基于调和模型神经网络的彩色图像复原研究

Research on color image restoration based on harmonic neural network

针对彩色图像复原提出了基于网络能量递减收敛的调和模型神经网络图像复原方法,研究了该方法在运动模糊图像复原上的应用。利用待复原图像重构出多幅模糊图像用于算法的实现,并首次提出基于图像局部方差的自适应正则化算子的实现方法。实验结果表明,该方法是有效的,复原效果优于有约束的最小二乘复原法和已有的传统神经网络图像复原法,对复原图像的信噪比有一定的提高。

A color image restoration algorithm using harmonic neural network is proposed in this paper, research the algorithm on motion-blurred image restoration. The algorithm utilizes the blurred image to reconstruct pieces of images used for the realization of the algorithm, and presents an adaptive regularization operator achieving method that based on image local variance. Experimental results show the efficiency of the new algorithm, its restoration results are better than the least square image restoration method with constraints and another image restoration based on traditional neural network,certain to improve the signal-to-noise ratio of the restored image.