follow us

Follow AITathens on Twitter faceebook in_logo

Happening now...

Human-Computer Interaction

Research @ AGC Group of AIT

Gesture and nod recognition

Gestures and nods are detected using motion: Information from both motion vectors in macroblocs with significant texture information and from blocks with significant frame-by-frame difference is combined to discriminate nods from gestures and understand their direction.

Green arrows correspond to motion vectors matched to the median motion (thick green arrow with yellow tip), that are used to recognise the gesture or nod. Red arrows do not match the median motion and are discarded. Blue circles mark blocks with significant frame-by-frame difference.

This is reseach conducted for the AGNES project. Contact Aristodemos Pnevmatikakis,

Back to top

Hand tracking for mixed reality applications

In A realtime mixed reality system for seamless interaction between real and virtual objects and the thesis Computer Vision Based Human Machine Interfaces by Achilleas Angagnostopoulos, we present a system for integrating actual objects into a virtual world. The system segments and tracks objects, especially hands, using an overhead camera stereo pair. The obtained 3D tracks are used in a mixed reality game, where digital and actual objects, as well as multiple hands interact. The process is as follows:

A realtime mixed reality system for seamless interaction between real and virtual objects

Contact Aristodemos Pnevmatikakis,

Back to top

Multi-touch surface

In A Low-Cost Multi-Touch Surface Device supporting Effective Ergonomic Cognitive Training for the Elderly we present a system that captures human input employing a multi-touch device.

The hardware setup for a multi-touch surface is as follows: an acrylic panel is placed on top of the TFT panel, the edges of which are illuminated by IR-Light Emitting Diodes. Due to the Frustrated Total Internal Reflection effect, a finger touch on the surface of the acrylic panel generates lighting blobs. This is due to the fact that the refractive index of skin is higher than that of air. The position of blobs that manage to penetrate the TFT panel is captured by a near-infrared (NIR) camera through an ultraviolet and visual cut optical filter.

Kalman filters are used to track the blobs corresponding to the fingertips. The multi-touch hardawe serves to attenuate background clutter, giving very clear measurement contacts. We have been applying the multi-touch device in games for aleviating the cognitive decline of the elderly in the HERMES project. The system is illustrated below:

A multi-touch surface for interaction between the elderly and computers

The infrared traces of fingers on a multi-touch surface are tracked using Kalman filters. Their state vectors comprise position, velocity and trace thickness. The following demo application draws the tracks as curves with thickness proportional to that of the trace and a random colour.

This is reseach conducted for the HERMES project. Contact Thodoris Petsatodis, (multi-touch hardware and game design) or Aristodemos Pnevmatikakis, (finger tracker).

Back to top

Pointing gestures

The rod is tracked in 3D using a pair of cameras. The direction of the rod is projected towards the wall, as shown in another pair of cameras.

This is the 2007 outcome of our students' project for the graduate course Ambinet Inelligence Laboratory. Contact Aristodemos Pnevmatikakis,

Back to top

Bookmark and Share
Affiliated with Aalborg University-CTiF, Harvard-Kennedy School Of Goverment © ATHENS INFORMATION TECHNOLOGY designed by {Linakis+Associates}