Historically, the evaluation of architectural beauty and urban decorum has relied on subjective assessments, but today the transition of urban planning towards data-driven approaches requires rational evaluations. Within this landscape lies the European Commission’s New European Bauhaus (NEB) initiative, which promotes the integration of Sustainability, Inclusion, and Beauty to address the social and cultural challenges of the green transition. This study focuses on the "Beauty" dimension of the NEB, with the primary objective of bridging the gap between qualitative architectural perception and objective measurement. The intent is to translate philosophical and perceptual concepts, such as aesthetic acceptance and sense of belonging (Genius Loci), into deterministic and computable Key Performance Indicators (KPIs), specifically KPIs B.7, B.10, and B.11 of the NEB framework. To avoid the use of "black-box" models, the research proposes a transparent, five-stage computational pipeline based on Computer Vision techniques. The developed system processes street-level building facade images to extract high-level semantic features, such as chromatic dominance, structural symmetry, and morphological regularity, leveraging Python libraries like OpenCV and algorithms like K-Means. The methodology was tested on a custom dataset of high-resolution images collected in the Lambrate quarter of Milan, specifically pre-processed through perspective rectification and spatial normalization. The extracted data allowed the computation of complex metrics such as Visual Richness (VisR) and the KPIs of the NEB framework. The results demonstrate that the system is capable of quantifying morphological affinity and architectural coherence within a neighborhood. Furthermore, it has demonstrated context awareness capabilities, allowing the inference of the construction chronology of buildings by cross-referencing chromatic and formal data. A rigorous empirical validation, conducted on a contemporary facade in the CityLife district, confirmed the framework’s sensitivity and accuracy in detecting stylistic differences and structural deviations. The true innovative value of this research lies in having transformed a traditional Computer Vision pipeline into a context-aware architecture. Through the application of taxonomies and ontologies based on the NEB guidelines, the system moves beyond the simple extraction of visual features to enable actual automated reasoning about the urban environment. The integration of context awareness within architectural analysis lays solid foundations for future applications in automated urban planning and heritage conservation.
Storicamente, la valutazione della bellezza architettonica e del decoro urbano si è basata su valutazioni soggettive, ma al giorno d’oggi la transizione dell’urbanistica verso approcci orientati ai dati richiede di rendere oggettive tali valutazioni. In questo panorama si inserisce l’iniziativa del New European Bauhaus (NEB) della Commissione Europea, che promuove l’integrazione di Sostenibilità, Inclusione e Bellezza per affrontare le sfide sociali e culturali della transizione ecologica. Questo studio si concentra sulla dimensione della "Bellezza" del NEB, con l’obiettivo primario di colmare il divario tra la percezione architettonica qualitativa e la misurazione oggettiva. L’intento è tradurre concetti filosofici e percettivi, come l’accettazione estetica e il senso di appartenenza (Genius Loci), in Key Performance Indicators (KPIs) deterministici e computabili, nello specifico i KPI B.7, B.10 e B.11 del progetto NEB. Per evitare l’utilizzo di modelli "black-box", la ricerca propone una pipeline computazionale trasparente organizzata in cinque fasi, basata su tecniche di Computer Vision. Il sistema sviluppato elabora immagini di facciate di palazzi a livello stradale per estrarne caratteristiche semantiche di alto livello, ad esempio dominanza cromatica, simmetria strutturale e regolarità morfologica, sfruttando librerie Python come OpenCV e algoritmi come K-Means. La metodologia è stata testata su un dataset personalizzato di immagini ad alta risoluzione raccolte nel quartiere Lambrate di Milano, appositamente pre-processato tramite correzione prospettica e normalizzazione degli spazi. I dati estratti hanno consentito il calcolo di metriche complesse come la Visual Richness e i KPI del framework NEB. I risultati dimostrano che il sistema è in grado di quantificare l’affinità morfologica e la coerenza architettonica all’interno di un quartiere. Inoltre, ha dimostrato capacità di "Context Awareness", permettendo di inferire la cronologia di costruzione degli edifici incrociando i dati cromatici e formali. Una rigorosa validazione empirica, condotta su una facciata contemporanea del distretto di CityLife, ha confermato la sensibilità e l’accuratezza del sistema nel rilevare le differenze stilistiche e le deviazioni strutturali. Il vero valore innovativo di questa ricerca risiede nell’aver trasformato una tradizionale pipeline di Computer Vision in un’architettura consapevole del contesto. Attraverso l’applicazione di tassonomie e ontologie basate sulle linee guida del NEB, il sistema supera la semplice estrazione di feature visive per abilitare un vero e proprio ragionamento automatico sull’ambiente urbano. L’integrazione della consapevolezza del contesto all’interno dell’analisi architettonica pone solide basi per future applicazioni nell’ambito della pianificazione urbana automatizzata e della conservazione del patrimonio.
Multimedia Data Analysis: a pipeline for the New European Bauhaus KPIs
CIUFFREDA, DAVIDE;CIAMMAICHELLA, MARCO
2025/2026
Abstract
Historically, the evaluation of architectural beauty and urban decorum has relied on subjective assessments, but today the transition of urban planning towards data-driven approaches requires rational evaluations. Within this landscape lies the European Commission’s New European Bauhaus (NEB) initiative, which promotes the integration of Sustainability, Inclusion, and Beauty to address the social and cultural challenges of the green transition. This study focuses on the "Beauty" dimension of the NEB, with the primary objective of bridging the gap between qualitative architectural perception and objective measurement. The intent is to translate philosophical and perceptual concepts, such as aesthetic acceptance and sense of belonging (Genius Loci), into deterministic and computable Key Performance Indicators (KPIs), specifically KPIs B.7, B.10, and B.11 of the NEB framework. To avoid the use of "black-box" models, the research proposes a transparent, five-stage computational pipeline based on Computer Vision techniques. The developed system processes street-level building facade images to extract high-level semantic features, such as chromatic dominance, structural symmetry, and morphological regularity, leveraging Python libraries like OpenCV and algorithms like K-Means. The methodology was tested on a custom dataset of high-resolution images collected in the Lambrate quarter of Milan, specifically pre-processed through perspective rectification and spatial normalization. The extracted data allowed the computation of complex metrics such as Visual Richness (VisR) and the KPIs of the NEB framework. The results demonstrate that the system is capable of quantifying morphological affinity and architectural coherence within a neighborhood. Furthermore, it has demonstrated context awareness capabilities, allowing the inference of the construction chronology of buildings by cross-referencing chromatic and formal data. A rigorous empirical validation, conducted on a contemporary facade in the CityLife district, confirmed the framework’s sensitivity and accuracy in detecting stylistic differences and structural deviations. The true innovative value of this research lies in having transformed a traditional Computer Vision pipeline into a context-aware architecture. Through the application of taxonomies and ontologies based on the NEB guidelines, the system moves beyond the simple extraction of visual features to enable actual automated reasoning about the urban environment. The integration of context awareness within architectural analysis lays solid foundations for future applications in automated urban planning and heritage conservation.| File | Dimensione | Formato | |
|---|---|---|---|
|
2026_03_Ciammaichella_Ciuffreda_Summary.pdf
accessibile in internet solo dagli utenti autorizzati
Descrizione: testo dell'executive summary
Dimensione
3.72 MB
Formato
Adobe PDF
|
3.72 MB | Adobe PDF | Visualizza/Apri |
|
2026_03_Ciammaichella_Ciuffreda_Thesis.pdf
accessibile in internet solo dagli utenti autorizzati
Descrizione: testo della tesi
Dimensione
49.73 MB
Formato
Adobe PDF
|
49.73 MB | Adobe PDF | Visualizza/Apri |
I documenti in POLITesi sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
https://hdl.handle.net/10589/252710