Gather the Forest
An Interactive Art Installation Aimed at Starting Conversation on Sustainable Technology Development
Gather the Forest is an interactive art installation created collaboratively by Stuart Romm, Amy Landesberg, Jordan Graves, and myself for the Extension of Community exhibition in Atlanta. The installation combines generative imagery, JavaScript, a Node.js server, and geolocation to enable visitors to plant virtual trees on an LED wall using their phones. Moreover, it offers them the opportunity to plant real trees through an organization called Trees Atlanta.
The project emphasizes the parallel dynamics of human communities and forest ecosystems, with the aim of bringing people together physically to take action towards building a more sustainable technology-driven world. The installation takes place at the Georgia Tech Media Bridge, a site which initiates a bridge between diverse positions and viewpoints. Our intent for Gather the Forest was to achieve working interaction between passers-by, to reward curiosity and exploration, to invite mutual introductions, and to do so through the example of the beneficial interdependency of a forest community. More information on Gather the Forest can be found on the project site here and the artist talk is available here.
Process
3D Modeling
The forest scene was initially designed using 3d modeling software to create the desired perspective and layout.
3D Models in 3 Tree Species
Perimeter Layout using 3D Modeling
Stable Diffusion Image-to-Image
Tree models were exported individually and fed into stable diffusion using image-to-image with canny edge detection. This process allowed for adding a degree of realism unachievable through standard modeling processes, uniquely to each tree.
Stable Diffusion UI
Perimeter Layout Post Stable Diffusion Image-to-Image
Interactive Javascript Sketch and MaxMSP Audio
The installation was written in JavaScript and runs in a web browser. The code removed backgrounds from the isolated AI tree assets and re-composited them with varying opacities. The installation used a Node.js server to handle communication between viewer’s devices and the installation display. To encourage gathering on site, geolocation was used to only allow participation when near the bridge. Visitors could scan the QR code on the display to interact with the installation. Lastly, a generative soundscape was produced in MaxMSP, providing an ever-evolving sonic backdrop and audio effects paired with user interactions.
