A New Multiface Target Detection Algorithm for Students in Class Based on Bayesian Optimized YOLOv3 Model

Table 4

Test results in COCO dataset.

Algorithm	Total number	Accurate detection number	Total detection number	Accuracy rate (%)	Recall (%)	Detection speed (Fps)	Parameter (MB)
Fast R-CNN	500	445	590	75.42	89.00	38.5	15.32
FaceNet	500	437	589	74.19	87.40	35.4	18.61
DeepID	500	456	601	75.87	91.20	35.6	17.85
SSD	500	453	586	77.30	90.60	37.8	14.83
YOLOv1	500	455	597	76.21	91.00	36.2	15.65
YOLOv2	500	463	585	79.15	92.60	37.5	16.21
YOLOv3	500	470	592	79.39	94.00	38.3	17.52
Ours	500	479	587	81.60	95.80	37.9	16.55