Output details

Automatic rigging for animation characters with 3D silhouette

Originality: Animating an articulated 3D character requires the specification of its interior skeleton structure which defines how the skin surface is deformed during animation. Currently this task is accomplished manually, which consumes a significant amount of the animators' time. This work presents animators with a fast, automatic, easily controlled rigging method. Given an animated character model, our method can automatically supply the animator with a fully structured skeleton in a very short time. This paper also presents a new geometric concept – the 3D silhouette - which extends the traditional 2D skeletonization techniques into 3D space. Our method is faster than existing methods for virtual character skeletonization.

Rigor: A prototype system was developed to test the idea in animation production. Extensive experiments on different virtual characters have proven the robustness and efficiency of our technique. The analysis of algorithm complexity proved in theory that our method is faster than existing methods. A comparison with the state-of-the-art techniques also shows that skeletons extracted by our method are more suitable for character animation.

Significance: This paper presents an efficient method for character rigging in animation production. It can save a large amount of animators’ time. This work also laid the foundation and provided an accurate skeleton structure for our later work on “automatic muscle modelling”. This is a key output of our three-year project “Retargeting Animated Human Characters” funded by Great Western Research (GWR) (£117,000, Zhang, Yang). In addition, this work has led to the successful funding from BIS on another project, “Intuitive Motion Data Retrieval and Synthesis” (£36,000, Zhang, Yang), whose aim is to provide the animator with an intuitive tool to reuse existing motion data.