Abstract
An algorithm for curves partial matching, up to a similarity transformation, is presented. The matching is based on the total-curvature graph, which is the direction of the first derivative of the curve. No higher derivatives are required. The algorithm uses dynamic programming. It is not dependent on reference or feature points. The base step consists of a comparison of the total-curvatures of neighborhoods. The neighborhoods are dynamically chosen, according to local characteristics of the curve. In the recursive step, ratios of lengths of matched curve segments are compared, in order to extend the matches. The matching is not sensitive to local deviations of the curves. The algorithm is fast and simple. Its worst-case time complexity is O(mn), where m and n are the number of points in the two digital curves. Experimental results are presented.
Original language | English |
---|---|
Pages (from-to) | 199-209 |
Number of pages | 11 |
Journal | Pattern Recognition |
Volume | 28 |
Issue number | 2 |
DOIs | |
State | Published - 1 Jan 1995 |
Keywords
- Dynamic programming
- Invariants
- Objects recognition
- Similarity transformations
- Total-curvature
- group
ASJC Scopus subject areas
- Software
- Signal Processing
- Computer Vision and Pattern Recognition
- Artificial Intelligence