TABLE 2B.

Classification precision and recall for the test sets per trial, for all three data augmentation strategies.

		(precision/recall)
Random data augmentation Set	T	Trial 1	Trial 2	Trial 3	Trial 4	Trial 5	Trial 6	Mean (S.D.)
crop	{T1} ∈ T	45.5%/55.6%	42.9%/66.7%	64.3%/100%	50.0%/100%	56.3%/100%	44.4%/88.9%	50.6% (8%)/85.2% (19%)
crop, flip	{T1, T2} ∈ T	36.4%/44.4%	50.0%/66.7%	52.9%/100%	53.3%/88.9%	52.9%/100%	50.0%/100%	49.3% (6%)/83.3% (23%)
crop, flip, and rotation	{T1, T2, T3} ∈ T	50.0%/66.7%	38.5%/55.6%	57.1%/88.9%	50.0%/100%	42.9%/66.7%	52.9%/100%	48.6% (7%)/79.7% (19%)

Data are shown as percentage, and each pair of numbers represents the precision and recall values, respectively.