TY - GEN
T1 - Archival Fragments to Immersive Worlds
T2 - 44th Conference on Education and Research in Computer Aided Architectural Design in Europe, eCAADe 2026
AU - Agkathidis, Asterios
AU - Varinlioglu, Guzden
AU - Sopher, Hadas
N1 - Publisher Copyright:
© 2026, Education and research in Computer Aided Architectural Design in Europ. All rights reserved.
PY - 2026
Y1 - 2026
N2 - This paper presents and verifies an AI-enhanced workflow for the digital reconstruction of lost urban heritage in contexts where archival documentation is scarce, often limited to low-quality photographs, historical maps, and fragmented architectural drawings. The workflow systematically addresses challenges of precision and interpretation occurring due to gaps and inconsistencies within such sources while enabling efficient three-dimensional modelling and immersive visualisation. The proposed four-phase workflow begins with archival research and evidence collection, followed by AI-assisted image enhancement and colourisation to complete missing architectural details. Processed imagery is then translated into three-dimensional geometry through image-to-mesh techniques and finally integrated into interactive virtual reality environments, allowing users to immerse into reconstructed urban spaces. The method is tested in a destroyed late 19th-century neighbourhood around the Arch ofGalerius in late 19th-century Thessaloniki's historic centre. Findings demonstrate that, despite challenges of precision and interpretation, the workflow offers a replicable approach for reconstructing lost architectural heritage and advancing immersive digital historiography.
AB - This paper presents and verifies an AI-enhanced workflow for the digital reconstruction of lost urban heritage in contexts where archival documentation is scarce, often limited to low-quality photographs, historical maps, and fragmented architectural drawings. The workflow systematically addresses challenges of precision and interpretation occurring due to gaps and inconsistencies within such sources while enabling efficient three-dimensional modelling and immersive visualisation. The proposed four-phase workflow begins with archival research and evidence collection, followed by AI-assisted image enhancement and colourisation to complete missing architectural details. Processed imagery is then translated into three-dimensional geometry through image-to-mesh techniques and finally integrated into interactive virtual reality environments, allowing users to immerse into reconstructed urban spaces. The method is tested in a destroyed late 19th-century neighbourhood around the Arch ofGalerius in late 19th-century Thessaloniki's historic centre. Findings demonstrate that, despite challenges of precision and interpretation, the workflow offers a replicable approach for reconstructing lost architectural heritage and advancing immersive digital historiography.
KW - AI-assisted Design
KW - Digital Heritage
KW - Digital Reconstruction
KW - Generative AI
KW - Virtual Reality
UR - https://www.scopus.com/pages/publications/105051222093
U2 - 10.52842/conf.ecaade.2026.3.317
DO - 10.52842/conf.ecaade.2026.3.317
M3 - ???researchoutput.researchoutputtypes.contributiontobookanthology.conference???
AN - SCOPUS:105051222093
SN - 9789491207433
T3 - Proceedings of the International Conference on Education and Research in Computer Aided Architectural Design in Europe
SP - 317
EP - 326
BT - Proceedings of the 44th Conference on Education and Research in Computer Aided Architectural Design in Europe, eCAADe 2026
A2 - Damtsas, Efstathios
A2 - Spaeth, A. Benjamin
PB - Education and research in Computer Aided Architectural Design in Europe
Y2 - 7 September 2026 through 11 September 2026
ER -