Object boundary edge selection for accurate contour tracking using multi-level canny edges

Abstract

We propose a method of selecting only tracked subject boundary edges in a video stream with changing background and a moving camera. Our boundary edge selection is done in two steps; first, remove background edges using an edge motion, second, from the output of the previous step, select boundary edges using a normal direction derivative of the tracked contour. In order to remove background edges, we compute edge motions and object motions. The edges with different motion direction than the subject motion are removed. In selecting boundary edges using the contour normal direction, we compute image gradient values on every edge pixels, and select edge pixels with large gradient values. We use multi-level Canny edge maps to get proper details of a scene. Detailed-level edge maps give us more scene information even though the tracked object boundary is not clear, because we can adjust the detail level of edge maps for a scene. We use Watersnake model to decide a new tracked contour. Our experimental results show that our approach is superior to Nguyen's.