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.