This is the repository for the predicting facial attractiveness using landmark features and gabor filters.
Features obtained from:
- Facial Distances
- Facial Ratios
- Gabor Filters
How to run:
- All features first must be generated by running generate_features.py
- Then you can train your own model by running train.py
- You can test with a single image by running demo.py (give the path at the beginning)
Required libraries: Dlib, OpenCV, numpy, scipy,sklearn,imutils
Validation results:
Demo:
Please see http://www.hcii-lab.net/data/SCUT-FBP/EN/introduce.html to download the entire dataset.