Modeling 3-D objects that are tightly constrained by real objects is a complex task. The usual process consists of iterations between the design stage that involve rigorous and tedious measurements, an optional simulation stage, followed by prototype testing which may lead to further iterations. By modeling in context (i.e., while using the real world to constrain the 3-D modeling process itself ), we can improve this iterative process by providing direct feedback to the user while modeling. The system could automatically suggest meaningful measurements, adjust part parameters according to the real object or even ensure that augmented parts satisfy fabrication, stability and other types of constraints that usually lead to multiple iterations when not met

I became interested in cameras, computer science, and the algorithms between them during a UROP in the Tangible Media group in the Media Lab while working on depth sensing and simulations. I continued exploring this topic in internships and class material. I’m excited for this project because I’ll have the opportunity to dive deeper into certain computer vision topics such as augmented reality, modeling, and scanning.