Research Article
Hotspot Detection with Machine Learning Based on Pixel-Based Feature Extraction
Table 2
Pixel-based feature extraction comparison results.
| Case | Accuracy (%) | False alarm rate | Efficiency (%) | Pixel | Edge | Fourier | Pixel | Edge | Fourier | Pixel | Edge | Fourier |
| B1 | 91.23% | 85.53% | 91.67% | 33.27 | 23.48 | 48.74 | 2.74% | 3.64% | 1.88% | B2 | 56.71% | 70.34% | 78.16% | 18.12 | 33.90 | 46.02 | 3.13% | 2.08% | 1.70% | B3 | 92.60% | 99.20% | 97.18% | 48.41 | 96.58 | 284.91 | 1.91% | 1.03% | 0.34% | B4 | 47.54% | 66.67% | 78.14% | 32.22 | 116.68 | 22.54 | 1.48% | 0.57% | 3.47% | B5 | 54.76% | 54.76% | 78.57% | 58.90 | 32.31 | 15.88 | 0.93% | 1.69% | 4.95% | Avg. | 0.80 | 0.88 | 1.00 | 1.28 | 1.75 | 1.00 | 1.90 | 1.34 | 1.00 |
|
|