ARTIST: Adaptive Humanoid Rigging by Transferring Individual Style with Optimal Transport

How to read this

Category

Eurographics short paper on auto rigging, learning an individual rigger's style and transferring it to new characters

Contributions

• Proposes style adaptive auto rigging that learns an artist's rigging conventions from a single example rig rather than a large training set
• Formalizes the control rig as a pose representation so rigging style can be compared and transferred using optimal transport
• Recovers usable rig controllers via rig inversion once the style has been transferred to a new character

Context

This sits in the auto rigging and rig retargeting line of research, the goal is closing the gap between fully automatic skeleton predictors, which tend to produce generic rigs, and the specific naming, hierarchy and control conventions a studio actually wants. It has no builds_on entry in the archive, the title and note suggest it argues against one size fits all learned rigging in favor of matching a given artist's setup.

Correctness

This is a short paper on the Eurographics short papers track. No PDF was found locally and web search did not surface an accessible abstract, only the authors, venue and dblp listing. Treat the contributions and context above as a plausible reading of the title and note, not confirmed against the paper's actual method or evaluation.

Clarity

Cannot be assessed directly without the text. Short papers in this track are typically compact, four to six pages, mixing a rigging formalism with a from example learning method, so expect it to be moderately technical.

How to read it

First pass, once you have the PDF, read the abstract and any figure showing an input rig plus a generated rig side by side, that alone will tell you whether the style transfer is convincing. Second pass, read the pose representation and optimal transport formulation to see how rigging style gets compared mathematically. Third pass is worth it only for someone building an in house auto rigger who wants to preserve a specific team's rig conventions, in which case track down the full short paper on the Eurographics digital library first.