An Object Detection Method Using Wavelet Optical Flow and Hybrid Linear-Nonlinear Classifier

Table 2

Classification performance of PCNN, FNN, GSVM, LDA, and HLNLC.

Video	PCNN	FNN	GSVM	LDA	HLNLC
Cars
M and SD	57.2 ± 16.7	55.4 ± 15.5	41.3 ± 28.5	37.8 ± 11.9	36.2 ± 10.4
AUC	0.65	0.53	0.77	0.78	0.81
DR	72.33%	71.96%	84.81%	88.27%	89.32%
Coastguard
M and SD	44.3 ± 13.2	42.9 ± 12.4	34.7 ± 25.6	32.6 ± 12.5	31.4 ± 11.7
AUC	0.72	0.69	0.81	0.80	0.85
DR	73.43%	72.16%	84.54%	87.29%	89.52%
Satellite
M and SD	32.3 ± 10.3	30.7 ± 10.7	25.8 ± 21.6	31.3 ± 13.8	30.2 ± 13.4
AUC	0.71	0.61	0.83	0.80	0.82
DR	72.33%	73.48%	84.49%	86.25%	88.64%
Running
M and SD	35.7 ± 11.4	33.6 ± 10.9	25.5 ± 22.6	33.4 ± 16.2	32.7 ± 15.9
AUC	0.76	0.64	0.85	0.83	0.84
DR	72.31%	73.14%	84.39%	88.17%	89.09%