← ArchivePaper2010
The Digital Emily Project: Achieving a Photorealistic Digital Actor
Oleg Alexander, Mike Rogers, William Lambeth, Jen-Yuan Chiang, Wan-Chun Ma, Chuan-Chang Wang, Paul Debevec
USC ICT and Image Metrics cross the uncanny valley with a fully digital photoreal face built from Light Stage scans.
Abstract
The Digital Emily project, a collaboration between Image Metrics and USC Institute for Creative Technologies, built a photorealistic animated digital face from high-resolution scans of actress Emily O'Brien. Her facial shape and reflectance were digitized in 33 expressions using the Light Stage capture system with polarized spherical-gradient illumination, recording diffuse and specular reflectance and surface normals accurate to the level of skin pores and fine wrinkles. The captured expressions were turned into a blendshape facial rig with displacement maps, eyes, and teeth, and animated using a video-based facial-animation system driven by an actor's recorded performance. Rendered with subsurface scattering and image-based lighting, the result aimed to cross the uncanny valley and was widely accepted by viewers as video of a real person.
How to read this
- Category
- Capture-to-rig pipeline: photoreal digital actor
- Contributions
- A photorealistic animated digital face built from high-resolution Light Stage scans of an actress in 33 expressions
- Capturing diffuse and specular reflectance plus surface normals down to skin pores and fine wrinkles via polarized spherical-gradient illumination
- Turning the scans into a blendshape rig with displacement maps, eyes and teeth, driven by a video-based facial-animation system and rendered with subsurface scattering and image-based lighting
- Context
- An end-to-end integration building on Blanz and Vetter's morphable face model and Debevec et al.'s reflectance-field acquisition, combining them with blendshape rigging and performance-driven animation.Builds on: A Morphable Model for the Synthesis of 3D Faces · Acquiring the Reflectance Field of a Human Face
- Correctness
- A collaboration-driven demonstration aimed at crossing the uncanny valley; the stated outcome is that viewers accepted the result as real video, which is a perceptual/audience claim rather than a controlled quantitative evaluation, and it rests on the specialized Light Stage capture rig.
- Clarity
- Very accessible as a system tech note; a single pass conveys the pipeline end to end, with deeper passes only if you chase the cited capture and morphable-model methods.
- How to read it
- Read once as a pipeline overview connecting capture, rigging and rendering; follow the Debevec reflectance-field and Blanz morphable-model citations if you need the underlying capture and modeling theory.
Builds on
Built upon by
- High-Quality Single-Shot Capture of Facial Geometry 2010
- Creating Realistic CG Humans: Skin, Facial Animation and Muscle Behaviour 2016
- Senua, Siren and Meetmike: Real-Time Character Performance and Mocap Solving 2016
- Emotion Challenge: Building a New Photoreal Facial Performance Pipeline for Games 2017
- MEETMIKE: Real-Time Digital Human Avatar Creation 2018
- Past, Present and Future of Digital Humans at ILM 2018
- Replicating Rachael for Blade Runner 2049 2018
- Digital Humans: Inside Epic's MetaHuman Creator 2021
- Rigging the MetaHumans | GDC 2021 2021
- Advances for Digital Humans in VFX Production at Goodbye Kansas Studios 2022
Related work
- High Fidelity Facial Animation Capture and Retargeting with Contours 2013 / SCA
- Acquiring the Reflectance Field of a Human Face 2000 / SIGGRAPH
- Facial Performance Synthesis using Deformation-Driven Polynomial Displacement Maps 2008 / SIGGRAPH Asia
- Democratizing the Creation of Animatable Facial Avatars 2024 / arXiv
Keywords
This page summarises the entry and links to its original source. The archive never hosts or redistributes the publication itself.Show it in the full archive list →