- author = "Lars Vidar Magnusson and Roland Olsson",
- title = "Improving the Canny Edge Detector Using Automatic Programming: Improving Non-Max Suppression",
- booktitle = "GECCO '16: Proceedings of the 2016 Annual Conference on Genetic and Evolutionary Computation",
- year = "2016",
- editor = "Tobias Friedrich and Frank Neumann and Andrew M. Sutton and Martin Middendorf and Xiaodong Li and Emma Hart and Mengjie Zhang and Youhei Akimoto and Peter A. N. Bosman and Terry Soule and Risto Miikkulainen and Daniele Loiacono and Julian Togelius and Manuel Lopez-Ibanez and Holger Hoos and Julia Handl and Faustino Gomez and Carlos M. Fonseca and Heike Trautmann and Alberto Moraglio and William F. Punch and Krzysztof Krawiec and Zdenek Vasicek and Thomas Jansen and Jim Smith and Simone Ludwig and JJ Merelo and Boris Naujoks and Enrique Alba and Gabriela Ochoa and Simon Poulding and Dirk Sudholt and Timo Koetzing",
- pages = "461--468",
- keywords = "genetic algorithms, genetic programming, ADATE",
- month = "20-24 " # jul,
- organisation = "SIGEVO",
- address = "Denver, USA",
- publisher = "ACM",
- publisher_address = "New York, NY, USA",
- isbn13 = "978-1-4503-4206-3",
-
DOI = "
doi:10.1145/2908812.2908926",
-
abstract = "we employ automatic programming, a relatively unknown
evolutionary computation strategy, to improve the
non-max suppression step in the popular Canny edge
detector. The new version of the algorithm has been
tested on a dataset widely used to benchmark edge
detection algorithms. The performance has increased by
1.9percent, and a pairwise student-t comparison with
the original algorithm gives a p-value of 6.45 x 10-9.
We show that the changes to the algorithm have made it
better at detecting weak edges, without increasing the
computational complexity or changing the overall
design. Previous attempts have been made to improve the
filter stage of the Canny algorithm using evolutionary
computation, but, to our knowledge, this is the first
time it has been used to improve the non-max
suppression algorithm.
The fact that we have found a heuristic improvement to the algorithm with significantly better performance on a dedicated test set of natural images suggests that our method should be used as a standard part of image analysis platforms, and that our methodology could be used to improve the performance of image analysis algorithms in general.",
- notes = "GECCO-2016 A Recombination of the 25th International Conference on Genetic Algorithms (ICGA-2016) and the 21st Annual Genetic Programming Conference (GP-2016)",
