The goal of this project is to develop a Python toolkit to assess urban visual restorative quality from street-level imagery. The toolbox will combine computer vision features with perception-based scores derived from Vision-Language Models, supporting machine learning analysis and interactive visualisation. A case study based on open street-view data will demonstrate the workflow.
Languages: Python
Suggested SW/Libraries: Jupyter Notebook, pandas, NumPy, scikit-learn, GeoPandas, matplotlib
Co-tutor: Prof. Lorenzo Gianquintieri