【摘要】 传统的活动轮廓模型具有处理速度慢 ,运算量大 ,对凹陷轮廓处理效果差等缺点 .本文把小波多分辨率技术应用于主动轮廓算法 ,结合梯度矢量流 (gradient vector flow,GVF)概念 ,首先在低分辨率的图像上进行处理 ,得到结果后 ,再在高分辨率层次上继续处理 ,由此解决了初始轮廓必须离真实轮廓很近的问题 ,提高了运算速度 ;同时利用小波分解的方向性 ,降低了搜索的范围 ,进一步提高了轮廓收敛的速度 ;由于使用了GVF所产生的外力场 ,对于凹陷轮廓同样可以保证正确的收敛 ,同时比原始的GVF方法处理速度要快得多更多还原

【Abstract】 The traditional active contour model has many defects such as low speed, huge computation, poor concavity processing. In this paper multi-resolution technique of wavelet and GVF (gradient vector flow) are applied to active contour model algorithm to solve these problems. Preliminary work is processed at low-resolution image. The results are then processed at a high resolution level. Thus the problem that the initial start points of the snakes must coincide with the actual contour mostly is solved. Simultaneously, the convergence speed of the snakes is accelerated due to the refined searching region using wavelet decomposition. As a result of using GVF, the algorithm can process concavity well and the speed is much faster than the initial GVF snake.更多还原