Fig. 2: Performance of AutoPhaseNN on simulated test data. | npj Computational Materials

Fig. 2: Performance of AutoPhaseNN on simulated test data.

From: AutoPhaseNN: unsupervised physics-aware deep learning of 3D nanoscale Bragg coherent diffraction imaging

Fig. 2

a Histogram of χ2 error for the modulus of the diffraction pattern. b Histogram of SSIM for the amplitude of the real space images. c Volume rendering of the input 3D diffraction patterns, d ground truth images, e phase retrieval results, f network predictions, and g refined results, for the four representative samples. d–g Show the 0.3 contour of the amplitude and the color represents the phase on the surfaces. Reciprocal χ2 errors and SSIMs for the amplitude of real space images are shown in the figures. The best figure of merit for each crystal is marked in red.

Back to article page