Computer Vision Art Gallery

ICCV 2019, Seoul, Korea

Julian Tapales

Naturata Naturans, (2019)


This work celebrates the life of a symbol, as it is upheld by a group of people in a specific time and place, and explores how its meaning changes when it is deconstructed, synthesized and represented digitally. Placing itself in an ancient tradition of art theory, this work revisits the vexed question: whether the task of art is to represent nature as it is, or to replicate the generative activity that constitutes the essence of nature itself.

My artistic practice explores the relationship between physical materials and code. With this work, I was interested in exploring symbolic frameworks in living culture as well formally in a computational process. The resulting work is the result of a cross disciplinary research process that enabled me to work with musicians and people from the Cordillera region of the Philippines, craftsmen, composers and philosophers.

For this work I used a GAN designed for texture synthesis and trained several instances on varying 2D textures. I took the weights from the resulting trained models and interpolated them to create new models that would generate synthesized textures reminiscent of the original inputs but simultaneously with strange digital properties. To the best of my knowledge, this is a novel technique. The subsequent textures were used to transform the surface of a mesh realtime with WebGL and presented as a projected render.

I have just completed an MA in Information Experience Design at the Royal College of Art. During and prior to this, I have worked as a researcher with Pedro Lopes at the Hasso Platzner Institute, participated in workshops including MIT Media Lab’s Signal and Noise and was invited to participate in discussions with Wayne Mcgregor studio as a creative technologist.