Alzette 2.0 Urban AI studies is a series of experiments aiming at exploring artificial intelligence capabilities in proposing alternative natural and urban structures. Using Paperspace and a predefined script which uses AI in style transfer from one image to another the output image can be a new natural or urban structure which can assist in design process and decision making.
This study was applied on a wider renaturing project of the Alzette river in Luxembourg. For the purpose of the project, two sites were selected to explore alternative geographies.

Location A and location B were chosen due to several features. This was the outcome of extensive research into the highest risk areas in the river. As an additional layer of study, a wider area of the entire southern flood plain of Alzette river was also used as a content image.

For the purpose of style transfer, we used a number of images of natural phenomenon. These were images of marshlands, river meanders and forests. As these three features are the focus of the renaturing project.After running hundreds of experiments we chose the most successful images with the most successful natural and urban outcome. These are displayed in this post along with some 3D views using Zbrush for better visualization in the 3D context.

Alzette River 2.0 – Urban AI studies is a project of IAAC, Institute for Advanced Architecture of Catalonia developed at Master in City & Technology in 2020/21 by student: Leyla Saadi, Marta Galdys, Sridhar Subramani, Ivan Reyes Cano  and faculty: Prof. Sandra Manninger & Dr. Matias del Campo