Street Rendering

Street Rendering investigates the limits of machine perception by transforming Google Street View images into an imperfect 3D model using Gaussian Splatting. Through a virtual "street photography" conducted from impossible perspectives, the project elevates distortions and reconstruction errors to an artistic language, reflecting on technology's inability to faithfully replicate reality.

360° images of the square in front of the Academy were downloaded from Google Street View and transformed into a 3D model using the Gaussian Splatting technique. The result is a fragmented space where the reconstruction oscillates between sharp details and severe distortions, reflecting the quality and coherence of the available data.

Within this reconstructed world, a series of street photography shots were produced, choosing subjects that highlight the contrast between what is faithfully rendered and what dissolves into error. The virtual camera was positioned at viewpoints impossible to obtain with Street View, expanding the possibilities of observation while simultaneously exaggerating the imperfections of the technology.

The work becomes a reflection on the way machines perceive and reconstruct the world, exposing their limits rather than their capabilities.