Adience Benchmark 数据集注释说明

Each fold is described by a csv file with 12 columns:

1. user_id - the folder in the dataset containing the image.
2. original_image - image name in the dataset.
3. face_id - the Face ID in the original Flickr image, can be ignored.
4. age - age label of the face.
5. gender - gender label of the face.
6. x, y, dx, dy - bounding box of the face in the original Flickr image, can be ignored.
7. tilt_ang, fiducial_yaw_angle - pose of the face in the original Flickr image, can be ignored.
8. fiducial_score - score of the landmark detector, can be ignored.