Virginia Tech
Download file
Download file
2 files

Trained Model for the Semantic Segmentation of Structural Material

posted on 2021-10-07, 15:01 authored by Eric Bianchi, Matthew Hebdon

This contains two trained DeepLabV3+ models for the semantic segmentation of structural material. The models all were trained using image sizes of 512x512. The two different models are described in the README.txt file included in the folder. The classes in the model are: [concrete, steel, metal decking and background]. Only primary structural concrete and steel were labeled and detected. Secondary concrete, like retaining walls, and non-structural steel, like pipes, were not labeled. This model was trained using the structural material dataset (DOI: 10.7294/16624648)After training, we were able to receive a F1-score of 94.2%. More details of the training, the results, the dataset, and the code may be referenced from the journal article. The repository of the training and testing code may be accessed using the GitHub repository referenced in the journal article.

If you are using the model in your work, please include both the journal article and the model citation.


National Science Foundation Grant No. IIS-1840044



University Libraries, Virginia Tech


  • English (US)


Virginia Tech