Patina

The problem

#

Italy has over 4,500 historic city centres and millions of protected buildings. Municipalities, soprintendenze, and conservation consultants regularly need to assess how a facade will age — but today this is done through manual inspection, expensive photogrammetry surveys, and expert intuition. There is no scalable tool that can predict and visualise material degradation across a city portfolio.

The idea

#

Patina is a decision-support platform that lets conservation professionals upload a photo or 3D scan of a building, select a material type and climate scenario, and receive a time-lapse visualisation of how the surface will age over 10, 25, or 50 years — along with a degradation risk score and recommended intervention timeline.

Technical stack

#

The system is built on three layers:

Perception — V-JEPA (Meta’s video joint-embedding predictive architecture) for learning latent representations of building surface textures and their temporal evolution from training data of aged facades.

Rendering — Gaussian Splatting for photorealistic 3D reconstruction from 2D inputs, enabling novel-view synthesis of the aged building from any angle.

Geospatial context — CesiumJS for placing the building in its urban context, overlaying degradation heatmaps, and exporting reports for GIS-integrated workflows.

Business model

#

Patina targets two customer segments:

• Municipalities (comuni with heritage districts): SaaS subscription for portfolio-level monitoring and report generation for PNRR-funded restoration projects
• Conservation consultants: per-project licensing for detailed assessments presented to soprintendenze

EU funding pathways under exploration include Horizon Europe EIC Pathfinder (deep tech) and PNRR Cultura (Italian cultural heritage digitisation).

Status

#

Proof-of-concept stage. Currently building the training dataset from publicly available facade photography of Italian historic centres, and prototyping the V-JEPA fine-tuning pipeline.