An efficient ant-based edge detector

  • Authors:
  • Doğan Aydin

  • Affiliations:
  • Department of Computer Engineering, Ege University, Izmir, Turkey

  • Venue:
  • Transactions on computational collective intelligence I
  • Year:
  • 2010

Quantified Score

Hi-index 0.02

Visualization

Abstract

An efficient ant-based edge detector is presented. It is based on the distribution of ants on an image, ants try to find possible edges by using a state transition function based on 5x5 edge structures. Visual comparisons show that the proposed method gives finer details and thinner edges at lesser computational times when compared to earlier ant-based approaches. When compared to standard edge detectors, it shows robustness to Gaussian and Salt & Pepper noise and provides finer details than others with same parameter set in both clear and noisy images.