Labeled Cracks in the Wild (LCW) Dataset
Labeled Cracks in the Wild (LCW) is a dataset which comprises of real images taken from Virginia Department of Transportation (VDOT) structural inspection reports. This dataset focuses on cracks in the global scene rather than zoomed-in concrete patch. The cracks for LCW were annotated using the GIMP software (The GIMP Development Team, 2019). The guidelines for the annotations are provided by the authors in the file folder. There are a total of 3,817 finely annotated images. The images were split into training and testing, 90% and 10% respectfully. The images were resized to 512x512 for training and testing the DeeplabV3+ model. The original and resized images are included. After training with the DeeplabV3+ model (DOI: 10.7294/16628707), we were able to correctly identify approximately 40% of the annotated ground truth cracks. More details of the training, the results, the dataset, and the code may be referenced in the journal article. The GitHub repository information may be found in the journal article.
If you are using the dataset in your work, please include both the journal article and the dataset citation.
National Science Foundation Grant No. IIS-1840044
PublisherUniversity Libraries, Virginia Tech
- English (US)