EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Fine-Tuned Visual Transformer Masked Autoencoder Applied for Anomaly Detection in Satellite Images

Jakub Gajda, Joanna Kwiecień

Applied Sciences · 03.06.2025 · 15 (11) · 6286

DOI: 10.3390/app15116286 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Applied Sciences, online erschienen am 03.06.2025. Verfasst von Jakub Gajda und 1 weiteren Autor:innen. Die Arbeit wurde bisher 3-mal zitiert (Crossref) und nennt 44 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: AGH University of Krakow.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Anomaly detection is a process in which outlier samples can be detected in a given dataset. The purpose of this study is to implement, test, and evaluate the possibility of using deep learning methods for outlier detection with the use of a fine-tuning approach. A Transformer Masked Autoencoder was fine-tuned for a custom satellite image dataset after being pre-trained on the ImageNet subset. The first process of training included building an internal representation of images from a normal class. After adjusting the model weights for this task, a custom dataset with normal and abnormal samples was used for the reconstruction error calculation. The results obtained in this study show that it is possible to distinguish between normal class representatives and outliers using the proposed approach. However, this is not sufficient for the model to be employed in real-life applications. With a given level of precision, the model requires additional knowledge about the subject to correctly classify the sample. To the best of our knowledge, this study is the first to apply ViTMAE for a custom satellite image database. An analysis of the misclassified samples shows that the model tends to generalize the image content and is not sufficiently robust for image noise. As a result of the analysis, a new anomaly indicator is proposed for further study.

Förderung

  • AGH University of Krakow

Zitieren

APA

Gajda, J., & Kwiecień, J. (2025). Fine-Tuned Visual Transformer Masked Autoencoder Applied for Anomaly Detection in Satellite Images. Applied Sciences, 15(11), 6286. https://doi.org/10.3390/app15116286

Vancouver

Gajda J, Kwiecień J. Fine-Tuned Visual Transformer Masked Autoencoder Applied for Anomaly Detection in Satellite Images. Applied Sciences. 2025;15(11):6286. doi:10.3390/app15116286

BibTeX
@article{Gajda2025,
  author = {Gajda, Jakub and Kwiecień, Joanna},
  title = {Fine-Tuned Visual Transformer Masked Autoencoder Applied for Anomaly Detection in Satellite Images},
  journal = {Applied Sciences},
  year = {2025},
  volume = {15},
  number = {11},
  pages = {6286},
  doi = {10.3390/app15116286}
}
Kostenlos starten →