Table 5 The statistics of the experimental data of image edge features extracted by four deep learning models.
Input Image AP（%） Model | Notebook Computer | Bicycle | Zebra | Flower |
|---|---|---|---|---|
DBN | 90.87 | 92.95 | 91.33 | 91.08 |
DS-DSA | 92.16 | 92.71 | 91.52 | 92.29 |
LDiffCNN | 92.45 | 91.27 | 92.02 | 91.46 |
IFDL | 94.77 | 93.08 | 95.83 | 94.27 |
Input Image F-measure [0,1] Model | Notebook Computer | Bicycle | Zebra | Flower |
DBN | 0.90 | 0.91 | 0.90 | 0.92 |
DS-DSA | 0.92 | 0.94 | 0.92 | 0.93 |
LDiffCNN | 0.94 | 0.92 | 0.91 | 0.93 |
IFDL | 0.95 | 0.94 | 0.93 | 0.95 |
