Machine learning is a great technology for classifying and interpreting huge amounts of data. One of the most interesting parts of this technology is the use of data that is not visible to the human eye.
There is much more in the world than we actively perceive. We are surrounded by many different invisible activities that can be measured in frequencies. Our goal was to create a general signal classifier that can be trained on all these different frequencies and interpret a generalized data stream.
As a test environment we wanted to build a water-touch-interface. The interface should be trained in real time to recognize if a finger, two fingers or the whole hand is in the water and thus control a Unity application.
Context
Elective: Creating Rich User Experiences with Applied Machine Learning
Student Project
Team
Maximilian Brandl
Philipp Kaltofen
Supervised by
Meghan Kane (@meghafon)
Buzzwords
Machine Learning, Tangible Interfaces
A sensor to generate training data for the signal classifier to turn objects into touch-sensitive interfaces. More infos about the sensor on Github.