Theres something captivating about our ability to imagine shapes in clouds.  For my Nature of code midterm I am to create a cloud mirror.  Heres a rough idea of how it will work.

One: An image is take by and the edges of the person are detected using edge-detection in Open CV.  I know this is very fickle at best so I am considering using max/msp’s OpenCV libraries to send coordinates to processing.

Two: These coordinates will then act as constraining boundaries. Using Toxilib, the edge points are then used to generate a particle system filling the boundaries to the shape in the image.

Three: Each particle on the system will use a Cloud node class, which will allow transparency shifts and generation of more smaller clouds to give a fog like effect.

In addition the user can change wind speed and direction, allowing your cloud to linger or blow away.   Each user will also be allowed to save their cloud photo.  I could see an application like this becoming very popular once used in mobile devices.

Another idea of implementation would be to measure the input camera’s the rate of change from each pixel to generate the boundaries.  In a previous project I used the concept in a previous project to effect an audio source.