A reproduction of Pathak et al. 2016 together with Edwin Holst. Semantic hole-filling/inpainting using convolutional autoencoders.
The notebooks are very messy and we might release a clean version in the future.
This is a PyTorch implementation.
Read the paper in pdf format or check out the overleaf source: https://www.overleaf.com/read/hhqwqscqjwrh
This was created as part of a university course at Chalmers University of Technology, SSY340 Deep Machine Learning by Lennart Svensson
https://arxiv.org/abs/1604.07379
Pathak, D., Krahenbuhl, P., Donahue, J., Darrell, T., & Efros, A. A. (2016). Context encoders: Feature learning by inpainting. In Proceedings of the IEEE conference on computer vision and pattern recognition (pp. 2536-2544).
