Recognition of Two-Dimensional Shapes.
Abstract
A new scheme for coding the boundary of two-dimensional shapes is proposed in this thesis. Random points on the boundary are paired for this coding. Using this scheme, an effective and efficient correlation technique to match two-dimensional shapes is developed. This technique has a number of very desirable characteristics. It is able to match shapes of arbitrary scale and orientation. The given shape may have closed or open boundary or even have portion of it obstructed from view. Matching can be performed at varying degrees of details, giving this technique an added robustness against geometric distortions. It also has the capability to discriminate between different shapes. Computation time on the IBM 3033 computer is typically 10 CPU seconds to generate one correlation curve between two shapes, each with a 500-point boundary curve. (Author)
Document Details
- Document Type
- Technical Report
- Publication Date
- Mar 01, 1986
- Accession Number
- ADA168403
Entities
People
- Gim P. Quek
Organizations
- Naval Postgraduate School