{"id":206,"date":"2015-06-01T00:00:08","date_gmt":"2015-05-31T23:00:08","guid":{"rendered":"http:\/\/preip.net\/?p=206"},"modified":"2015-06-01T00:00:08","modified_gmt":"2015-05-31T23:00:08","slug":"youtouch","status":"publish","type":"post","link":"https:\/\/preip.net\/?p=206","title":{"rendered":"YouTouch!"},"content":{"rendered":"\n
\n
\n

While large display walls are support multiple users interacting at the same time, most system can’t recognize which touch belongs to which user. YouTouch!<\/em> tracks users interacting in front of the wall can distinguished touches of each individual user. It therefore enables applications to use personalizes touch support. To demonstrate this principle, we developed a multi user paint application where each user has their own individual color palette. It was developed at the Interactive Media Lab Dresden<\/a> and resulted in a publication <\/a>at the ACM AVI 2016<\/a>.<\/p>\n<\/div>\n\n\n\n

\n
\"\"<\/figure>\n<\/div>\n<\/div>\n\n\n\n\n\n\n\n

General<\/h2>\n\n\n\n

YouTouch!<\/em> was developed by a team of three, including me.<\/p>\n\n\n\n

The systems consists of several applications. The users are tracked using the Microsoft Kince<\/a>t<\/a>. Because the Kinect loses the Id of people whenever they leave its field of view, further processing is done by an application running on a dedicated tracking PC. It is written in C++ and identifies persons by comparing color histograms of body parts and skeleton-based biometric measurements. Short-term occlusion of users is also handled. The result is sent to another component running on the display wall, also written in C++. It uses a combination of the skeleton and image data to associate touches to specific users. The resulting personalized touches are then send to the client applications using the TUIO <\/a>protocol, which we extended to support the additional information. The client application can be written in any framework. Our multi-user paint is written in Python <\/a>using libavg<\/a>. YouTouch!<\/em> requires no user instrumentation nor custom hardware, and there is no user registration nor learning phase.<\/p>\n\n\n\n

\n