The Project 'Hallucinating Culture' is an exploration into AI and aims to experiment with Machine Learning to create artificial but maybe inspirational images of architecture. Within the recend years many papers around the realm of Artificial Intelligence were released. The use of Machine Learning to generate images by training and running your own generative adversial network became very popular.
The Project 'Hallucinating Culture' was developed during the Hardware II Seminar in 2021. The aim is the imagining of unbuild architecture with help of Machine Learning. It is divided into three main branches:
The first part will cover the preparation of our dataset. We were exploring the city of Barcelona and takind snapshots of architectural columns.
To create a kind of streamlined process we are proposing some tools for a process-pipeline: Camera footage - GrabCut_Opencv with Computer Vision - Resize and rename with Python
Here the full process of the Preparation of the Dataset.
This part is dedicated to the Machine Learning task itself. We are using the StyleGAN2 Repository of NVIDIA Labs in cmbination with running the training in Google Colab. The cloning of the Github Repository happens inside of the Goolge Colab file into your own Google Drive.
Here the full process of setting up the Machine Learning with Machine Learning Network with the use of StyleGAN2 in Google Colab.
This part of this Repository is covering the building of a semi-holographic display to visualize the generated architectural features in a more appealing way.
Here the full process of the preparation for the Hologram Screen.
Hallucinating Culture // HARDWARE II is a project of IAAC, Institute for Advanced Architecture of Catalonia developed at the Master in Robotics and Advanced Construction in 2020/2021 by:
Students: Hendrik Benz, Alberto Browne, Michael DiCarlo // Faculty: Óscar González // Faculty Assistant: Antoine Jaunard //