Precise Contour Extraction Based on Moving Object

摘要

The contour extraction method of moving object is one of the most challenging problems in the area of machine vision. We present a method for moving object in video tracking binding the improved active contour models and Mean-shift. It uses the mean-shift method to track and determine the range of the moving object. Then it uses the improved active contour model to extract the object edge, which makes the distance among the control-points continuously reducing by using of the curve minimum energy function. The curve has the contraction phenomenon by iterates after many times and the function causes the control points to gather the position of large curvature in the contour line and determine the edge more accurately. The method has achieved to the moving object edge accurate extraction.