Volker Blanz
Baback Moghaddam
Hanspeter Pfister
Dimitris Samaras
William Smith

Dates (all 2008)

Submission: March 15th
Extended submission:
March 22nd
(After ECCV Deadline)

Notification: April 15th
Camera ready: May 1st
Workshop: June 27th


Accepted papers will appear in CVPR proceedings and on IEEE Xplore

Authors will also be invited to submit to a special issue of the journal IET Computer Vision

A workshop of CVPR 2008

Programme Announced

The final workshop programme is now available. We are delighted to announce that there will be two invited talks from world renowned speakers: Demetri Terzopoulos (University of California) and Parag Havaldar (Sony Imageworks).

Call For Papers

Estimating 3D face shape from one or more images is a longstanding goal of computer vision. In the earliest work on shape-from-shading, researchers applied their algorithms to face images with little success. Advances during the last decade have seen the development of techniques that offer robust performance on real world images. Meanwhile, advances in structured light scanning have made high-end acquisition of 3D structure and motion a reality, albeit in very controlled settings, thus making statistical techniques attractive. A clear result to come from this work is that processing 3D face shape in images requires techniques that span a number of fields. These include statistical shape modelling, non-linear optimisation, reflectance modelling, illumination estimation and shape-from-shading.

These advances hold out the hope of estimating intrinsic properties of a face from single images or video streams. This is clearly attractive in the domain of face recognition where modelling appearance variation caused by large changes in pose, illumination and expression remains a key problem. Applications also lie in model acquisition for graphics applications, retouching faces in images (for example adjusting expressions or illumination conditions) or even exchanging faces between images.

There is also a strong link between this work and one of the key questions in psychological studies of human face processing, that of the role played (if any) by 3D shape information. This has led to an exchange of ideas between machine vision and psychology/neuropsychology in this area which is of mutual benefit.

Topics of interest include, but are not limited to, the following:

  • 3D morphable face models
  • 2D+3D active appearance models
  • Facial shape-from-shading and photometric stereo
  • Stereo for face images
  • Face/skin reflectance modelling
  • Psychological or neuropsychological investigations into the role 3D information plays in face processing in humans
  • Modelling variation in appearance due to 3D shape using spherical harmonics, light fields etc
  • Dynamic 3D face processing in video images, e.g. tracking, modelling of expressions in 3D, use of motion capture data
  • Real-time 3D face scanning from video
  • Colour information for 3D face processing
  • Structured light/Shape-from-X for face shape recovery
  • Estimation of illumination or shadowing from images
  • Data management for large 3D face data sets
  • Matching of partial or deformed scans
  • Fusion of multimodal face information, e.g. 3D scans, high-speed video, high-resolution imaging

Applications of interest include:

  • Facial shape estimation
  • Recognition/classification using 3D information estimates from images
  • Facial retouching, expression/texture transfer, relighting using 3D models
  • Medical applications of 3D face modelling and facial expression analysis

Submission Policy

Papers must describe high-quality, original research. By submitting a manuscript to this workshop, authors assert that no paper substantially similar in content has been submitted to another conference or workshop during the review period.

For all enquiries contact: