Facial Appearance Modifications using SKPCA-Derived Features Extracted from Convolutional Autoencoder's Latent Space
labels.csv - contains additional information about images from the Celeb data set.
The project has been divided into six parts
- Training CAE on Celeb data set.
- Encoding images using trained CAE and merging it with labels
- Choosing the representative part of the encoded images to learn SKPCA (SKPCA consumes a lot of memory and therefore has to operate on a limited number of samples. During experiment we used 14662 samples)
- Calculating SKPCA
- Calculating Inverse of SKPCA
- Sample images modyfication based on SKPCA + CAE model