EDGE DETECTION TECHNIQUES USING NONLINEAR DIFFUSION-BASED MODELS.

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 disscused 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 disscused. Our own contributions in this computer vision domain are finally described. [ABSTRACT FROM AUTHOR]