Scalability in Building Component Data Annotation: Enhancing Facade Material Classification with Synthetic Data
Paper i proceeding, 2024

Computer vision models trained on Google Street View images can create material cadastres. However, current approaches need manually annotated datasets that are difficult to obtain and often have class imbalance. To address these challenges, this paper fine-tuned a Swin Transformer model on a synthetic dataset generated with OpenAI’s DALL E and compared the performance to a similar manually annotated dataset. Although manual annotation remains the gold standard, the synthetic dataset performance demonstrates a reasonable alternative. The findings will ease annotation needed to develop material cadastres, offering architects insights into opportunities for material reuse, thus contributing to the reduction of demolition waste.

Författare

Josie Harrison

Chalmers, Arkitektur och samhällsbyggnadsteknik, Byggnadsteknologi

Alexander Hollberg

Chalmers, Arkitektur och samhällsbyggnadsteknik, Byggnadsteknologi

Yinan Yu

Chalmers, Data- och informationsteknik, Funktionell programmering

Proceedings of the European Conference on Computing in Construction

26841150 (eISSN)

Vol. 2024 120-127
9789083451305 (ISBN)

European Conference on Computing in Construction, EC3 2024
Chania, Greece,

Digitala materialinventeringar för hållbar urban mining

Stiftelsen för Strategisk forskning (SSF) (FFL21-0082), 2022-08-01 -- 2027-12-31.

Ämneskategorier

Samhällsbyggnadsteknik

Datorsystem

DOI

10.35490/EC3.2024.197

Mer information

Senast uppdaterat

2024-09-20