Table 4 Comparison of the F1 measure applied on the spectrogram
From: A novel approach to extracting useful information from noisy TFDs using 2D local entropy measures
SNR | Proposed method | Local entropy-based algorithm | RICI TFD threshold | 5% | 10% | 15% |
|---|---|---|---|---|---|---|
Signal 1 | ||||||
−3 dB | F1= 0.865 | F1= 0.778 | F1= 0.868 | F1= 0.673 | F1= 0.785 | F1= 0.792 |
0 dB | F1= 0.909 | F1= 0.869 | F1= 0.896 | F1= 0.706 | F1= 0.85 | F1= 0.872 |
3 dB | F1= 0.873 | F1= 0.884 | F1= 0.939 | F1= 0.936 | F1= 0.91 | F1= 0.853 |
6 dB | F1= 0.924 | F1= 0.887 | F1= 0.94 | F1= 0.479 | F1= 0.479 | F1= 0.479 |
10 dB | F1= 0.932 | F1= 0.939 | F1= 0.918 | F1= 0.479 | F1= 0.479 | F1= 0.479 |
Signal 2 | ||||||
−3 dB | F1= 0.769 | F1= 0.763 | F1= 0.768 | F1= 0.699 | F1= 0.764 | F1= 0.742 |
0 dB | F1= 0.864 | F1= 0.823 | F1= 0.864 | F1= 0.772 | F1= 0.838 | F1= 0.798 |
3 dB | F1= 0.913 | F1= 0.793 | F1 =0.9 | F1= 0.9 | F1= 0.829 | F1= 0.74 |
6 dB | F1= 0.943 | F1= 0.915 | F1= 0.93 | F1= 0.477 | F1= 0.477 | F1= 0.477 |
10 dB | F1= 0.959 | F1= 0.942 | F1= 0.943 | F1= 0.477 | F1= 0.477 | F1= 0.477 |