Figure 3
From: Single-step phase identification and phase locking for coherent beam combination using deep learning
Prediction accuracy for (a) camera A (µ = 0.26 ± 0.08 radians) and (b) camera B (µ = 0.41 ± 0.21 radians), and relative importance of camera pixels in making phase predictions calculated using gradient analysis for (c) camera A and (d) camera B. Orange arrows highlight the asymmetry in (c).
