Demo Video - https://www.youtube.com/watch?v=_fRp408fauI
- Dataset include JAFFE, CK+ and a personalised dataset of illumination variant images
- Pre-processing included cropping and alignment to face using Multi-Task Cascaded Neural Networks (MTCNNs).
- Images are converted to Grayscale with application of Contrast Limited Adaptive Histogram Equalization (CLAHE).
- CNN used is VGG-16.
- Trained on CPU mode on Caffe. Used Python 2.7.
https://drive.google.com/drive/folders/183DCLh7DfJ-pF-ktyROjbPaxPmbd69UC?usp=sharing
Email me at [email protected] for questions.