Object Detection and Movement Tracking Using Tubelets and Faster RCNN Algorithm with Anchor Generation

Table 7

Performance comparison of the proposed approach with the existing approaches.

Methods	No. of frames	Specificity (%)	Precision (%)	Recall (%)	Accuracy (%)	TP rate (%)	FP rate (%)	Error detection rate (%)
Faster RCNN	572	0.7384	0.8513	0.8411	0.9235	0.9185	0.8331	0.9076
Kai et al. [25]	463	0.7572	0.8358	0.8824	0.9088	0.8938	0.8578	0.9012
Mihir et al. [9]	546	0.7264	0.8488	0.8517	0.9124	0.8754	0.8841	0.0876
Proposed approach	575	0.7841	0.9133	0.9357	0.9788	0.9341	0.9544	0.0212