Fuzzy contour characterization by subspace based methods of array processing and direct method

Blurred contours are often sought in digital image processing. Most of the existing techniques try to detect blurred contours by fuzzy techniques. In this article, we propose to model the blurred contours by generating the virtual signals from the image, and provide a new viewpoint to detect blurred contour. Especially, the blurred contour with exponential distribution can be presented by three parameters. Then, array processing methods are adopted to estimate these parameters, and optimization method is also considered when the spread parameter is estimated. Experiments finally prove that the proposed methods estimate the parameters of blurred contours accurately and with a reduced computational load.