Modeling and Rendering for Realistic Facial Animation

Abstract

Abstract.Rendering realistic faces and facial expressions requires good models for the reflectance of skin and the motion of the face. We describe a system for modeling, animating, and rendering a face using measured data for geometry, motion, and reflectance, which realistically reproduces the appearance of a particular person's face and facial expressions. Because we build a complete model that includes geometry and bidirectional reflectance, the face can be rendered under any illumination and viewing conditions. Our face modeling system creates structured face models with correspondences across different faces, which provide a foundation for a variety of facial animation operations.

Recovering Facial Shape Using a Statistical Model of Surface Normal Direction

“…The first of these is the three-dimensional shape of the face [6]. The second is the albedo map which captures effects such as local variations in skin pigmentation and the distribution of facial hair [24]. Finally, there is the process by which light is reflected from the skin.…”

Techniques of Facial Synthesis: A Comprehensive Literature Review

“…Another work that started at the end of 1990s was the reflectance and illumination recovery of face images.In 1997,Marschner and Greenberg [43] developed an lighting system for face relighting.Given a camera model,the face geometry and albedo,and a set of basis lights,their system produces a set of basis images for the face under the basis lights.Given a new photograph,their system searches for a linear coefficients are essentially the lighting coefficients.In 1999,Marchner et al [45] developed an image-based system to measure the Bidirectional Reflectance Distribution Function(BRDF).They assumed that the surface area of an object is curved and all the point on the measured object have the same BRDF.In 2000,Debevec et al [44] developed a system to capture spatially varying reflectance properties of a human face's skin area.It require 2D array of light sources and a number of synchronized cameras,and the light sources and the cameras need to be calibrated.In 2005,Weyrich et al [46,47] developed a system to capture and measure a more general bidirectional surface scattering distribution function,which takes into account the translucent component of the skin reflectance.In 2007,Wang et al [48,49] proposed a spatially varying texture morphable model.They divide the image into multiple subrrgions and have a separate texture morphable model for each subregion.The spatial coherence between subregions is modeled as a Markov random field.Their technique is capable of handling harsh lighting conditions that cause cast shadows and saturations,as well as partial occlusions.…”

Mesh Parameterization Methods and Their Applications

“…For example on a database of human shapes [3], the distinguishing traits may be gender, height, and weight, while a database of human faces may add facial expressions [10][11][12]85]. Objects can be compared against the database and scored against each of these dimensions, and the database can be used to create new plausible object instances by interpolation or extrapolation of existing ones.…”