Example-Based-Image-Colorization-Using-Primal-Dual-Algorithm-and-Saliency-Screening

This matlab pseudo code is for "Example-Based Image Colorization Using Primal-Dual Algorithm and Saliency Screening".

Requirements

Matlab R2018b

Usage

1. Put the test image into the "input" directory. Grayscale images are named with the prefix "gray_" and the corresponding reference images are with prefix "ref_". The images pairs in the directory are collected from previous publications.
2. Run 'demo.m'.
3. Results will be generated in the "output" directory.

Third-party source codes

The third-party source codes in the directory "3rdparty" are from these sources.

• A. Levinshtein, A. Stere, K. N. Kutulakos, D. J. Fleet, S. J. Dickinson, and K. Siddiqi, “Turbopixels: Fast superpixels using geometric flows,” IEEE transactions on pattern analysis and machine intelligence, vol. 31, no. 12, pp. 2290– 2297, 2009.

• E. Tola, V. Lepetit, and P. Fua, “Daisy: An efficient dense descriptor applied to wide-baseline stereo,” IEEE transactions on pattern analysis and machine intelligence, vol. 32, no. 5, pp. 815–830, 2009.

• W. Zhu, S. Liang, Y. Wei, and J. Sun, “Saliency optimization from robust background detection,” in Proceedings of the IEEE conference on computer vision and pattern recognition, 2014, pp. 2814–2821.

• X. Guo, Y. Li, J. Ma, and H. Ling, “Mutually guided image filtering,” IEEE transactions on pattern analysis and machine intelligence, 2018.