Tile Damage Detection

Published: 24 July 2023| Version 1 | DOI: 10.17632/3t3dk43bv9.1
Contributors:
Narges Karimi,
,

Description

Our dataset comprises 5080 images, categorized into three folders: "train," "valid," and "test." All of these images were captured by contributors in Portugal, showcasing damages on tiles that cover historical buildings in the country. This dataset was compiled in 2023 and is intended for the purpose of detecting damages, such as cracks, craters, tile lacunae, and glaze detachment, using the YOLOv7 model.

Files

Institutions

Universidade do Minho Escola de Engenharia

Categories

Machine Learning, Historical Building, Structural Deterioration, Deep Learning

Licence