求解无约束凸优化问题的广义压缩邻近算法

图像作为一种信息载体,在现代生活中占据了十分重要的地位,但往往由于成像设备或拍摄距离等原因,导致获取的图像质量非常差。因此图像恢复问题显得尤为重要。图像恢复问题可以转换成求解两个凸函数的极小化问题,本文提出了一种新的压缩邻近点算法来解决这个问题,并且进一步提出了具有有界误差的迭代算法。在适当的参数条件下,证明了算法的强收敛性。

As an information carrier,the image occupies an important position in modern life,but often results in poor image quality due to imaging equipment or shooting distance.Generally,the image recovery problem can be transformed into an unconstrained convex optimization problem of the sum of two functions.In this paper,we get new contraction-proximal point algorithms for solving the problem.Furthermore,we present the algorithm with bounded error.We also get the relative strong convergence under some conditions.