Strolling Cities: AI-Generated Italian Urban Landscapes
Strolling Cities, directed by Mauro Martino (founder of the Visual Artificial Intelligence Lab at MIT-IBM Watson AI Lab), is a project that uses generative AI to create real-time urban landscapes in the style of specific Italian cities. Developed in collaboration with students from Politecnico di Milano, MindEarth (a company specializing in city mapping via street-level photographic surveys), and researcher Dalila Colucci of Harvard University, the project employs a Voice-to-City AI system. The AI understands spoken words and generates corresponding cityscapes, drawing from over two million images collected by Politecnico students who walked empty streets during lockdown with cameras. The software produces scenes that evoke the atmosphere of a particular city, akin to how readers mentally construct settings from a book's descriptions. The project is showcased at www.strollingcities.com.
Key facts
- Directed by Mauro Martino, founder of the Visual Artificial Intelligence Lab at MIT-IBM Watson AI Lab
- Collaborators include Politecnico di Milano students, MindEarth, and Harvard researcher Dalila Colucci
- Uses generative AI Voice-to-City system to create real-time urban landscapes
- Based on over two million images collected during lockdown by students walking empty streets
- AI generates cityscapes in the style of a particular city based on spoken words
- Project website: www.strollingcities.com
- MindEarth specializes in city mapping via street-level photographic surveys
- Images were captured during lockdown when streets were empty
Entities
Artists
- Mauro Martino
- Dalila Colucci
Institutions
- Visual Artificial Intelligence Lab
- MIT-IBM Watson AI Lab
- Politecnico di Milano
- MindEarth
- Harvard University
Locations
- Italy