Table 5.
DL-based mammography for breast tumor classification.
| Reference | Year | Method | Database | Number of images | Accuracy | AUC | Sensitivity | Precision | F1-Score |
|---|---|---|---|---|---|---|---|---|---|
| (95) | 2017 | Transfer learning, Random Forest | INbreast | 108 | 90% | NA | 98% | 70% | NA |
| (124) | 2018 | Deep GeneRAtive Multitask | CBIS-DDSM | NA | 89% | 0.884 | NA | NA | NA |
| (125) | 2019 | VGG, Residual Network | CBIS-DDSM | NA | NA | NA | 86.10% | 80.10% | NA |
| (126) | 2019 | DCNN, Alexnet | CBIS-DDSM | 1696 | 75.0% | 0.80 | NA | NA | NA |
| (127) | 2019 | MA-CNN | MIAS | 322 | 96.47% | 0.99 | 96.00% | NA | NA |
| (128) | 2019 | DCNN, MSVM | MIAS | 322 | 96.90% | 0.99 | NA | NA | NA |
| (129) | 2019 | CNN Improvement (CNNI-BCC) |
MIAS | NA | 90.50% | 0.90 | 89.47% | 90.71% | NA |
| (130) | 2020 | MobileNet, VGG, Resnet, Xception |
CBIS-DDSM | 1696 | 84.4% | 0.84 | NA | NA | 85.0% |
| (131) | 2020 | MobilenetV1, MobilenetV2 | CBIS-DDSM | 1696 | 74.5% | NA | NA | 70.00% | 76.00% |
| (132) | 2020 | DE-Ada* | CBIS-DDSM | NA | 87.05% | 0.9219 | NA | NA | NA |
| (133) | 2020 | AlexNet | MIAS | 68 | 98.53% | 0.98 | 100% | 97.37% | 98.3% |
| (133) | 2020 | GoogleNet | MIAS | 68 | 88.24% | 0.94 | 80% | 94.74% | 85.71% |
| (134) | 2020 | Inception ResNet V2 | INbreast | 107 | 95.32% | 0.95 | NA | NA | NA |
| (132) | 2020 | De-ada* | INbreast | NA | 87.93% | 0.9265 | NA | NA | NA |
| (135) | 2021 | CNN | CBIS-DDSM | 1592 | 91.2% | 0.92 | 92.31% | 90.00% | 91.76% |
| (136) | 2021 | MobilenetV2, Nasnet Mobile, MEWOM | CBIS-DDSM | 1696 | 93.8% | 0.98 | 93.75% | 93.80% | 93.77% |
| (137) | 2021 | ResNet-18, (ICS-ELM) | MIAS | 322 | 98.13% | NA | NA | NA | NA |
| (135) | 2021 | CNN | MIAS | 322 | 93.39% | 0.94 | 92.72% | 94.12% | 93.58% |
| (136) | 2021 | Mobilenet V2 & NasNet Mobile, MEWOA |
MIAS | 300 | 99.80% | 1.00 | 99.00% | 99.33% | 99.16% |
| (137) | 2021 | ResNet-18, (ICS-ELM) | INbreast | 179 | 98.26% | NA | NA | NA | NA |
| (135) | 2021 | CNN | INbreast | 387 | 93.04% | 0.94 | 94.83% | 91.23% | 93.22% |
| (136) | 2021 | Fine-tuned MobilenetV2, Nasnet, MEWOM | INbreast | 108 | 99.7% | 1.00 | 99.0% | 99.0% | 99.0% |
| (123) | 2022 | DualCoreNet | CBIS-DDSM | NA | NA | 0.85± 0.021 | NA | NA | NA |
| (138) | 2022 | CNN classifier with different fine-tuning |
DDSM | 13128 | 99.96% | 1.00 | 100% | 99.92% | 99.96% |
N/A, Not Applicate.