection
- parallel image processing
- template matching