Extended Data Fig. 2: Noise and Resolution estimation for AFM-imaged particles of RNase P RNA.
From: Determining structures of RNA conformers using AFM and deep neural networks
a, Cropped raw AFM images of the selected particles: P1, P2 and P3. The empty area around the perimeter of each cropped particle (gray) was used to calculate the background noise. b, Noise assessment (z-value distribution) of the gray areas in the raw images. c, AFM images after processing (leveling, spike-filtering, FFT-filtering). d, Noise assessment of the gray areas in the processed images. e, FFT images of the three particles (P1, P2 and P3) after image processing. A low-pass Fourier filter was applied as concentric rings (black) spaced in 5-pixel increments. f, Inverse FFT (IFFT) images, in real space, of the Fourier-filtered particles at each respective Fourier-ring cutoff radius (r). g, Auto-correlation value (ACV) profiles, in Fourier space, after the applied low-pass cut-off filters.
