A Simple and Efficient Multiscale Edge Detection Method Based on Image Diffusion

In this paper, a fast multi-scale edge detection approach is proposed based on the theories of image diffusion and curve evolution. In comparison with the previous edge detection approaches, edge detection is performed in two steps: beginning by detecting the initial contours at a courser scale that is performed by using a linear diffusion coupling with the denoising effect and gray transformation, then the obtained contour curves are mapped and further refined up to finer scales using the fast Hermes algorithm. By this way, the experimental results show that the proposed edge detection approach are more promising than the existing methods for object detection on general images, especially on medical images.