This repo is organized as follows:
├── 54532-0.jpg
├── checkpoints
│ └── coral_age_epoch_33.pth
├── config
│ └── config_coral.py
├── dataset
│ └── dataset_coral.py
├── models
│ ├── __pycache__
│ │ └── resnet_coral.cpython-36.pyc
│ └── resnet_coral.py
├── test.py
└── train_age.py1.pytorch >=0.4
2.opencv >= 3.4
3.numpy ...
- CACD: http://bcsiriuschen.github.io/CARC/
- UTKFace: https://susanqq.github.io/UTKFace/
- AFAD: https://github.com/afad-dataset/tarball
- MORPH-2: https://www.faceaginggroup.com/morph/
Download dataset into dataset dir,and then use face detector and landmark do face alignment like insightface.
| AFAD | Morph II | |
|---|---|---|
| SSR | 3.16 | |
| C3AE | 2.75 | |
| ORCNN | 3.66 | 3.27 |
| Ours | 3.40 | 2.65 |
https://github.com/derronqi/C3AE_Age_Estimation.git