A Pruning-and-Voting Strategy to Speed up the Detection for Lines, Circles, and Ellipses.

Shape analysis is a fundamental problem in image processing field. In shape analysis, lines, circles, and ellipses are three important features since they often occur in the image. Based on the determined windows on the edge map, this paper first presents a novel pruning-and-voting strategy to speed up the detection for lines, circles, and ellipses. Especially, our proposed strategy can be plugged into several existing randomized algorithms to reduce the required computation time while preserving the same robustness. In addition, some related time complexity analyses are provided to show the computational advantage of our proposed strategy. Under some real images, experimental results confirm our theoretic analyses. [ABSTRACT FROM AUTHOR]