Surfnet logo

Using Machine Learning Concepts for Learning gestures

This is a completed project.

Description

Industrial Problem

This study addresses a new challenge in the design of gesture based applications for multi-touch devices: to design gestures that users consider natural, understandable and easy to use.

Research Goal

The study aims to provide designers with an easy way to create gestures that does not require any programming skills. In order to learn custom gestures, we developed IGT, a machine learning tool based on the Gesture Toolkit that uses an approximation of the anti-unification modulo theory to learn based on samples of the gesture provided by the designer.

 

Images and Videos


 

 

Publications

Learning gestures for interacting with low-fidelity prototypes. Accepted to the Workshop on Realizing Artificial Intelligence Synergies in Software Engineering (RAISE’12). To be held June 22—29, 2012, in Zurich, Switzerland

Partners

NRC Institute for Biodiagnostics

 

Workshops and Symposiums

Surfnet workshop 2011 – Tutorial: Using Machine Learning to Recognize Gestures

 

Research Focus Area(s)

 

Researchers

  • Frank Maurer (Supervisor)
  • Tulio Alcantara (MSc Student)