Table 3 Aberrations and resulting dataset size used in simulated training datasets.
From: nNPipe: a neural network pipeline for automated analysis of morphologically diverse catalyst systems
Applied aberration values | C0 = [−50; −500] Å | C0 = [−300; −1600] Å |
C3,0 = [−0.05; 0.5] mm | C3,0 = [−0.05; 0.5] mm | |
A1 = [0; 250] Å | A1 = [0; 250] Å | |
A1,θ = [0; 360] ° | A1,θ = [0; 360] ° | |
B2 = [0; 500] Å | B2 = [0; 500] Å | |
B2,θ = [0; 360] ° | B2,θ = [0; 360] ° | |
A2 = [0; 1200] Å | A2 = [0; 1200] Å | |
A2,θ = [0; 360] ° | A2,θ = [0; 360] ° | |
σ = [90; 160] Å | σ = [90; 160] Å | |
GLD criterion | 0.025 | 0.025 |
Images for object detection | 140,000 | 140,000 |
Images for semantic segmentation | 480,730 | 618,090 |
