Regions of Interest

BACKGROUND The medium of collage has drawn critical attention as art practice since early 20th Century movements of Surrealism and Dadaism (Gavin, 2020). Collage techniques foster juxtaposition and melange of non-sequential imagery, connecting photographic and graphic gestures across time and place, defying notions of scale and logic. Contemporary collage continues to serve as a valuable tool in working critically with emerging technologies such as digital image manipulation (Rosler, 1989). Computer vision, machine learning and AI processes offer opportunity for collage; to use large datasets of existing images, to imagine new worlds, to unlock alternate realities. CONTRIBUTION This research comprises the generation of original collage works from public domain and found images using our custom-made automated-art-system based on artificial Intelligence. Rather than following a GANS pathway, where each pixel of the new image is invented, this project makes collages of existing images through algorithmic addition and subtraction based on computer vision and sentiment analysis. Exploring notions of chance, authorship, and the changing relationships to the image, the project expands on the ground-breaking spatial organisation of historical genres of collage. This automated-art-system and its custom-code is designed by Simionato and Donnachie. SIGNIFICANCE The Regions of Interest project premiered with large-scale installations of the triptych “Delayed Rays of a Star”, in the City of Viborg (DK) as part of the inaugural UNESCO sponsored LESS Festival of Contemporary Collage, 2020, curated by Faye Dowling, after a competitive curatorial review process. Press reviews include Elephant art, and Kolaj Magazine. Other works from this research were later exhibited in a juried showcase for NeurIPS aiartonline.com (2020). The research, published by Atomic Activity Books, was awarded “Excellent Work” by the Tokyo Type Directors Club and featured in their annual book (2021).