This repository contains the source code for my proposed colorization method. It consists of an exemplar-based method which is based on a classification framework and works at superpixel level for enhanced coherence and performance.
It was built upon the project I forked from, but at this stage it holds no resemblance.
The method's pipeline includes third party code:
- The superpixel segmentation in:
Levinshtein, A., Stere, A., Kutulakos, K. N., Fleet, D. J., Dickinson, S. J., & Siddiqi, K. (2009). Turbopixels: Fast superpixels using geometric flows
- The saliency maps in:
Yang, C., Zhang, L., Lu, H., Ruan, X., & Yang, M. H. (2013). Saliency detection via graph-based manifold ranking.
- Histogram matching functions
- The scribble propagation algorithm:
Levin, A., Lischinski, D., & Weiss, Y. (2004, August). Colorization using optimization.
- Parts from the colorization algorithm of:
Gupta, R. K., Chia, A. Y. S., Rajan, D., Ng, E. S., & Zhiyong, H. (2012, October). Image colorization using similar images
- The script
single_colorizationperforms the colorization for the input pair and parameters defined in the file <./input/single.in> - Features weights should be adjusted in order to achieve better results for each input pair.
The code was developed for experimental purposes, do not expect more than this.