Fringe pattern analysis using deep learning and physical priors

We propose a deep-learning-based fringe pattern analysis method that employs physical priors. It can achieve accurate and efficient single-shot phase reconstruction and exhibits strong generalization capability to unseen samples.