Wavefront reconstruction with artificial neural networks

In this work, a new approach, a method using artificial neural networks was applied to reconstruct the wavefront. First, the optimal structure of neural networks was found. Then, the networks were trained on both noise-free and noisy spot patterns. The results of the wavefront reconstruction with artificial neural networks were compared to those obtained through the least square fit and singular value decomposition method. (c) 2006 Optical Society of America.