Research Article

Enhancement of Background Subtraction Techniques Using a Second Derivative in Gradient Direction Filter

Table 3

Performance of each technique.

NameMethodEvaluation performance based on average value
-scoreAccuracy (%)
Without SDGDWith SDGD%im*Without SDGDWith SDGD%im*

A1FD0.750.78398.4598.470.02
AM0.690.75698.1698.290.13
RA0.740.77398.4298.440.02
RGA0.790.81298.4298.470.03

A2FD0.780.79196.7496.65−0.09
AM0.800.81197.3197.14−0.17
RA0.760.78296.9196.83−0.08
RGA0.760.77195.7595.65−0.10

B1FD0.680.74697.9998.120.13
AM0.710.76598.2298.370.15
RA0.710.76598.2598.380.13
RGA0.640.68497.4797.510.04

C1FD0.840.85195.6495.690.05
AM0.790.80194.8795.030.16
RA0.570.65891.6992.170.48
RGA0.80.81194.9595.070.12

D1FD0.680.71397.8498.070.23
AM0.710.75498.1398.370.22
RA0.710.74398.1498.360.22
RGA0.700.74498.0498.250.21

E1FD0.800.81196.6696.690.03
AM0.800.81196.7096.710.01
RA0.800.81196.6996.700.01
RGA0.680.77996.4596.500.05

im: percent improvement.