Projective Motion Correction with Contact Optimization.

Area of Science:

• Computer Graphics
• Animation
• Robotics

Background:

• Motion capture data often contains geometric and physical errors when applied to virtual characters.
• Manual refinement of these errors is time-consuming and complex.

Purpose of the Study:

• To present a novel framework for efficiently obtaining corrected motion and contact information from multi-contact motion capture data.
• To improve the naturalness and accuracy of virtual character animations.

Main Methods:

• Developed a projective dynamics-based method for optimizing character motions using differential representations.
• Created a contact optimizer that iteratively refines contact information based on feedback from the motion optimizer.
• Integrated motion and contact optimization to ensure adherence to dynamic and geometric constraints.

Main Results:

• The proposed framework efficiently corrects geometric and physical errors in motion capture data.
• Significantly improved the naturalness of virtual character motions across various activities like walking and rolling.
• Provided accurate supporting contact information alongside the corrected motion.

Conclusions:

• The novel framework offers an efficient solution for refining motion capture data for virtual characters.
• The integrated approach of motion and contact optimization enhances animation realism and accuracy.
• This method reduces the need for cumbersome manual adjustments in character animation workflows.