Residual Forms combines LiDAR, AI, and robotics to generate and paint depth-based patterns on sculptures. A LiDAR camera captures an object’s topography, and a 6-axis robot arm translates this data into intricate, site-specific designs.

The process begins with a plinth, created by transforming a human-like image into a 3D model. A computer vision script then maps the height variations of the sculpture, generating patterns and color schemes tailored to its surface. AI-generated images define the silhouette of the painted areas, while visuals from CycleGAN refine the details. The robotic arm, equipped with a brush, applies these patterns in real time with precision.

By automating both form generation and surface detailing, the project explores new ways to integrate generative AI processes with physical fabrication.