Martin Guay



I am a research scientist in visual computing currently working on markerless motion capture. I am currently excited about combining vision with graphics, and did a few recent projects on AR such as automatically putting water tight clothes onto people in pictures.

Last year I got cought up in the personal assistant / chat bot hype and did a project on animating an autonomous 3D conversation avatar.

While I am doing more vision and machine learning now, my original background is graphics, and games. I spent two years in video games before doing a master's in fluid simulation and a ph.d. on character animation.



martin[dot]guay[at]disneyresearch[dot]com


Students and Collaborators

Dominik Borer, Ph.D. Student on Motion Models for 3D Computer Animation
Jakob Buhmann, Post-Doc on Motion Capture
Prof. Bob Sumner, Supervisor

Past:
Christoph Maurhofer, Master's on Poser and Costumes in AR.
Gökcen Cimen, Ph.D. on Character Motion Models for Interaction in AR and VR.
Victor Cornillere, Autonomous 3D Avatar Chat Bots.
Dominic Lutz, 3D animator on chat bot avatars.
Stelian Coros, Prof. on Synthetising Quadruped Motion.
Ye Yuan, Synthetising Quadruped Motion.
Simone Guggiari, Bachelor's on Interacting with a self-Balancing Character in AR and VR.
Dominik Borer, Master's on Trajectory Optimization for Simulated Character Control
Loïc Ciccone, Ph.D. on Character Animation Authoring.
Renzo Roth, Master's on Animated Music Band.
Marcel Lüdi, Master's on Deep Learning for Character Animation.

Remi Ronfard, Ph.D. co-advisor at INRIA.
Marie-Paule Cani, Ph.D. co-advisor at INRIA.
Richard Egli, Master's advisor at Sherbrooke U.


Publications and Projects


Pose and shape estimation from images and video

Marker-less motion capture for virtual production (multi-view and single view).
We are pushing the limits of pose and shape estimation from video data, leveraging 3D motion capture and special sauces.

Cloth


Animal pose estimation from a single image.
There are no large datasets of animal poses that are labelled.
We are working on methods to learn from other sources of data.
lion



Augmented Reality

Cloth

AR Costumes: Automatically Augmenting Watertight Costumes from a single RGB Image
ACM SIGGRAPH Conference on Virtual Reality and Applications in Industry (VRCAI), 2018.
Christoph Maurhofer, Gokcen Cimen, Mattia Ryffel, Robert W. Sumner, Martin Guay
PaperVideo


Cloth

AR Poser: Automatically Augmenting Mobile Pictures with Digital Avatars Imitating Poses
Computer Graphics, Visualization, Computer Vision and Image Processing (CGVCVIP), 2018.
Gokcen Cimen, Christoph Maurhofer, Robert W. Sumner, Martin Guay
PaperVideo


Cloth

Interacting with a Fully Simulated, Self-Balancing Bipedal Character in Augmented and Virtual Reality
Encyclopedia of Computer Graphics and Games, 2018.
Dominic Borer, Simone Giguarri, Robert W. Sumner, Martin Guay


Cloth

Modeling Physical Interactions with Real World Objects:
Interacting with Intelligent Characters in AR

Workshop on Artificial Intelligence in Virtual and Augmented Reality, 2017.
Gokcen Cimen, Ye Yuan, Stelian Coros, Robert W. Sumner, Martin Guay




Control of Digital Avatars, Parametric Character Motion

Cloth

Animating an Autonomous 3D Conversation Avatar

Computer Graphics, Visualization, Computer Vision and Image Processing (CGVCVIP), 2018.
Dominik Borer, Dominic Lutz, Robert W. Sumner, Martin Guay




Automatic Animation of a Digital Band from Midi Signals
Live Demo @ Symposium on Computer Animation (SCA), 2016.
Renzo Roth & CGL Ph.D. Students, Martin Guay





Variational Autoencoder Applied to Human Walking Motion: Automatically Learning
an Intuitive Animation Interface from a Collection of Human Motion Clips

Computer Graphics, Visualization, Computer Vision and Image Processing (CGVCVIP), 2017.
Marcel Ludi, Martin Guay, Brian McWilliams, Robert W. Sumner


Cloth

Keys-to-Sim: Transferring Hand-Crafted Key-framed Animations to Simulated Figures using
Wide Band Stochastic Trajectory Optimization

Computer Graphics and Visual Computing (CGVC), 2018.
Dominic Borer, Robert W. Sumner, Martin Guay
PaperVideoOnline Policy Learning PDFOnline Policy Video




An Intuitive Interface for Human Performance Tracking with Simulated Characters
Computer Graphics, Visualization, Computer Vision and Image Processing (CGVCVIP), 2017.
Gokcen Cimen, Martin Guay, Stelian Coros, Robert W. Sumner




Authoring Character Motion


The Line of Action: an Intuitive Interface for Expressive Character Posing
SIGGRAPH Asia 2013.
Martin Guay, Marie-Paule Cani, Remi Ronfard





Space-time sketching of character animation
SIGGRAPH 2015.
Martin Guay, Remi Ronfard, Michael Gleicher, Marie-Paule Cani




Adding dynamics to sketch-based character animation
Sketch-based modeling and interfaces (SBIM) 2015.
Martin Guay, Remi Ronfard, Michael Gleicher, Marie-Paule Cani





Authoring Motion Cycles
Symposium on Computer Animation (SCA), 2017.
Loïc Ciccone, Martin Guay, Mauri Nitti, Robert W. Sumner




Flow Curves: an Intuitive Interface for Coherent Scene Deformation
Pacific Graphics, 2016.
Loïc Ciccone, Martin Guay, Robert W. Sumner


Cloth
Programmable Animation Texturing using Motion Stamps
Pacific Graphics, 2016.
Antoine Milliez, Martin Guay, Marie-Paule Cani, Markus Gross, Robert W. Sumner




Physics Simulations on the GPU



Simple and Fast Fluids
GPU Pro 2, 2011.
Martin Guay, Fabrice Colin, Richard Egli






Simple rasterization-based liquids
GPU PRo 5, 2013.
Martin Guay



Cloth

Screen space animation of fire
SIGGRAPH Asia, 2011.
Martin Guay, Fabrice Colin, Richard Egli





Multi-dimensional Unconditionally Stable Alternate Direction Explict (ADE) Scheme for the Convection-Diffusion Equation
Conference on Nonlinear Evolution Equations and Wave Phenomena: Computation and Theory, 2010.
Martin Guay, Fabrice Colin, Richard Egli





Video Games I worked on




A Game of Thrones: Genesis (PC)
Cyanide Studio Montreal, 2011.







Blood Bowl
Cyanide Studio Montreal, 2009.