CV模型是一种重要的图像分割模型，本代码针对其收敛速度慢、效率低的缺点，提出一种求解CV模型的新方法。将CV模型的能量泛函改写为改写为与原来有稳定解的总变分公式形式，然后使用对偶公式法求总变分公式的极小值，再在其中引入一速度向以加快模型的收敛速度。新方法克服了梯度下降法，要求时间步长小，迭代次数多的缺点。-CV model is a kind of important image segmentation model, the code for its slow convergence speed, the shortcomings of low efficiency, put forward a new method for solving the CV model.Will change the form of the energy functional to CV model with the original total variational formula form of stable solution, and then use the dual formula method of minimum total variational formula, in which introduce another speed to to speed up the convergence rate of the model.New method overcomes the gradient descent method for small time step, the disadvantages of the more iterative times.