Skip to main content

Table 3 Performance comparison between CytoSpectre (CS) and FibrilTool (FT) using synthetic images of cell clusters

From: CytoSpectre: a tool for spectral analysis of oriented structures on cellular and subcellular levels

  CS orientation, error mean ± std (deg) FT orientation, error mean ± std (deg) Wilcoxon signed rank test p-value CS circular variance, Pearson’s r FT isotropy index, Pearson’s r
Non-degraded 1.7884 ± 3.6128 1.3904 ± 3.5042 0.0111 0.39844 0.31512
Blurred 1.8748 ± 3.9533 1.3432 ± 3.0325 0.0069 0.39645 0.31258
Gaussian noise 1.7101 ± 3.4064 6.4346 ± 10.067 4.0933E-14 0.38642 0.15947
Poisson noise 1.8325 ± 3.623 3.3052 ± 3.0916 1.0166E-08 0.38501 0.22644
  1. Absolute errors in degrees (mean ± standard deviation) between true mean orientation values and values estimated by CS and FT are shown for images of the cell cluster dataset (N = 100) without degradations, with moderate blurring (kernel std 2.5 pixels), with moderate Gaussian noise (var. 1 %) and with moderate Poisson noise (level 10). Paired, two-sided Wilcoxon signed rank tests were performed to compare the orientation errors of CS and FT and the resulting p-values are given for each case. Pearson’s linear correlation coefficients between true and estimated measures of isotropy are also shown for the corresponding cases