Portal to Other Dimensions: use of Computer Vision to create art work from day life images

• Luísa Ferreira UFV http://orcid.org/0009-0002-9587-9603
• Ana Osias UFV https://orcid.org/0009-0004-2562-372X
• Lucas Vieira UFV
• Michel Silva UFV https://orcid.org/0000-0002-2499-9619

Abstract

This work aims to combine classic and contemporary paintings with the external world by altering the interior of a door into another style in real-time photos and/or videos, creating a portal effect to provide an interactive experience with artworks. To achieve this, various techniques from Artificial Intelligence and Computer Vision were employed, primarily focusing on convolutional neural networks (CNNs) with supervised training. Models proposed for semantic segmentation problems were used for door detection and style transfer to stylize the input image. Finally, we conclude with an analysis of the results for each network individually and in combination, ensuring that the metrics and approaches were satisfactory. With our results, we were able to offer an accessible and interactive way to bring art into the daily lives of many people.

Keywords: Component, Formatting, Style, Styling