Font Size: a A A

Research Of The Image Sequence Segmentation Based On The Active Contour

Posted on:2003-02-05Degree:MasterType:Thesis
Country:ChinaCandidate:X Y JiangFull Text:PDF
GTID:2168360095950744Subject:Computer application technology
Abstract/Summary:PDF Full Text Request
The segmentation of a given image is a key step from the Image Processing to Image Understanding, which is also a basic technique in the Computer Vision. It is for the reason that the image can change to be more abstract and impacted after segmenting the image, separating the objects, extracting the features, measuring the parameters and so on. That's made the high-level Computer Vision tasks possible, such as Image Analysis and Understanding. The image segmentation has been absorbing many attentions for many years and a great amount of the segmentation algorithms have been proposed.In this paper a system for segmenting and tracking the objects from sequences is designed. This system is mainly made up of three parts that are the motion estimation giving out the motion vectors, the edge detection finding the edge points and contour and the deformable contour achieving the segmentation and tracking for the objects. Through testing many different Block Matching Algorithms in the motion estimation, the New Three Step Search Algorithm is chosen, which can fulfill the requirement of the system. For in general the motion in the sequences is small and the New Three Step Search is especially effective for capturing the small motion.hi designing the edge detection algorithm, the multi-scales character of the wavelet change and the character of the B-Spline function as a best smoothing filter are combined. Based on the Scheme of Local Modulus Maxima, a B-Spline wavelet function is designed to detect the edge points. In the post-processing for the edge points, the contour is found through the methods based on the Mathematical Morphology.The active contour is chosen to be the method for segmentation. The basic idea of the active contour is introduced first. A curve can move to the destination with the force from itself and outside, whenever its energy become minimum it arrives the aim points. The internal force that is the elasticity and the strength makes the curve regular and smooth. The external force given by the edge map makes the curve moving in the direction of the edge, hi traditional active contour the external force is the potential energy of the edge map, which has some insurmountable difficulties. Therefore the GVF(Gradient Vector Field) is used to substitute the traditional potential field and itovercomes the defects of the traditional active contour.Through the experiments with the standard test sequences a satisfactory result is got.
Keywords/Search Tags:Image Segmentation, Motion Estimation, Block Matching Algorithm, Edge Detection, B-Spline wavelet function, Mathematical Morphology, Active Contour
PDF Full Text Request
Related items