Deep-learning-based cross-modality translation from Stokes image to bright-field contrast.

By comparing the cross-modality translation performance with MM images, we found that the Stokes images, with the advantages of faster acquisition and independence from light intensity and image registration, can be well translated to bright-field images.