Learning Diverse Image Colorization
Authors:
- Aditya DeshpandeADAditya Deshpande
- Jiajun LuJLJiajun Lu
- Mao-Chuang YehMYMao-Chuang Yeh
- Min Jin ChongMCMin Jin Chong
- David ForsythDFDavid Forsyth
- 1 July 2017
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- No. 10636919,p. 2877-2885
Abstract
Colorization is an ambiguous problem, with multiple viable colorizations for a single grey-level image. However, previous methods only produce the single most probable colorization. Our goal is to model the diversity intrinsic to the problem of colorization and produce multiple colorizations that display long-scale spatial co-ordination. We learn a low dimensional embedding of color fields using a variational autoencoder (VAE). We construct loss terms for the VAE decoder that avoid blurry outputs and take into account the uneven distribution of pixel colors. Finally, we build a conditional model for the multi-modal distribution between grey-level image and the color field embeddings. Samples from this conditional model result in diverse colorization. We demonstrate that our method obtains better diverse colorizations than a standard conditional variational autoencoder (CVAE) model, as well as a recently proposed conditional generative adversarial network (cGAN).Keywords
All Related Versions
This publication has 11 references indexed in Scilit:
- Attribute2Image: Conditional Image Generation from Visual Attributes2016
- Let there be color!ACM Transactions on Graphics, 2016
- Learning Large-Scale Automatic Image Colorization2015
- Deep Colorization2015
- ImageNet Large Scale Visual Recognition ChallengeInternational Journal of Computer Vision, 2015
- Loss-Specific Training of Non-Parametric Image Restoration Models: A New State of the Art2012
- Diverse M-Best Solutions in Markov Random Fields2012
- Automatic Image Colorization Via Multimodal Predictions2008
- Colorization using optimizationACM Transactions on Graphics, 2004
- Transferring color to greyscale images2002