Distortion studies






What happens when human control meets machine interpretation?

This project began as an exploration of ControlNet-based image generation, using existing visual structures to guide a diffusion model while leaving space for unexpected transformations. I became interested in the tension between control and unpredictability - how a recognizable form can be preserved, distorted, or reinterpreted through the model’s generative process.

Working with bodies, movement, sports, and cultural imagery, I explored how familiar forms could shift between recognition and abstraction while retaining traces of their original structures.

These experiments shaped my ongoing interest in machine interpretation and computational imperfection: how human intention and algorithmic processes negotiate the final image, and what unexpected visual languages can emerge between them.








Fencers, 2024




Monroe, 2024










2024-2026

Invited exhibitions
<Art Alive>, Dead End AI Gallery, NL, 2025
<Ceci n’est pas de l’IA>, Capital-C, NL, 2024


Featured
<Yewonji> / de Volkskrant







©2026 Hyerin Im