Table 2.

Experiment results

Name	Acc	Acc Avg|Std	F1	F1 Avg|Std	AUC	AUC Avg|Std
FC, Mobilenet v2, augmented with no color	100.0–82.0%	94.8% | 5.3	100.0–81.8	94.8 | 5.3	100.0–95.5	99.2 | 1.3
FC, Mobilenet v2, augmented	100.0–87.1%	93.8% | 3.8	100.0–86.8	93.7 | 3.9	100.0–93.7	99.0 | 1.6
FC, Mobilenet v2, no augmentation	100.0–76.9%	92.8% | 6.1	100.0–76.8	92.7 | 6.1	100.0–94.1	98.7 | 1.6
Random Forest, Mobilenet v2 augmented	97.4–84.6%	92.1% | 4.0	97.4–84.5	92.0 | 4.1	100.0–95.6	98.7 | 1.3
XGBoost, Mobilenet v2, augmented with no color	97.4–82.0%	91.1% | 5.1	97.4–81.2	90.9 | 5.2	100.0–93.2	97.8 | 2.2
XGBoost, Mobilenet v2, augmented	97.4–76.9%	91.1% | 5.5	97.4–76.8	91.0 | 5.5	100.0–89.7	98.0 | 2.8
Random Forest, Mobilenet v2, augmented with no color	94.8–82.0%	89.9% | 4.3	94.8–81.6	89.8 | 4.4	99.7–92.8	97.9 | 1.9
Decision Tree, Mobilenet v2, augmented	89.7–64.1%	80.0% | 7.7	89.5–63.8	79.7 | 7.8	89.5–63.9	80.1 | 8.0
Decision Tree, Mobilenet v2, augmented with no color	89.7–66.6%	79.3% | 6.6	89.6–65.8	79.0 | 6.8	90.3–67.7	79.8 | 6.6
Random Forest, Mobilenet v2, no augmentation	87.1–64.1%	76.9% | 6.9	87.1–63.8	76.7 | 7.0	94.8–75.6	85.2 | 5.4
XGBoost, Mobilenet v2, no augmentation	89.7–56.4%	75.3% | 11.6	89.7–55.9	75.1 | 11.7	95.7–65.7	83.0 | 9.4
Decision Tree, Mobilenet v2, no augmentation	84.6–46.1%	62.3% | 10.7	84.5–44.8	62.0 | 10.8	85.0–45.4	62.8 | 10.8