Gamasutra

Intrinsic Images

Wednesday, 27 July, 2:00 pm - 3:30 pm, Anaheim Convention Center, Ballroom E
Session Chair: Piotr Didyk, Universität des Saarlandes, Max-Planck-Institut für Informatik

Live Intrinsic Video

Intrinsic video decomposition is the basis for a variety of video manipulation applications, such as realistic recoloring, retexturing, material editing, and stylization of objects. This paper present sa novel variational approach to tackle this underconstrained inverse problem, which enables online processing of live video footage.

Abhimitra Meka
Max-Planck-Institut für Informatik

Michael Zollhoefer
Max-Planck-Institut für Informatik

Christian Richardt
Max-Planck-Institut für Informatik

Christian Theobalt
Max-Planck-Institut für Informatik

Multi-View Intrinsic Images of Outdoors Scenes With an Application to Relighting

This paper computes intrinsic images for multi-view sets of outdoor photos. Parameters are estimated using lighting simulation and optimization, and a robust visibility classifier computes intrinsic layers for relighting.

Sylvain Duchêne
INRIA

Clement Riant
INRIA

Gaurav Chaurasia
INRIA

Jorge Lopez-Moreno
INRIA

Pierre-Yves Laffont
ETH Zürich

Stefan Popov
INRIA

Adrien Bousseau
INRIA

George Drettakis
INRIA

Let There Be Color!: Joint End-to-End Learning of Global and Local Image Priors for Automatic Image Colorization With Simultaneous Classification

A novel end-to-end convolutional neural network-based model that combines both global image priors with local image features in order to automatically colorize greyscale images. The method realistically colorizes a diverse range of images.

Satoshi Iizuka
Waseda University

Edgar Simo-Serra
Waseda University

Hiroshi Ishikawa
Waseda University

Approximate Translational Building Blocks for Image Decomposition and Synthesis

This paper introduces approximate translational building blocks for unsupervised image decomposition, which is used to augment a conventional Markov random-field model for image synthesis.

Chuan Li
Johannes Gutenberg-Universität Mainz

Michael Wand
Johannes Gutenberg-Universität Mainz