This paper builds volume-appearance models based on density data from X-ray computed tomography (CT) scans and appearance data from photographs under uncontrolled illumination. The resulting models show extreme detail in closeups, while at larger scales the distinctive textures and highlights of a range of fabrics emerge automatically.

A simple, fast solution for reflectance acquisition with a BRDF chart, a linear light source, and a video camera on a mobile phone. Convincing svBRDF results are obtained after less than 30 seconds of data capture in casual environments.

This paper describes a system for measuring microscopic surface geometry using an elastomeric sensor and a single camera. The system is not affected by the optical characteristics of the surface and can be configured as a portable device. Results from a variety of real-world materials are presented.

A novel interactive method to assess cataracts in the human eye by crafting an optical solution that measures the perceptual impact of forward scattering on the foveal region. The method goes beyond traditional evaluation procedures by computing maps for opacity, attenuation, contrast, and point-spread function of cataracts.

This paper addresses the problem of transferring facial components between images with different expressions. Expression Flow is a 2D flow field that warps the target face globally so that the warped face becomes compatible with the new feature to be blended in.

With this new technique, weathering data can be extracted from photographs of real scenes and applied to generate similar weathering appearance in new environments. Stains produced by fluid flow illustrate a representative case.

A method for generating face animations from large image collections of the same person, by computing an optimized, aligned subsequence. This approach is the basis for the Picasa's Face Movies feature. A key contribution of this paper is proving why the cross dissolve produces a strong motion effect.

A data-driven method for synthesizing phenomena characterized by repetitive discrete elements, such as a stack of fresh produce, a plate of noodles, or a mosaic pattern. The method is general and easy to use, requiring only a small input exemplar and a coarse output domain.

A robust solution for integrating photon beams into PhotoRealistic RenderMan for the animated feature "Tangled".

A method to effectively importance sample the single scattering integral in homogeneous participating media from point lights and area lights of arbitrary shapes.

A method for efficient ray tracing of heterogeneous participating media using ray marching by decoupling the lighting calculations from the sampling of volume properties.

A new demand-driven volume rendering approach for extremely large electron microscopy (EM) volume data that for the first time allows neurobiologists to view terabyte-sized EM volumes interactively, without pre-computing a 3D multi-resolution representation beforehand.

Modular Radiance Transfer is a recent technique that computes approximate direct-to-indirect transfer by warping and combining transport from a small library of simple shapes. This talk discusses implementation issues that arose when scaling the algorithm from mobile devices to high-end GPUs.

An overview of a systems pipeline for high-dynamic-range video capture, reconstruction, and modeling of real-world scenes with strong spatial variations in the lighting environment, which enables photo-realistic rendering of synthetic objects placed in the scene.

A simple and efficient method for pixel-scale occlusion culling in arbitrarily complex scenes.

This solution to the ill-posed, non-uniform point-spread-function (PSF) estimation problem transforms it into a well-posed image-registration problem that estimates homographies representing the motions of camera shakes in 3D.

Movies captured at the equivalent of a trillion frames per second show the motion of light pulses through scenes, bouncing off walls and getting trapped inside fruits due to sub-surface scattering.

Display pixel caching fills empty borders that occur when videos are displayed on a screen with mismatching aspect ratios and/or resolutions with spatially and temporally consistent information while preserving the original video format. It merges the motion information from subsequent frames to generate high-resolution panoramas in a consistent manner.

A device-independent imaging system for accurately capturing and faithfully reproducing colors.

Presented by David Attenborough, "Who do you think you really are?" is an interactive film that uses cutting-edge technology, including location-based augmented reality and real-time audience participation, to take audiences on an immersive, interactive journey into their evolutionary past.

Google Body provides an intuitive 3D context for learning and discussing human anatomy on a browser, smart phone or tablet. It allows the user to peel away layers of anatomy, rotate and zoom, navigate, and search, all without a plug-in because of the WebGL standard and efficient data transmission.

This novel algorithm to interactively compute indirect illumination is based on a filtered voxel octree representation of the scene used as a proxy for lighting information and coupled with an approximate voxel cone tracing that allows a fast estimation of the visibility and incoming energy.

CANCELLED

From Dust is a new game for PlayStation 3, Xbox 360, and PC developed at Ubisoft Montpellier. It features a fully dynamic world where sand, water, lava, and vegetation are constantly evolving and ruled by a realistic simulation.

An interactive desktop GPU ray tracer for large models composed of several hundred million polygons.

This talk summarizes the components of an extremely fast fluid-simulation framework for animated special effects, including the specific numerical methods used to achieve high performance and good visual quality, and the setup and control framework that allowed artists to work efficiently at high resolution.

Exploding a super hero in DreamWorks Animation's “Megamind” was a large-scale task. This talk presents solutions to some of the technical hurdles and the collaborative work flow to that facilitated them.

How and why Animal Logic created Snap,a grid-based simulation and volumetric rendering framework used in "Legend of the Guardians: The Owls of Ga'Hoole" and "Sucker Punch".

Practical techniques for simulating fluid dynamics and occlusion in a dynamic 2D environment, utilizing PlayStation 3 SPUs to achieve 60fps at 1080p in PixelJunk Shooter 2.

This talk discusses implementation of the occlusion culling system used in Alan Wake and presents details on how the system efficiently supports shadow-caster culling for dynamic light sources in large environments.

A new solution and approach to culling complex game scenes on parallel architectures at high performance.

KIllzone 3 features complex occluded environments. To cull non-visible geometry early in the frame, the game uses PlayStation 3 SPUs to rasterize a conservative depth buffer and perform fast synchronous occlusion queries against it. This talk presents an overview of the approach and key lessons learned during its development.

This novel content creation framework, consisting of Guerrilla's own game-rendering engine embedded inside Autodesk Maya, achieves high-fidelity previews of assets while preserving flexibility, so work can continue during the preview.

"Legend of the Guardians: The Owls of Ga’Hoole" features over 60 distinctive, art-directed, hyper-realistic feathered characters. Animal Logic developed a procedural feathering pipeline, Quill, that allows a procedural representation of feathers with automatic de-intersection, animation, character effects, and dynamics, and rendered with extensive level-of-detail support.

A simulation-based technique to prevent animated feather penetrations that preserves the artist's visual development as closely as possible while providing a collision-free state for subsequent animation and simulation.

An accurate, robust, and efficient method to handle penetration between interpolated hairs and subdivision surfaces during render time. The method works by converting the subdivision surfaces to Bezier patches and performing accurate hair-patch intersections using Newton's method. The method works on every frame independently.

Kami is a geometry-instancing pipeline to generate and deform simple primitives like curves (for fur, hair, and grass) and arbitrarily complex shaded hierarchical geometry. The system has been most recently used in "The Smurfs" movie to populate a village with plants, flowers, and grass.

Tthe portrait system that drives Eve Online's new high-fidelity avatars. Breaking down the final image into its components, this talk surveys the various techniques (including some that didn't work out) for real-time skin, eyes, hair, and shadows on commodity graphics hardware.

SpeedFur is a node-based procedural hair and fur modeling system that provides a GPU-accelerated real-time preview of the hair and renders on a traditional CPU-only render node.

A language and heterogeneous compiler for efficiently solving structured grids on the CPU or the GPU, designed from the ground up with parallelism in mind.

The tools developed at Digital Domain to create unique character destruction effects that are art directable.

The Moving Picture Company introduces Kali, a finite-element-based simulation toolkit for large-scale and detailed destruction, developed in collaboration with Pixelux. The toolkit is described with examples from Zack Snyder's "Sucker Punch".

This hybrid approach to combining the power of custom hair dynamics with the artistic control of key-framed animation was key to the success of directing hair motion on "Tangled".

Creation of a highly art directed "dam breaking with resulting flood" sequence for Disney's "Tangled" by applying the 12 basic principles of animation to a fluid-based simulation through post manipulation of timing and shape.

The stereoscopic conversions of Disney's "The Lion King" and "Beauty and the Beast" required development of novel techniques for art direction, editing, and visualization of 3D stereoscopic depth without the use of geometric models.

CANCELLED

A concrete survey of stereo film-making techniques based on three recent Moving Picture Company projects.

Production of "Inazuma Eleven: the Movie" required a set of new tools to enable conversion from Japanese animation to 3D stereoscopic movie. This talk summarizes those in-house tools and how OLM used Japanese animation specificity to build up an efficient conversion system.

Processing.js 1.0, a complete port of the Processing language, brings interactive 2D and 3D graphics to the web without the use of plug-ins.