Table 1.
List of studies used different deep segmentation methods of breast cancer mammograms images.
| Ref# | Year | Segmentation method | Segmentation accuracy (dice coefficient index) | Classifier | Dataset | Classification accuracy |
|---|---|---|---|---|---|---|
| [37] | 2020 | Vanilla U-net | 95.1% | VGG-16 | CBIS-DDSM, INbreast, UCHCDM, BCDR-01 | 92.6% |
| [32] | 2019 | RU-Net | 98.3% | ResNet | INbreast | 98.7% |
| [34] | 2019 | U-Net integrated AGs | 82.24% | — | DDSM | 78.38% |
| [29] | 2018 | FrCN | 92.69% | CNN | INbreast | 95.64% |
| [35] | 2015 | CRF | 90% | — | DDSM-BCRP and INbreast | — |
| [33] | 2020 | cGAN | 98% | CNN based on BI-RADS | Abreast | 97.85% |
| [39] | 2020 | DSPAE | — | Linear classifier | MIAS DDSM |
97.54% 98.13% |