Tell Me the Truth

We are thrilled to have received the generous support of Seeing Machines
for the use of the licensed version of FaceAPI  for the project.
The team would like to extend a huge thank you!
This will allow us to move forward on the project.

Tell me the Truth is an interactive artwork that consists of an intelligent mirror interface detecting presence. When approached, the mirror transforms the viewer’s reflection into that of Andy Warhol.

Hanging on a wall, the oval-shaped mirror, about 20 inches high, is cast over an LCD display and is large enough to reflect a viewer’s face and the upper body.

When a person approaches the mirror a video tracking system is activated, gradually turns the mirror transparent and displays the avatar.

The viewer’s face morphs in real time, replaced by a video image of the iconic artist, Andy Warhol, who reflects and mimics the viewer’s movements. When the gestures, faces or behavior of the viewer are out of range with the Andy Warhol character then the iconic star takes over and plays his own persona, staring straight ahead, occasionally reciting a short, typically Warholian sentences like: “That’s great”, “Wow, that’s amazing”, “I can make you a superstar”, ”Who is this?”, “You are fantastic”, and “Take a look at that, I’ve never seen that before!”. The mimicking resumes quickly, catching up with the viewer’s behavior.

Key words: Tangible Computing, Computer Vision, Facial tracking, Optics simulation,

Reflections of Simon’s face using the linear gradient mirror, which will let us determine the degree of transparency that a partial mirror will need. This will be quantitatively measured, and specified for a partially silvered mirror to be used for our prototype.

Demo of a Maya head model with blend shapes for facial animation being animated in Panda 3D while FaceAPI non-commercial version drives the xyz position and rotation of the head via OSC

Tell Me the Truth prototype articulating a Maya 3D face model running in Panda3D. FaceAPI communicates to Panda3D via OSC.

Demo

I finally got the non-commercial version of FaceAPI working inside OF.
Right now, the API runs inside an independent thread.  I’m working in windows on of my mac, and I hope that in a real windows machine the application started faster.

I’m using the OF version for visual studio,  to make it work with the 2008 version it’s necessary to remove GLaux.lib from the linked libraries. You can find more information and the link to a 2008 friendly version in the OpenFrameworks forum

Inside of the tools folder of the API there is a Camera Calibration Tool, CamCal.
The lens calibration files can be used for accurate tracking in world-coordinates.

Seeing machines specifications are  Intel Core-2 Duo, 2.4GHz, 4MB Cache, Logitech Quickcam Deluxe for Notebooks

Face detection setup options:

- IR setup: tracking region position and size correspondences detected frontal face patterns

- Stereoscopic Setup: depth estimation,  segmentation of the user from other people and background objects
optical axis as used by the display and send estimate of the 3-D head position of the user to the application.

- Zcamera: depth estimation,  segmentation of the user from other people and background objects

Camera position:

placing camera so it shares the same optical axis as the video display.
-half silvered mirror? 45 degree half mirror so the user can view the screen also looking straight into the camera.

Video display

- Size of the screen and model of the screen.
- lenticular magnifier and a smaller flat panel display just an inch or so behind.
- Other materials

Content of the display

- 3D deformable face model
-  3D model from a 2D picture
- Use prerecorded video

Additional light control

to do:

Short video of ideal interaction   to work on the first stage of the tracking system