Fig. 17: Flowchart of the one-step deep-learning-based phase unwrapping approach and the unwrapping results of different methods. | Light: Science & Applications

Fig. 17: Flowchart of the one-step deep-learning-based phase unwrapping approach and the unwrapping results of different methods.

From: Deep learning in optical metrology: a review

Fig. 17

a The flowchart of the one-step deep-learning-based phase unwrapping approach: the absolute phase can be predicted directly from a noisy wrapped phase based on the trained CNN. b The wrapped phase map of living mouse osteoblast. c Unwrapped phase of (b) obtained by deep learning. d Phase errors of (c). e Unwrapped phase of (b) obtained by the conventional LS method322. f Phase errors of (e). af Adapted with permission from ref. 321, Optica publishing

Back to article page