TABLE 1.
Comparison of results of other models before and after combining modulation-enhancement network (MENet) on BSDS500 data set.
| Method | MENet | ODS | OIS | AP | |
| HED (Xie and Tu, 2015) | × | SS | 0.788 | 0.808 | 0.840 |
| √ | 0.791 (↑0.003) | 0.811 (↑0.003) | 0.809 | ||
| RCF (Liu et al., 2017) | × | MS | 0.811 | 0.830 | – |
| √ | 0.812 (↑0.001) | 0.833 (↑0.003) | 0.871 | ||
| LRC (Lin et al., 2020) | × | MS | 0.816 | 0.836 | 0.864 |
| √ | 0.817 (↑0.001) | 0.839 (↑0.003) | 0.867 (↑0.003) |
“×” means without MENet and “√” means with MENet. SS, single-scale results; MS, multi-scale results. Except for HED (Xie and Tu, 2015), the results are on BSDS500-VOC. “↑”indicates improved results.