BIRD

Bayesian Inference for Reflection Detection (BIRD)

  • Our technique BIRD detects reflections formed by two uncorrelated layers. The layers must have distinctive feature points.
  • Parameter values for Manual Configuration (T_1,T_s) are shown in caption. Here, T_s is set to 3.15 in 11 out of 15 sequences
  • Reflections detected are shown in green.

SelimK1

Figure 1.a: (T_1,T_s) = (-0.14,3.15)
Figure 1.b: Automated Parameters Configuration

SelimK2

Figure 2.a: (T_1,T_s) = (-0.12,3.15)
Figure 2.b: Automated Parameters Configuration

SelimC

Figure 3.a: (T_1,T_s) = (0,4)
Figure 3.b: Automated Parameters Configuration

SelimH

Figure 4.a: (T_1,T_s) = (0.-13,3.15)
Figure 4.5: Automated Parameters Configuration

BuildOnWind

Figure 5.a: (T_1,T_s) = (0,3.15)
Figure 5.b: Automated Parameters Configuration

PHouse

Figure 6.a: (T_1,T_s) = (0.00,3.15)
Figure 6.b: Automated Parameters Configuration

GirlRef

Figure 7.a: (T_1,T_s) = (-0.01,3.15)

BKelss

Figure 8.a: (T_1,T_s) = (-0.01,3.15)
Figure 8.b: Automated Parameters Configuration

BluWind

Figure 9.a: (T_1,T_s) = (-0.01,3.15)
Figure 9.b: Automated Parameters Configuration

BusRef

Figure 10: (T_1,T_s) = (0.01,3)

GaryMik

Figure 11.a: (T_1,T_s) = (0.05,3.15)
Figure 11.b: Automated Parameters Configuration

RedPPL

Figure 12: (T_1,T_s) = (-0.01,3.15)

CarRef

Figure 13: (T_1,T_s) = (-0.21,4.15)

PortraitC

Figure 14.a: (T_1,T_s) = (-0.05,3.15)
Figure 14.b: Automated Parameters Configuration

SHADOW: An Example of Shadow Detection

Figure 15.a: (T_1,T_s) = (-0.17,3.85)
Figure 15.b: Automated Parameters Configuration

Page last modified on November 03, 2010