Poster Statement

Abstract

This project aims to incite an aesthetic transference between Impressionism and emerging media art practices. How might we design a framework that is able to perceive and capture the characteristics of sceneries to portray its impression, much like Impressionists did? Inspired by the philosophy of Impressionism,this paper explores a way of depicting an impression of the world made by computers using deep neural networks (DNN). First, we built a conceptual framework adopting the insights acquired from Impressionism from a historical context. Then, we reconstruct an original image through the DNN as an approximation of human perception of the world. Our results demonstrated an opportunity for adopting a machine learning approach to generate visualizations for emerging digital media art purposes conceptually as well as practically

• Kyungho Lee, Illinois Informatics Institute University of Illinois at Urbana-Champaign, USA

Full text (PDF) p. 668-669

This research was supported by the Fiddler Innovation Research Fellowship. We wish to thank Jerry Fiddler and Melissa Alden for their generous support. The authors also gratefully acknowledge the use of the services and facilities of the National Center for Supercomputing Applications (NCSA) at the University of Illinois at Urbana-Champaign.