Skip to main content
. 2019 May 25;23:101872. doi: 10.1016/j.nicl.2019.101872

Table 3.

Previous studies reporting weighted kappa (κw) values for intra- and inter-rater agreements together with the test set agreement between L.C. and AVRA (in bold text), and L.C and VGG16 as a reference. The interval given refers to the minimum and maximum κw value reported in the referenced study. The N column refers to the number of images used for the intra- and inter-rater assessment (if two values are given the number of images rated were different for the intra- and the inter-rater analysis). denotes if L.C. (whose ratings was used for training in this study) was one of the raters in the reported agreements.

Study Scale N Intra-rater agreement (κw) Inter-rater agreement (κw)
Cavallin et al. (2012) MTA 100 0.83–0.94 0.72–0.84
Cavallin et al. (2012b) MTA 100 0.84–0.85
Westman et al. (2011) MTA 100 0.93
Velickaite et al. (2017) MTA 20/50 0.79–0.84 0.6–0.65
Ferreira et al. (2017) MTA 120 0.89–0.94 0.70–0.71
Koedam et al. (2011) MTA 29/118 0.91–0.95 0.82–0.90
VGG16 MTA 464 1 0.58–0.59
AVRA MTA 464 1 0.72–0.74
Koedam et al. (2011) PA 29/118 0.93–0.95 0.65–0.84
Ferreira et al. (2017) PA 120 0.88 0.88
VGG16 PA 464 1 0.63
AVRA PA 464 1 0.74
Ferreira et al. (2016) GCA-F 100 0.70 0.59
Ferreira et al. (2017) GCA-F 120 0.83 0.79
VGG16 GCA-F 464 1 0.56
AVRA GCA-F 464 1 0.62