Michelangelo is a revolutionary tool for 3D artists. It connects the power of procedural modeling with crowd modeling to produce 3D content very efficiently. With Michelangelo artists can accomplish projects of a complexity they would never dare before.

Procedural modeling

When modeling manually the artist adds geometric primitives to the scene and shapes them using geometric operations in order to achieve the desired result. The goal of procedural modeling is to create a template which describes a whole class of objects at once. Templates are like recipes capturing the sequence of operations to be applied to the initial primitives to transform them into the final 3D objects. Each template can be easily instantiated to many similar objects by varying its input parameters. Artists control the parameters either directly or by specifying valid ranges for random generators.

Crowd modeling

Similar to crowd-sourcing, collaborative effort in 3D modeling can easily beat the work of individuals. Michelangelo utilizes this idea in its core as the main distinctive feature towards any other 3D modeling system. Our vision is to connect all artists to virtual team without any additional management effort. While working with Michelangelo, each result is automatically assembled from many small contributions of different users. The system collects semantics of the created models and learns to connect them in a meaningful, context-sensitive way. Crowd modeling enables complex 3D scenes to be produced easily and very fast.


Users create grammars which contain procedural rules. Rules process semantics and geometry of objects to refine their appearance or to divide it into specific parts. The design process goes always from a high-level concept with a roughly given volume down to fine local details. Each rule either creates new geometry itself or queries rules stored in Michelangelo's cloud. Up to the whole 3D scene can be created without actual modeling, just by casting queries. The artist choses how much control gets delegated to Michelangelo and how much stays in the hands of the human. Resulting 3D models can be exported and used in movies, games, commercials and VR applications without any restrictions.


Michelangelo started as a research project at TU Wien in fall 2014. The current alpha version offers a scripting interface which is a serious barrier for artists. It will be replaced by a visual interface within the next months. Further improvements are planed to break the current limitations of geometric operations and materials settings. We are looking for developers, mainly with with computer graphics background, to extend our team.

Powered by