Feature Driven Rule Based Framework for Automatic Modeling of Organic Shapes in the Design of Personalized Medical Prosthetics

Abstract

We propose a novel framework for the personalized design of organic shapes that are constrained to exhibit conformity with the underlying anatomy. Such constrained design is significant for several applications such as the design of implants and prosthetics, which need to be adapted to the anatomy of a patient. In such applications, vaguely defined work instructions are usually employed by expert designers to carry out a sequence of surface modification operations using interactive CAD tools. Our approach involves the abstraction of the work instructions and the expert knowledge into feature dependent machine interpretable rules in a Knowledge Base. Robustly identified canonical set of anatomical features are then employed to determine concrete surface sha** operations by a Smart Shape Modeler. These operations are eventually performed sequentially to adapt a surface to a target shape. The versatility of our approach lies in a priori defining an entire design workflow through a scripting language, thereby yielding a high degree of automation that is completely flexible and customizable via scriptable rules. Consequently, it eliminates tedious manual intervention and offers desirable precision and reproducibility. We validate this framework with a practical application – automatic modeling of shells in hearing aid (HA) manufacturing (HAM).