Research On Image Segmentation Method Based On Active Contour Model

The main task of image segmentation is to divide the target and background with their own unique attributes.In the field of computer vision,image segmentation technology has been playing an important role as the leading operation of image.The image processing program generally uses image segmentation to locate the target area,reduce the interference of redundant data,and further operate the target area accurately,so as to accurately identify the target.Active contour model-based image segmentation method is a flexible and effective image segmentation method.Because the method itself is easy to combine with the cross domain,it has attracted the attention of many experts and scholars in the field,resulting in a variety of active contour model-based image segmentation methods.In this thesis,the main research work of image segmentation method based on active contour model is as follows:Firstly,this thesis introduces the basic theory of image segmentation based on active contour model,including curve evolution theory and level set theory.At the same time,the principle of various active contour models is analyzed,and the characteristics of these classic active contour models are verified by experimental simulation.According to the experimental results,the advantages and problems of various models are analyzed.Secondly,aiming at the problem that the classical active contour model is not ideal for the segmentation of gray-scale uneven image,this thesis proposes a new image segmentation method based on active contour model.In this method,local information and global information are deeply mined,global information of image is fitted by deviation correction method,local information of image is fitted by Gaussian kernel function,and the fitting weight of global and local information is calculated adaptively by introducing weight coefficient,so as to improve the accuracy of image segmentation.The experimental results show that the proposed method can get more accurate segmentation results.Finally,the real-time processing of color image by active contour model is studied.Because color image has two more channels of gray-scale information than ordinary gray-scale image,the number of image pixels and the diversity of gray-scale values have changed in magnitude.Therefore,with the increase of data information,the segmentation speed is also affected.In order to solve the real-time problem of segmentation method,a fast image segmentation method based on natural gradient descent method is proposed in this thesis.By introducing threshold segmentation method to roughly segment the image,the result of rough segmentation can effectively eliminate some interference areas,reduce the amount of data in subsequent segmentation processing,and the result of threshold segmentation is used as the initial contour of segmentation.At the same time,combined with the theory of statistical model,the natural gradient descent method is used to optimize the energy function of the segmentation method,thus effectively accelerating the convergence speed of the segmentation method.The simulation results show that the improved method improves the segmentation speed significantly on the basis of ensuring the segmentation accuracy.