Note: One of the our paper is under review now. After the review process is concluded, we will share our codes, pretrained models, and other necessary information/files.
For FERET dataset, there are 8 text files that contain ear and profile face image names with age and gender labels.
Codes, geometric features for our dataset, cropped ear and profile face coordinates will be included.
If you use these codes, please refers our papers:
[1] D. Yaman, F. I. Eyiokur, H. K. Ekenel, "Multimodal Age and Gender Classification Using Ear and Profile Face Images", Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition Workshops (CVPRW), 2019.
[2] D. Yaman, F. I. Eyiokur, N. Sezgin, H. K. Ekenel, "Age and Gender Classification from Ear Images", 6th IAPR/IEEE International Workshop on Biometrics and Forensics (IWBF), 2018.
Dataset | Train data | Test data |
---|---|---|
FERET | ||
UND-F | ||
UND-J2 |
We used VGG-16 and ResNet-50 deep CNN models.