Neuromorphic Vision Based Multivehicle Detection and Tracking for Intelligent Transportation System

Table 6

The tracking performance for EventSeq-Vehicle3 using different tracking methods (SORT, GM-PHD, GM-CPHD, and PDAF) with detection through DBSCAN clustering. The impact of time intervals from 10ms to 30ms was studied.

Tis	Tracker	MOTA	MOTP	MT	PT	ML	FP	FN	IDs	FM
10ms	SORT	24.6%	69.5%	1	34	24	1258	10838	109	611
10ms	GM-PHD	12.9%	68.9%	1	42	16	2484	10956	666	1903
10ms	GM-CPHD	12.5%	69.1%	2	41	16	3844	9957	363	1646
10ms	PDAF	13.8%	69.0%	0	38	21	2710	11173	70	2166
20ms	SORT	10.1%	69.3%	5	33	21	1990	5225	64	235
20ms	GM-PHD	21.4%	70.1%	5	38	16	1475	4699	188	490
20ms	GM-CPHD	13.4%	70.4%	5	36	18	2218	4698	94	402
20ms	PDAF	17.4%	70.2%	1	40	18	1682	4948	59	707
30ms	SORT	4.0%	69.0%	1	32	26	1329	3815	41	141
30ms	GM-PHD	14.3%	70.3%	7	35	17	1199	3311	120	285
30ms	GM-CPHD	6.1%	70.4%	4	34	21	1712	3300	58	240
30ms	PDAF	12.9%	70.6%	3	34	22	1112	3548	46	406