Skip to main content
. 2016 Nov 20;13:264–270. doi: 10.1016/j.nicl.2016.11.020

Table 1.

Mean and standard deviation of dice coefficients (DC).

All
(n = 50)
< 5 ml
(n = 26)
5–15 ml
(n = 14)
> 15 ml
(n = 10)
Ham-Syd 0.67 (0.12) 0.63 (0.15) 0.69 (0.06) 0.74 (0.08)
Ham-Zur 0.66 (0.13) 0.61 (0.16) 0.68 (0.06) 0.74 (0.07)
Zur-Syd 0.67 (0.12) 0.63 (0.14) 0.69 (0.07) 0.72 (0.09)
Raters 0.66 (0.12) 0.62 (0.14) 0.69 (0.06) 0.73 (0.08)
Raters-LGA SPM 8 0.60 (0.15) 0.53 (0.16) 0.65 (0.08) 0.70 (0.09)
Raters-LGA SPM 12 0.53* (0.16) 0.45* (0.18) 0.59* (0.09) 0.63 (0.10)
Raters-LPA 0.57* (0.16) 0.49* (0.17) 0.63 (0.10) 0.68 (0.11)

Means and standard deviations (in brackets) of DC between lesion masks generated by manual or automated segmentation. The numbers marked with an asterisk indicate DC which are significantly (p = 0.05) different from the DC of the manual raters ("Raters"). In columns 3–5 the same analysis was performed but restricted to groups with different total lesion volumes.