Final Project Schedule

Splash Wall:

Dates:

March

31

• Determine which language we're using - try to implement kinect and Jos Stam's stable fluids in C++.

April

Tue. 5, Thur. 7,

• Naureen: Determine how to output silhouette boundaries from figures from kinect as 640x480 RGBA images. Determine a protocol to pass information between objects.
• Austin: Write fluid simulation to accept collision objects

Tue. 12, Thr. 14

• Austin: Accept image information from kinect stream object
• Naureen: Examine how to determine velocity of feet to pass to simulation
• If fluid processing turns out to be processor restrictive, change kinect code into another project (detect hand prints on wall).

Tue. 19, Thur, 21,

• Reserve projector
• Fine tune distance needed for kinect

Tue. 26, Thur, 28

• Fine tune fluid simulation behaviors
• Optimize speed

May 3rd

• Done just in time!

Final Project Proposal: Co-dependant Flowers

Since the other idea is so complicated, I thought I would come up with a back-up idea quickly. So here you go:

I remember some kind of plant at viz-a-gogo last year that reacted to user's presence, but I couldn't recall if it used the same kind of thought process. If it did, oh well.

Here's something similar using a new thing called muscle wire:

Final Project Proposal: Not a Bang but a Ripple

[Image coming soon]

Aesthetic Goals

Users step up to a body-sized projected image on the wall and stomp their feet to introduce a swirl of blue fluid onto the screen. The fluid flows upward, bouncing off of the sides of the screen, as well as off of silhouettes of the user’s body. The system can accommodate multiple people as well. With every step, more fluid is emitted into the volume. As the user moves their body, they will be able to trap the fluid against the wall, and introduce further velocities.

This will hopefully create a sense of playful serenity and mystery. With the absence of detected movement, color fades over time to leave the wall empty and peaceful.

The faster the fluid moves, the lighter in color it becomes.

Conceptual

The user’s feet sets off ripples that reveal that their own silhouettes are part of the pond. The turbulence created by the user’s silhouettes is designed to help users contemplate the complex chain of events that people set off with each step they take, every move in life. With two or more people, the display becomes a complex intermingling of cause and effect between the actions of two people.

Technical

This is technically nontrivial. It will likely require a Kinect for object recognition and depth perception, as well as a projector with a ceiling or rear projection orientation, depending on installation space. A fluid simulation engine complete with object collision detection will need to be designed and implemented. Object collision will need to be able to be defined based on input images from the Kinect. Everything must also run in real time.

Skills required

- Kinect

o What kind of information does the Kinect driver return for you? Velocity of point movement?

- Determine depth plane

- Determine user proximity to presentation screen

o Handled with depth channel of connect

o Requires an array of proximity sensors otherwise

- Determine user outline.

o Kinect: use depth plane

o Use a difference matte of original image vs image with viewer close to it.

o Determine velocity using optical flow.

- Fluid simulation engine

o Probably best to use Eulerian fluid sim methods. Will need to build/find such an engine for processing.

Kinect?

Sillhouette Mirror: http://www.youtube.com/watch?v=hCBUKC9wc_0

Art Possibilities: http://www.youtube.com/watch?v=Z7ZJpwOBl_A

Kinect/iPad: http://www.youtube.com/watch?v=ocoiov2aot4