Image & Shape Manipulation

Wednesday, 27 July, 10:45 am - 12:35 pm, Anaheim Convention Center, Ballroom C
Session Chair: Connelly Barnes, University of Virginia

cSculpt: A System for Collaborative Sculpting

cSculpt is a real-time collaborative modellng system for digital sculpting and appearance painting, taking advantage of a multi-resolution algorithm for merging users' edits.

Claudio Calabrese
Sapienza – Università di Roma

Gabriele Salvati
Sapienza – Università di Roma

Marco Tarini
Università degli Studi dell'Insubria

Fabio Pellacini
Sapienza – Università di Roma

StyLit: Illumination-Guided Example-Based Stylization of 3D Renderings

StyLit is a new approach to example-based stylization of 3D scenes. It incorporates illumination-specific effects and an improved synthesis algorithm, and it better preserves the visual richness of hand-created style exemplars.

Jakub Fišer
Czech Technical University in Prague

Ondřej Jamriška
Czech Technical University in Prague

Michal Lukáč
Czech Technical University in Prague

Eli Shechtman
Adobe Research

Paul Asente
Adobe Research

Jingwan Lu
Adobe Research

Daniel Sykora
Czech Technical University in Prague

Flow-Guided Warping for Image-Based Shape Manipulation

This paper presents a method that manipulates perceived object shape from a single input image. It produces the illusion of shape sharpening or rounding by modifying orientation patterns, strongly correlated to surface curvature. The warping algorithm produces convincing shape manipulation results for a variety of materials and lighting environments.

Romain Vergne
Université Grenoble Alpes, Centre national de la recherche scientifique, INRIA

Pascal Barla

Georges-Pierre Bonneau
Université Grenoble Alpes, Centre national de la recherche scientifique, INRIA

Roland W. Fleming
Justus-Liebig-Universität Gießen

Transfiguring Portraits

This system's goal is to let people imagine and explore how they may look in a different country, era, hairstyle, hair color, age, and practically anything else that can be queried in an image search engine.

Ira Kemelmacher-Shlizerman
University of Washington

Band-Sifting Decomposition for Image-Based Material Editing

This paper introduces a set of image filters to alter the perceived properties of the observed materials in a natural way. The filters construct a multi-scale decomposition and modify three axes: scale, sign, and amplitude of the sub-band coefficients. User studies validate the perceptual effects of the filters.

Ivaylo Boyadzhiev
Cornell University

Kavita Bala
Cornell University

Sylvain Paris
Adobe Research

Edward H. Adelson
Massachusetts Institute of Technology