Edge detection techniques using nonlinear diffusion-based models

Abstract

An overview of the edge detection techniques based on partial differential equations (PDE) is presented in this work. Nonlinear anisotropic diffusion-based boundary extraction approaches, like the influential Perona- Malik model and some improved variants of it, are described first. Anisotropic diffusion-based detection schemes using the mean curvature motion and nonlinear PDE-based approaches combining anisotropic diffusion to the bilateral filter are then discussed here. Some nonlinear reaction-diffusion-based edge detection methods are described next. Variational edge detection solutions using the total variation (TV) regularization or combining the anisotropic diffusion to the TV-based models are then presented. Directional diffusion-based image edge extraction algorithms are also discussed. Our own contributions in this computer vision domain are finally described.