Megan Hofman
Carnegie Mellon University
meganh@cs.cmu.edu
Bio
Megan Hofmann is a PhD Candidate at the Human Computer Interaction Institute at Carnegie Mellon University. Her research focuses on the intersection of digital fabrication disability and healthcare. Her research has explored 3D printed prosthetics and assistive technology and developed novel 3D modeling frameworks that support 3D printed model reusability. Additionally she develops design tools for Industrial Knitting machines rooted in the traditions of hand-knitters. Her current research is exploring the design of Clinical CAD tools that will enable healthcare professionals to design and print a wide variety of medical devices using generative design. Her research is currently put into action to support healthcare professionals in the COVID-19 crisis. Much of this research has been published at top HCI conferences such as CHI UIST ASSETS and CSCW. Megan has been awarded multiple fellowships and best paper awards for her work.
A Framework for Producing Clinical CAD Tools with Generative Design
A Framework for Producing Clinical CAD Tools with Generative Design
In recent years there has been a surge in digital fabrication research. Specialized design tools enable us to create new things with 3D printers laser cutters and even automatic knitting machines. We are only limited by what designers can model. However modeling complex designs from scratch is tedious and requires extensive skill from end-users. In many cases designers would prefer to provide a set of requirements to a tool and have it generate a matching design automatically. Based on six studies of disabled makers and medical practitioners we find that this preference is particularly salient when designing customized medical devices. Unfortunately domain specific generative design tools are difficult to build and require programmers to have extensive knowledge of the domain and optimization methods a rare combination. Despite extensive work on generative design tools for digital fabrication and the optimization methods that enable them there are no frameworks that enable the rapid development of generative design tools in new domains. In Megan Hofmann’s thesis work she summarizes two systems that characterize an alternative domain specific approach to building design tools. The first system PARTs (Parameterized Abstractions of Reusable Things) helps expert designers to craft reusable models and share them with novice modelers. The second system OPTIMUM (Optimization Programming Toolkit Integrating Metaheuristic User-driven Methods) enables domain experts and programmers to collaboratively implement domain specific generative design tools. She has demonstrated OPTIMUM by replicating prior work to generate “fabricable tile-decors” and by developing two novel tools: 1) KnitGIST a generative design tool for machine knitting; and 2) Maptimizer a tool for generating customized tactile maps for blind users.