Nonparametric methods for image segmentation using active contour

This paper applies Parzen-window nonparametric estimation of active contour methods for image segmentation which can be performed without regarding image regions as any of typical probability distributions, so it expands the application range of the algorithm. In order to use the information of gray value congruence in a particular region, the Parzen-window parameter is alterable in estimating the densities. This will be more suit for practical image segmentation. Many examples of synthetical and real image segmentation prove that the method is rationality, better for curve evolution and has good accuracy and antinoise function.