Spencer Barton

28 Jan 2014

Exploring ofx

Capacitive touch sensing: ofxTouche

ofxTouche made at resonate.io 2013 by Brett Renfer, Bjorn Staal, Thijs Bernolet

Touché caught my attention. This OF plugin provides functionality for the Disney Research capacitive touch sensing project.

I have messed around with a plat based sensor (Disney’s Botanicus Interactivus) and would like to explore this concept with more objects. One thought is to attach to device to a person and detect what it touching the person/ where the person is being touched (wikipedia says the human body can be charged to 3kV!). This may work well with dance as different parts of the body could come into contact with ground instead of simply hands and feet.

Capacitive sensing has the potential to add interactivity to a number of inanimate objects. I would curious about measuring frequency responses off of trees and then mapping this to sounds. Threes in the wind have subtle motions that may be detectable. Buildings may also be interesting. In this case detection would focus on material changes throughout the seasons.

Machine Learning: ofxlearn

ofxlearn by genekogan

Machine learning would be a necessary component of the capactivie touch sensing so it would be nice to have alternative machine learning options. This library is built on dlib which contains a number of tools.

I am currently rather excited by the potentials of machine learning. I like the idea that I could spend a little time training a computer and then it can solve stuff on its on (though its not necessarily that easy). This library will hopefully make the process simpler.

ML can be applied in a few projects that I have in mind. On the practical side I would like to implement smart power consumption monitor which has been trained to recognize the various devices drawing power.

I have also been thinking that it would be great to use Amazon Turk to create training data. It might be fun to have Turk identify awkward smiles.

Identifying faces ofxFaceTracker:

ofxFaceTracker by Kyle McDonald

If I am to identify awkward smiles then this face recognition plug-in will be necessary. This module is included because there have been some cool projects in OF using this software.

It would be interesting to map human face features to various animals. An avatar would be controlled by facial gestures. It may also be interesting to create a classification system for face features where images or words could be mapped to responses or speech to the face that produces it.

On a practical note: ofx

ofx by prusnak

This is a plug-in manager so I will install it first. This will likely save many headaches.