Based On Active Contour Model, Contour Extraction

In the fields of image processing and computer vision, extracting the contour from the moving or stillness objective is an important problem. It is a vital scientific research values and application values. In many researching and application domain, the contour of the target object needs being founding, and location the part of the image which you have interested. There are many methods for the contour extracting in the literature. For example, the boundary detection algorithm and image segmentation algorithm, threshold segmentation, region growing, clustering, mathematical morphological watershed algorithm etc. There are less or more demerits when these methods are used for extracting the contour. It isn't good for extracting the contour of the objective. At a later time many scholars propose many improving and new methods. Such as Active Contour Model and searching algorithm based on Dynamical Programming. In which Active Contour Model are applied abroad. In 1987 the 1st international visual conference, M. Kass, A. Witkin, and D. Terzopoulos propose Active Contour Model which has been quick development and applied to segmentation in the later of 1980s. It is provided a kind of unified frame for solving many questions of visual such as image segmentation and extracting contour of the objective. It is designed an energy function which is consisted of the internal energy, external energy and image energy. Minimizing the energy function we can get the contour of the objective. Now it is applied widely in image segmentation and the contour extracting.The paper mostly discuss two aspects of the extracting the contour of the object. One is that using the edge detecting algorithm to extract the contour directness. The paper proposes a new algorithm for edge line detection and uses it extracting the contour of a simple object from the image. The other one is combining the new algorithm with active contour model. Using the Multiresolution analysis and achieves the initial contour automatically. Firstly use the new edge line detection algorithm on the multiresolution images and get the contour. Let the contour as the model's initial contour, then we can get the precision location contour.