：通过对梯度矢量流(GVF)力场迭代和轮廓收敛过程中时间开销大、不能逼近较复杂轮

廓等不足之处的分析，提出了一种新的初始轮廓线的设置方法。基于小波变换的多尺度轮廓检测算

法，能够较精确地定位出各类轮廓，在此基础上设置的初始轮廓线能够非常好地靠近图像中的各种真

实轮廓。实验结果表明，此方法有效地缩小了搜索的范围，减少了GVF迭代次数，提高了轮廓收敛的

速度，并保留了GVF Snake模型的所有优点。-: The gradient vector flow (GVF) field iterative convergence process and outline the time cost, could not approach the more complex contours of such deficiencies, a new method of setting the initial contour. Wavelet-based multi-scale contour detection algorithm can accurately locate the various types of contours, on the basis set to a very good initial contour close to the image in a variety of real contours. Experimental results show that this method effectively reduces the search and reduce the number of iterations of the GVF to improve the convergence speed profile, and retained all the advantages of GVF Snake model.