ACSProject reference: 6413
Funded under :
Algorithms for Complex Shapes with certified topology and numerics
Increasingly demanding applications of geometric computing, for example in computer aided surgery, CAD/CAM, realistic virtual environments, robotics, and molecular modelling in drug design and structural biology, require efficient and robust methods for computing with complex shapes. This project aims at advancing the state of the art in this field. Current technology can cope well with curves in the plane and smooth surfaces in three-dimensional space. We want to deal with a larger class of shapes, including piecewise smooth and singular surfaces. Topics that we address are shape approximation (including meshing and simplification), shape learning (including reconstruction and feature extraction), as well as robust modeling (including boolean operations). Our work on these topics will be closely intertwined with basic research on shape representations, certified geometric calculus and algorithms producing output with guaranteed topology. A distinctive feature is the design and implementation of novel algorithmic solutions with certified topology and numerics as an alternative for heuristics and ad hoc methods, and to develop an experimental geometry kernel for modeling and computing with complex shapes as a proof-of-concept justifying our approach. The results of this project should be directly useful to the application areas mentioned above. We intend to disseminate our work by publication in the appropriate applied research forums, by organizing multidisciplinary workshops aimed at exchange of knowledge and discussion of our work. Moreover, we aim at transferring our new technology by producing high quality software, demonstrating the feasibility of our techniques in practice. Cooperation with our industrial partner includes the assessment, trial, validation and packaging of the software developed in the project, thus guaranteeing a smooth transfer of new technology to application areas.