Table 1.

Synthetic mistriggering artefact data classification results for mean accuracy (A), precision (P), recall (R) and balanced accuracy (BA) results. A 10-fold cross validation was used and each image was labelled once over all folds and standard deviation over folds is reported (mean ± std). All results are multiplied by 1000 and the bold font highlights the best results.

Methods	A	P	R	BA
K-Nearest Neighbours	742 ± 25	742 ± 33	746 ± 40	744 ± 37
Linear SVM	748 ± 36	743 ± 89	749 ± 41	746 ± 73
Decision Tree	756 ± 42	757 ± 46	751 ± 33	754 ± 41
Random Forests	787 ± 45	782 ± 78	786 ± 62	784 ± 67
Adaboost	783 ± 37	781 ± 60	778 ± 73	779 ± 66
Naive Bayesian	809 ± 65	796 ± 48	804 ± 57	800 ± 52
Variance of Laplacian	802 ± 42	799 ± 62	803 ± 79	802 ± 41
NIQE	922 ± 56	919 ± 72	925 ± 82	923 ± 71
Lorch et al. (2017)	893 ± 62	892 ± 83	894 ± 49	893 ± 22
3D CNN	961 ± 79	957 ± 101	959 ± 87	958 ± 74
LRCN	963 ± 45	963 ± 33	965 ± 41	964 ± 38