High Fidelity Facial Animation Capture and Retargeting with Contours

How to read this

Category

Capture system: contour-based facial performance capture and retargeting

Contributions

• Performance capture that exploits contour features (eyelids and inner mouth) to recover subtle expressions missed by marker-only approaches
• Two-step optimization combining blendshape fitting with per-frame corrective shapes to match tracked contours and markers
• Enables high-fidelity retargeting of subtle expressions to CG character rigs

Context

Builds on high-resolution facial performance capture (Bradley et al. High Resolution Passive Facial Performance Capture), adding contour cues to improve fidelity in regions markers cannot track.Builds on: High Resolution Passive Facial Performance Capture

Correctness

Demonstrated on subtle expressions around eyes and mouth that markers miss; relies on robust contour tracking and on the corrective-shape step to absorb residual error, so capture quality depends on contour-detection reliability.

Clarity

Accessible; a first pass conveys the contours-plus-markers idea, a second pass clarifies the two-step blendshape-then-correctives optimization.

How to read it

Focus on which contours are tracked and how the corrective step complements blendshape fitting; a second pass pays off if facial-capture fidelity or retargeting is your goal.