Image restoration with a neural network
M. Kurimoto, Sajad Ahmad Kawa, R. Ishii · 2003
The authors propose an image restoration algorithm by which the sharpness of blurred images can be restored. The restoration algorithm can be applied when characteristics of the degrading system are unknown. A neural network is used as a kind of nonlinear filter. This restoration algorithm was implemented with a neural network. This neural network has a layered structure, and values of connection weights are obtained by training with the backpropagation algorithm. Simulation results show that blurred images can be restored by using this algorithm.>