As 3D printing technology advances and becomes cheaper, homemade fabrication of digital objects becomes more and more common.
On the other hand, this approach does not fit well industrial production's requirement. Printing speed and resin's cost simply make the technique unscalable to large volumes of product.
Thus, despite being not so innovative, casting molds are still the preferred technology for assembly line production: their production is cheap, and they support a large variety of materials and ensure a virtually unlimited geometric precision.
However, designing castin molds turns out to be a large hidden cost. This process is difficult and really time consuming. Also, a good design that allows for clean casting and material saving requires the work of very specialized people, hence definitely introducing other costs.
While a variety of softwares are available that provide help in mold design, they still require human supervision and manual interaction to produce acceptable results. Instead, we want to find algorithms based on differential geometry and shape analysis that provide a fully automated pipeline for designing casting mold.
Support to rigid molds is a key step in automating the design process. Having this possibility will provide a great generalization and allows for the use of a wider range of casting materials, like plastic and metal.
Other than the great benefit that many enterprises would experience, the ability to 3D print molds would provide the possibility of casting with cheap materials even to amateurs and low budget companies.
The results achieved by Alderighi et al.[1] prove the possibility of using differential geometry and shape analysis to design silicone molds for almost arbitrary shapes. The studies presented in this work offer a solid basis to start with, and the problem formulation can be tuned and enriched to fit the more restrictive problem of rigid molds.
An higher generalization to multi-component casting can be addressed via manifold segmentation. This is a well known problem in literature, useful for many applications[2,3]. Partitioning offers a valid basis for calculating the components of a mesh. In addition, connection points and structures can be designed by taking into account shapes and dimensions of single components.
[1] T. Alderighi et al. "Metamolds: computational design of silicone molds." ACM Trans. Graph. 37 (2018): 136:1-136:13.
[2] I. Baran et al. "Automatic rigging and animation of 3d characters." ACM Transactions on graphics (TOG) 26.3 (2007): 72-es.
[3] A. Shamir "A survey on mesh segmentation techniques." Computer graphics forum. Vol. 27. No. 6. Oxford, UK: Blackwell Publishing Ltd, 2008.