PixColor: Pixel Recursive Colorization

Sergio Guadarrama, Ryan Dahl, David Bieber, Jonathon Shlens, Mohammad Norouzi and Kevin Murphy

Abstract

We propose a novel approach to automatically produce multiple colorized versions of a grayscale image. Our method results from the observation that the task of automated colorization is relatively easy given a low-resolution version of the color image. We first train a conditional PixelCNN to generate a low resolution color for a given grayscale image. Then, given the generated low-resolution color image and the original grayscale image as inputs, we train a second CNN to generate a high-resolution colorization of an image.

Session

Posters

Files

PDF iconPaper (PDF)
PDF iconSupplementary (PDF)

DOI

10.5244/C.31.112
https://dx.doi.org/10.5244/C.31.112

Citation

Sergio Guadarrama, Ryan Dahl, David Bieber, Jonathon Shlens, Mohammad Norouzi and Kevin Murphy. PixColor: Pixel Recursive Colorization. In T.K. Kim, S. Zafeiriou, G. Brostow and K. Mikolajczyk, editors, Proceedings of the British Machine Vision Conference (BMVC), pages 112.1-112.13. BMVA Press, September 2017.

Bibtex

            @inproceedings{BMVC2017_112,
                title={PixColor: Pixel Recursive Colorization},
                author={Sergio Guadarrama, Ryan Dahl, David Bieber, Jonathon Shlens, Mohammad Norouzi and Kevin Murphy},
                year={2017},
                month={September},
                pages={112.1-112.13},
                articleno={112},
                numpages={13},
                booktitle={Proceedings of the British Machine Vision Conference (BMVC)},
                publisher={BMVA Press},
                editor={Tae-Kyun Kim, Stefanos Zafeiriou, Gabriel Brostow and Krystian Mikolajczyk},
                doi={10.5244/C.31.112},
                isbn={1-901725-60-X},
                url={https://dx.doi.org/10.5244/C.31.112}
            }