Abstract
A trained deep neural network generates computational polarization-sensitive contrast from OCT intensity images. The deep-learning-enabled images were evaluated based on the structural similarity with the ground truth.
© 2021 The Author(s)
PDF Article