Man-made Object Detection Based on Fractal Theory and Active Contour Model

A novel man-made object detection method based on fractal characteristic is presented in this paper. Firstly, the proposed method gets the saliency map with the difference of fractal characteristic between man-made objects and natural background. Then, a threshold is adaptively calculated using the maximum entropy method to binarize the saliency map and get the candidate area. After that, the minimum enclosing rectangle that contains the candidate area is taken as the initial profile curve and finally the geometric active contour model is applied to make the curve evolving to the object contour. The proposed method can effectively solve the problem of active contour model associated with curve initialization. Also, experimental results on different image sequences demonstrate that it can accurately detect and segment the man-made objects and is robust to the perturbations of the background.