This repository contains code to reproduce the face rotation experiments in section 4.1 of the paper Towards Automatic Image Editing: Learning to See another You, a BMVC 2016 paper by Amir Ghodrati*, Xu Jia*, Marco Pedersoli, Tinne Tuytelaars (* Amir and Xu contributed equally).

We propose a method that aims at automatically editing an image by altering its attributes. More specifically, given an image of a certain class (e.g. a human face), the method should generate a new image as similar as possible to the given one, but with an altered visual attribute (e.g. the same face with a new pose or a different illumination).

If you use our code in your research, please cite following paper:

@inproceedings{face_bmvc16,
  author    = {Amir Ghodrati and Xu Jia and Marco Pedersoli and Tinne Tuytelaars},
  title     = {Towards Automatic Image Editing: Learning to See another You},
  booktitle = {BMVC},
  year      = {2016}
}

Face Rotation:

Change Illumination
![face illumination] (./images/fig_ill.png)
Image Inpainting

• Install cuda 7.5, [Lasagne with small modification] (https://github.com/stephenjia/Lasagne), Theano 0.9.0dev2.dev-e4e08782d3a10d010d3a99bc87fd0fc3b0465405, cuDNN 5.0

• Run the experiments for face rotation:
  First change the configuration of the experiment in config_stageX_color.py Then run

python train_stage1_color.py

to train the first stage model and write checkpoint files to the checkpoints directory.
Then run

python train_stage2_color.py

to train the second stage model and write checkpoint files to the checkpoints directory.
To evaluate the trained models on test data, run

python generate_triplet_demo_color.py

You need to first download MultiPIE dataset. To crop and align faces, we use the code provided by Junho Yim for their cvpr15 paper titled Rotating Your Face Using Multi-task Deep Neural Network. Please cite their paper if you use this code for face cropping and alignment.