Minimal-Cut Model Composition T. Hassner 1 L. Zelnik-Manor 2 G. Leifman 3 R. Basri 1 1. Dept. of Computer Science & Applied Math 2. Faculty of Electrical Engineering 3. Faculty of Electrical Engineering The Weizmann Institute of Science California Institute of Technology Technion - Israel Institute of Technology Rehovot, 76100 Israel Pasadena CA, 91125 USA Haifa 32000, Israel Abstract Constructing new, complex models is often done by re- using parts of existing models, typically by applying a se- quence of segmentation, alignment and composition opera- tions. Segmentation, either manual or automatic, is rarely adequate for this task, since it is applied to each model in- dependently, leaving it to the user to trim the models and determine where to connect them. In this paper we propose a new composition tool. Our tool obtains as input two mod- els, aligned either manually or automatically, and a small set of constraints indicating which portions of the two mod- els should be preserved in the final output. It then automat- ically negotiates the best location to connect the models, trimming and stitching them as required to produce a seam- less result. We offer a method based on the graph theoretic minimal cut as a means of implementing this new tool. We describe a system intended for both expert and novice users, allowing easy and flexible control over the composition re- sult. In addition, we show our method to be well suited for a variety of model processing applications such as model repair, hole filling, and piecewise rigid deformations. 1. Introduction Computer graphics systems are becoming ubiquitous, posing a growing demand for both realistic and fictitious 3D models. Constructing new models from scratch is a tedious process, requiring either a careful scan of real objects or the artistic skills of trained graphics experts. This process can potentially be enhanced, as more and more models become available, by reusing parts of existing models. With current methods, however, the process of composing new models from existing ones is still laborious, requiring a user to man- ually segment the input models, align them, and determine where to connect their parts. Automatic segmentation tools exist, but they are largely inadequate for this task. Segmen- tation tools are applied to each of the input models inde- pendently, and so they often produce results that require the user to further trim the parts to eliminate undesired protru- Figure 1. Centaur. An output of our model com- position framework. sions or to significantly extend the parts so that they can be connected properly. In this paper we present a new, model composition tool. This tool, which is intended for use by both novice and expert modelers, automates much of the manual labor often associated with creating complex mod- els. We further show the new operation to be particularly suitable for an assortment of model processing applications. To illustrate our motivation consider for example the centaur in Fig. 1. Using our system, creating such a model from those of a horse and a man (Fig. 2) is as easy as (1) stating how the two models are to be positioned with re- spect to each other and (2) indicating to the system that the man’s head and the horse’s legs must be included in the fi- nal composition. The computer then automatically negoti- ates the best location for cutting and stitching the two mod- els, in order to produce a seamless result. With existing methods, creating such a model is rarely that easy. Cur- Proceedings of the International Conference on Shape Modeling and Applications (SMI’05) 0-7695-2379-X/05 $20.00 © 2005 IEEE