Geometry-Aware Retargeting for Two-Skinned Characters Interaction
DescriptionInteractive motion between multiple characters is widely utilized in games and movies. However, the method for transitioning interactive motions considering the character's diverse mesh shape has yet to be studied. We propose a Spatio Cooperative Transformer (SCT) to retarget the interacting motions of two characters having arbitrary mesh topologies. SCT predicts residual of root position and joint rotations considering the shape difference between the source and target of interacting characters. In addition, we introduce an anchor loss function for SCT to maintain geometric distance between interacting characters from source to target characters. We also propose a motion augmentation method with deformation-based adaptation to prepare a source-target paired dataset with identical mesh topologies. In experiments, our method shows higher accuracy for semantic preservation and lower artifacts of inter-penetration between interacting characters for the unseen character and motion than the baselines. Moreover, we conduct a user evaluation across various shapes of characters, spanning low-to-high interaction levels to show better semantic preservation of our method compared to previous works.
Event Type
Technical Papers
TimeTuesday, 3 December 20249:00am - 12:00pm JST
LocationHall C, C Block, Level 4
