Scott Eaton

Drawing experiment - A GAN (inspired by pix2pixHD) is trained with a custom loss function on a dataset of primitive forms to establish the ‘shape language’ for the network. After training a model, one of the joys of working with neural networks is exploring the latent space for aesthetic possibilities. Here, I play with the network to understand how it reacts to different silhouettes, line weights, and styles.

The final drawing, the “blueprint”, encodes all the information needed to recreate the final sculpture (as long as one has the neural network to uncompress/expand it). The process of developing the final piece involved many iterations of line placement, weigh, and silhouette refinement—effectively drawing the sculpture I wanted. It’s instructive to compare the drawing with the final piece to understand how the neural network interprets the lines as instructions for establishing plane changes.

This step takes the piece from a purely digital creation into the real world through the age-old process of bronze casting. The neural network output was extracted into polygons using standard 3D software, then 3D-printed, molded, cast in wax, and invested in a ceramic shell before the bronze pour shown in the video. The processes used to realize the work compress thousands of years of technological progress from primitive casting and forging to generative adversarial networks trained on state-of-the-art GPUs.

Scott’s ‘Figures’ dataset comprises 30,000+ unique photographs that he shot in the studio from a diverse set of volunteers over a two-year period. The usability of a neural network is often directly related to the quality of the training inputs, so carefully curating these was critical to building AI tools in his artistic practice.

A selection of time lapse videos from his drawing sessions is translated by the “Figures” network—an NVIDIA pix2pixHD, image-to-image translation generative adversarial (GAN) network. The network continually assesses the lines, shapes, and contours of each drawing for body patterns it ‘recognizes,’ then shades and renders them as appropriately as possible.

The drawing process for the master image for Fall of the Damned is underway. The final artwork is 2.2 meters tall, so the preparatory drawing had to be incredibly detailed. The final drawing was too big to fit in GPU memory, so was processed through the neural network in chunks of 8192x4096 before being combined into its final size of 20500x15200 pixels.