Figure 8
Two-point Pearson autocorrelation maps of four segmented images (inset): (a) random noise at the one-pixel scale, (b) random noise at the ten-pixel scale, (c) square pattern with periodicity of five pixels, and (d) square pattern with periodicity of 20 pixels. Features (white) in (c)–(d) have the same size as 3 pixels. It can be seen in (a), (b) that the initial descent in the Pearson correlation (from g = 1 at = 0 to for ) corresponds to the feature size rc; meanwhile, the high-frequency oscillations in (c), (d) correspond to the distances between features in the images (separated by ). Both scale bars in (d) measure 20 pixels. |