TopoCap enables video-to-animation retargeting across diverse skeletal structures

Researchers have developed TopoCap, a novel framework for extracting motion from monocular video and applying it to characters with diverse skeletal structures. The system learns a universal motion manifold that disentangles motion dynamics from topology, enabling retargeting onto arbitrary rigs without test-time optimization. This is achieved through a two-stage generative pipeline that compresses kinematic chains into a shared latent code and predicts these codes from visual features. TopoCap was trained on Mobjaverse, a large-scale dataset featuring over 5,000 unique skeletal topologies, significantly outperforming specialist models and enabling zero-shot retargeting for a wide range of 3D creatures. AI

IMPACT Enables animation of diverse 3D characters from monocular video, potentially streamlining content creation pipelines.

RANK_REASON The cluster contains an academic paper detailing a new method for motion retargeting. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.CV →

• Mobjaverse
• TopoCap

• paper
• other

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →