Figure 2 | Scientific Reports

Figure 2

From: Physics constrained unsupervised deep learning for rapid, high resolution scanning coherent diffraction reconstruction

Figure 2

Datasets. Examples of amplitude images from three distinct dataset types used in this study. The ‘Lines’ and ‘Fines Features’ (Gaussian Random Field - GRF) datasets provide contrasting conditions of local symmetry (i.e., isotropy). The Gaussian random field process produces characteristic speckle, while ‘Lines’ consists of sharp, oriented edges. ‘Large Features’ is anisotropic but coarser than both ‘Lines’ and ‘Fine Features’.

Back to article page