*                                                                                                                                                *
*                        For engineers that moonlight as artists and artists that moonlight as engineers                                         *
*                                                                                                                                                *
*                                                                                                                                                *
*   CCC    OOO    L      L     IIIII    SSS     IIIII    OOO    N    N                   l     l                      t      i                   *
*  C   C  O   O   L      L       I     S   S      I     O   O   NN   N                  ll    ll                      t                          *
*  C      O   O   L      L       I     S          I     O   O   N N  N    ccc     ooo    l     l     eee     ccc    ttttt    i    v   v    eee   *
*  C      O   O   L      L       I       S        I     O   O   N  N N   c   c   o   o   l     l    e   e   c   c     t      i    v   v   e   e  *
*  C      O   O   L      L       I         S      I     O   O   N   NN   c       o   o   l     l    eeeee   c         t      i    v   v   eeeee  *
*  C   C  O   O   L      L       I     S   S      I     O   O   N    N   c   c   o   o   l     l    e       c   c     t      i     v v    e      *
*   CCC    OOO    LLLLL  LLLLL IIIII    SSS     IIIII    OOO    N    N    ccc     ooo   lll   lll    eee     ccc      tt    iii     v      eee   *
*                                                                                                                                                *

Social Turkers: Crowdsourced Relationships

computer, website archive and video
Artwork by collisi1

What if we could receive real-time feedback on our social interactions? I developed a system like this for myself using Amazon Mechanical Turk to explore in the form of a performance. During a month of continuous dates with new people I met on OkCupid, I streamed the interaction to the web using an iPhone app. Turk workers were paid to watch the stream, interpret what was happening, and offer feedback as to what I should do or say next. These directions were communicated to me via text message.

The Verge – OK, Cupid: giving your love life to Google Glass and the hive mind
Fast Company – Lauren McCarthy let the crowd control her date, and soon you can too

Dial-A-Style: An Algorithmic Portrait Studio

wood, computer with custom software, webcam, video projector
Dial-A-Style: An Algorithmic Portrait Studio
Dial-A-Style: An Algorithmic Portrait Studio
Dial-A-Style: An Algorithmic Portrait Studio
Dial-A-Style: An Algorithmic Portrait Studio
Dial-A-Style: An Algorithmic Portrait Studio

Dial-A-Style - An Algorithmic Portrait Studio is an interactive video installation that allows visitors to create digital self-portraits in a variety of painterly styles. Through experimentation, viewers will come to understand and appreciate how various styles of painting impact the emotional connection to the artwork.

For online images created by this system, please see


  1. Click the Start Button
  2. Position the Webcam
  3. Spin the Wheel
  4. Click the Upload Button

There are four major styles represented on the wheel:

  • Impressionism - in the style of Van Gogh
  • Cubism - in the style of Picasso and Braque
  • Pointillism - in the style of Chuck Close
  • Anime - the Japanese comic style

The wheel can stop in between neighboring styles, which results in a hybrid styled portrait.

Open Source Components:

  • openFrameworks
  • OpenCV library
  • XDoG: eXtended difference-of-Gaussians flandmark face detector

I would like to thank Jennifer Lim for her help with this installation.


web, computer, urethane
Artwork by [user-name]
polyurethane hippo

Minimals are low resolution animals constructed out of digital material building blocks which form strong structures and can be readily assembled with a low cost robot. The concept of this work is to scientifically determine the number of parts where a shape turns from abstract to concrete. The number is determined to be the number of parts that yields 50% object recognition. We determine this number using an online survey that asks users to identify the given animal rendered with a certain number of parts. We employ a CAM system for constructing a model derived from a given shape mesh and php scripts for running the automated online survey.

Thanks to Sarah Hovsepian for help building parts and MIT CBA group for their support.

LumaTouch Synesthesia

wood, Plexiglas, fluorescent lights, computer with custom software, webcam, video projector, headphones, speakers
Artwork by [user-name]

LumaTouch Synesthesia is an interactive system for creating abstract artwork with electronic music by manipulating tangible objects. The system consists of a light table with five movable objects on the surface, four small cubes and a small cylinder. By manipulating these objects, the user can simultaneously create an abstract painting and compose electronic music to complement the painting. The phenomenon of synesthesia is experienced literally in the creation of color and sound.
The location and rotation of the four cubes are detected by a webcam in the light table. The custom software uses these inputs to choose up to eight of 16 reference images to be morphed and combined for the painting and up to eight of 16 music loops to be mixed for the song. The cylinder is used to change the color of the images and the musical key of the song.
While a viewer is interacting with the piece, it will show the digital painting projected on the wall behind the light table, and play music at a moderate volume level in the room. Headphones are also available to allow the operator to listen more closely. The user can retrieve his/her painting by sliding the “ROBGON” gadget to the center of the light table. The image will be uploaded to and instructions for retrieval will appear on the screen.
LumaTouch uses OpenFrameworks, an open-source C++ library for creative coding. The tracking used by LumaTouch is based on the concepts from TrackMate, an open-source tracking system. The reference paintings are from the Smithsonian Institute Website,, and are used in accordance with the Smithsonian's terms of use. Some of the music loops are original and some are from and are used in accordance with their terms of use. I would like to thank Jennifer Lim for her help with this project.

Masked Thoughts

wood, mirror, computer with custom software, webcam, video projector, foam board, mic. stands
Artwork by [user-name]

Masked Thoughts is an interactive video installation that allows viewers to try on virtual masks, and think virtual thoughts. The installation is comprised of the following components: a large mirror mounted in a wooden frame, a video projector and video camera mounted on top of the frame, and two projection surfaces mounted on mic stands in front of the mirror. One of the projection surfaces is a mask with eyeholes cut out. The other surface is a thought balloon - comic book style. Both surfaces are front/back symmetric and can be swiveled around to face the other side.

A webcam provides a video feed to the CPU, which scans the area and detects changes to the mask and thought balloon. Swiveling the surfaces cause the projected images to change in real-time. Turn the mask around, see a different face. Turn the thought balloon around, see a different thought. Note that the surfaces can be changed at any time, in any order.

A variety of recognizable faces are available to try on: politicians, historical figures, entertainers, etc.


Acrylic modeling resin, foam board, wood, steel, servo motor, micro-controller, custom software, webcam, video projectors
Artwork by [user-name]

FaceLifter is an interactive video installation that allows the viewer to see his/her face projected on a 3d mask. The mask is mechanically raised and lowered to allow the viewer to see him/herself eye-to-eye.
An outline of shoes marks the spot on the floor where the viewer can get a closer look at FaceLifter (and FaceLifter can get a closer look at the viewer). The viewer's faces is illuminated by lights mounted on the column. A web camera mounted above the mask captures the viewer's image, which is algorithmically identified and processed with a hidden computer.
The mask, rendered in white acrylic resin, is mounted on the surface of a column attached to vertical rails. A computer controlled motor inside the column lifts the mask to the height of the viewer.
A video stream of the face is projected via two ceiling-mounted projectors. The projectors are mounted diagonally to allow the viewer to get close to the mask without casting a shadow. The images are adjusted vertically by the computer to track with the mask.
The overall effect of the installation is to allow the viewers to see themselves as they appear to others.
The face finding algorithm is by Philip Abbet, from the IDIAP Research Institute, in Valais, Switzerland.
Thanks to Jennifer Lim, Vivien Leone, and William Tremblay.

User login
Latest Blob Posts
Syndicate content