Natural user interfaces for engineering sketch understanding

Sundar Murugappan, Purdue University

Abstract

Over the last decade, post-WIMP interfaces started to gain acceptance among the Engineering Design community. The increased interest in freehand sketching paved the way for creating `sketch-based interfaces' with digital pen as the main tool for interaction. The first part of this thesis focused on the development of robust sketch understanding techniques that would enable pen-based interfaces to be the preferred computing platform of choice. The goal of providing greater and better affordances for natural human-computer interactions (HCI) has largely motivated computational research and technological advancements. As a result, developments of novel algorithms for better virtual interfaces and innovations in hardware technology have become mutual partners, driving each other towards making HCI more intuitive and accessible. Recent innovations like iPad™ and Kinect ™ have created a paradigm shift in how people view and interact with computing devices especially using their fingers and whole body. The second part of this thesis focuses on turning any static physical surface such as a wall or a table into a multi touch enabled interactive surface using the depth sensing camera without instrumenting the surface itself. This thesis also focuses on developing algorithms in the area of combined pen and multi touch interaction.

Degree

Ph.D.

Advisors

Ramani, Purdue University.

Subject Area

Computer Engineering|Mechanical engineering|Computer science

COinS