Gaussian Three-Dimensional kernel SVM for Edge Detection Applications

Schematic of the method: training edge images train a 3D Gaussian-kernel SVM, which produces edge maps compared against Canny and Sobel. This is a schematic, not the paper's figure.

Abstract

This paper presents a novel and uniform algorithm for edge detection based on SVM (support vector machine) with Three-dimensional Gaussian radial basis function with kernel. Because of disadvantages in traditional edge detection such as inaccurate edge location, rough edge and careless on detect soft edge. The experimental results indicate how the SVM can detect edge in efficient way. The performance of the proposed algorithm is compared with existing methods, including Sobel and canny detectors. The results show that this method is better than classical algorithm such as canny and Sobel detector.

Framework

3D Gaussian kernel

A support vector machine uses a three-dimensional Gaussian radial basis function kernel.

Trained on edges

Training images teach the SVM where edges are, including soft edges that classical detectors miss.

Beats the classics

Compared with Sobel and Canny detectors, the method detects edges more accurately.

Resources

Preprint: arXiv:1710.01260
DOI: 10.48550/arXiv.1710.01260

BibTeX

@misc{irandoust2017gaussian,
  title = {Gaussian Three-Dimensional kernel SVM for Edge Detection Applications},
  author = {Irandoust-Pakchin, Safar and Ayanzadeh, Aydin and Beikzadeh, Siamak},
  year = {2017},
  eprint = {1710.01260},
  archivePrefix = {arXiv},
  primaryClass = {cs.CV},
  doi = {10.48550/arXiv.1710.01260},
  url = {https://arxiv.org/abs/1710.01260},
  note = {Originally presented at the International Conference on New Research Findings in Electrical Engineering and Computer Science, Tehran, 2015}
}