Normalized Euclidean Distance Matrices for Human Motion Retargeting

How to read this

Category

Method: a motion retargeting algorithm

Contributions

• A frame-based posture representation using normalized Euclidean Distance Matrices of all inter-joint distances
• Normalization and denormalization procedures based on kinematic-chain lengths to adapt distance matrices across different skeletal morphologies
• A Distance Geometry Problem solver with spectral gradient optimization to recover retargeted joint positions

Context

Belongs to the motion retargeting lineage opened by Gleicher's Retargeting Motion to New Characters, recasting the cross-morphology transfer as a distance-geometry problem.Builds on: Retargeting Motion to New Characters

Correctness

The approach is posture (frame) based on inter-joint distances; readers should keep in mind that a purely per-frame distance formulation may not by itself guarantee temporal smoothness or enforce constraints like foot contacts unless handled separately.

Clarity

Moderately technical; a first pass conveys the distance-matrix idea, a second pass is needed for the normalization scheme and the DGP solver.

How to read it

Focus on how postures become distance matrices and how chain-length normalization bridges morphologies; a second pass on the DGP solver pays off only if you implement or extend it.