Table 7 Skin detection scores obtained using different methods for the HGR data set
| Â | Acceptance threshold | Acceptance threshold | ||||||
|---|---|---|---|---|---|---|---|---|
Method | set to maximize F-measure | set to minimize | ||||||
| Â | F-measure | prec | rec |
|
| rec | δ fp |
|
Global Bayesian classifier [16] | 0.9031 | 89.72% | 90.92% | 5.11% | 7.05% | 91.53% | 5.63% | 5.63% |
Global model in ESS [14] | 0.9090 | 90.76% | 91.04% | 4.31% | 6.56% | 91.81% | 4.92% | 4.9 2 % |
Chen’s global model [15] | 0.8607 | 89.33% | 83.03% | 7.26% | 12.11% | 83.03% | 7.26% | 7.26% |
Wavelet-based hybrid detector [61] | 0.8991 | 91.38% | 88.49% | 3.9 1 % | 7.50% | 90.24% | 5.25% | 5.38% |
Spatial analysis using RP [59] | 0.9086 | 87.77% | 94.17% | 29.83% | 5.91% | 94.49% | 6.30% | 43.18% |
Spatial analysis using DSPF [9] | 0.9391 | 92.90% | 94.94% | 8.39% | 3.06% | 96.24% | 2.37% | 8.94% |
Proposed method (RP-based) | 0.9220 | 92.66% | 91.74% | 26.52% | 5.78% | 91.85% | 3.41% | 31.88% |
Proposed method (DSPF-based) | 0.9 5 6 2 | 9 5.3 2 % | 9 5.9 2 % | 10.51% | 2.5 2 % | 9 6.9 2 % | 1.9 6 % | 11.71% |