一种改进的Hopfield神经网络SST图像恢复方法

A New Image Restoration Algorithm of SST Using Hopfield Neural Network

提出了采用三元存储结构和图像分块的改进算法,同时利用模拟退火算法和遗传算法来提高恢复算法的收敛速度,从而进一步提高图像恢复的实时性。最后利用所提出的基于Hopfield神经网络的快速恢复方法进行模拟计算,其运算效率和恢复结果都能达到SST图像恢复的使用要求。

The algorithm is modified to reduce its space and time complexities. And the simulated annealing algorithm and genetic algorithm are used to make the neural network stable quickly. At last an image is processed using the new restoration algorithm based on the Hopfield neural network produced in this paper. The result and the computation cost are both better than the traditional algorithm and the faster algorithm is proved to be effective in processing SST image.