MIT Labs Develops Kinect-Based Teleconference Interface

MIT wunderkids demo awesome telco interface using SAR and Kinect.

Published
Updated
1 minute & 30 seconds read time
MIT Labs Develops Kinect-Based Teleconference Interface | TweakTown.com


Lining Yao, Anthony DeVincenzi, Ramesh Raskar, and Hiroshi Ishii from the Massachusetts Institute of Technology's Media Lab have developed a prototypal teleconference interface demonstration using Microsoft's Kinect sensor array.

Featured in their demo video, Lining (Lizzie) and Anthony (Tony) show off some of the features they managed to successfully implement. In their interactive interface, speakers will be given time-bubbles that pop up over their respective heads, tracking the length of time that each is speaking. Amazingly, the clock stops as soon as the person stops speaking, meaning the program is recognizing voices individually. A cool feature (that may need a bit of tweaking, but cool nonetheless) they also included is the automation of focus- when a person is speaking, the focus of the camera changes, ensuring that everything but the speaker is blurry. This wasn't as successfully implemented as the time-tracker, but a very interesting idea, especially for someone like me with rampant ADD.

Speaking of ADD, I have many problems when it comes to not answering or responding to portable phone vibrations, so this next feature made me sigh in relief that someone was actually working on it. The MIT team developed a way that a person in a teleconference can actually freeze an image of themselves, for instance sitting at a table with a rapt expression, while the person (unbeknownst to the people on the other side of the video chat) can check their phones, scratch their armpit or play Angry Birds (picture above).

The team also managed to integrate Spatial Augmented Reality (SAR) into their demo, illustrating the uses an interface like that could have for interaction with objects on the other side of the screen. Want to display that new architectural CAD model your team just finished? Instead of sending everyone a bulky file they all have to download, why not just display it from an image marker on a block of wood? Also, why not just give each user the ability to manipulate and select marker objects at will? Why not, indeed, asks MIT Media Labs.

We're really looking forward to future progress on this- the Kinect is a wonderful development tool and projects this are the best part. Keep up the good work guys! Video below:


Newsletter Subscription

Related Tags