Trained Model for the Semantic Segmentation of Corrosion Condition States
This contains four trained DeepLabV3+ models for the semantic segmentation of corrosion condition states. The models all were trained with a batch size of two, horizontal flip augmentations, and a resnet50 backbone. The models all were trained using image sizes of 512x512. The four models were trained with four different loss functions [cross entropy, L1-loss, L2-loss, and weighted cross-entropy]. The classes for this model are: [good (background), fair, poor, severe]. The classes are the condition states for corrosion. This model was trained using the corrosion dataset (DOI: 10.7294/16624663). After training, we were able to receive a F1 score of 86.67% for the l2 loss function model. More details of the training, the results, the dataset, and the code may be referenced in 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.
PublisherUniversity Libraries, Virginia Tech
- English (US)