Skip to main content

Table 4 F -measure and minimal error δ min (given in brackets) obtained using different local skin color models

From: Self-adaptive algorithm for segmenting skin regions

γ p →

RP-based

DSPF-based

↓ Local skin color model

ECU data set

HGR data set

ECU data set

HGR data set

Gaussian

0.8535 (7.96%)

0.9135 (6.74%)

0.8 4 1 1 (7.2 2 %)

0.9 5 6 2 (2.5 2 %)

Histogram, 256 bins

0.8308 (9.42%)

0.9215 (5.80%)

0.8347 (8.00%)

0.9556 (2.51%)

Histogram, 128 bins

0.8 3 4 8 (9.2 0 %)

0.9 2 2 0 (5.7 8 %)

0.8353 (7.84%)

0.9555 (2.50%)

Histogram, 64 bins

0.8400 (8.82%)

0.9217 (5.86%)

0.8348 (7.56%)

0.9563 (2.43%)

Histogram, 32 bins

0.8328 (9.01%)

0.9145 (6.59%)

0.8257 (7.90%)

0.9559 (2.49%)

Histogram, 16 bins

0.8107 (9.95%)

0.9134 (6.67%)

0.8125 (8.62%)

0.9533 (2.69%)

Histogram, 8 bins

0.7906 (10.96%)

0.9138 (6.67%)

0.7982 (9.62%)

0.9481 (3.06%)

  1. The arrows indicate the column or row. Italicized values indicate the selected configuration.