Loading...

MaskChanger: A Transformer-Based Model Tailoring Change Detection with Mask Classification

Ebrahimzadeh, M ; Sharif University of Technology | 2024

10 Viewed
  1. Type of Document: Article
  2. DOI: 10.1109/MVIP62238.2024.10491166
  3. Publisher: 2024
  4. Abstract:
  5. Change detection in multi-temporal remote sensing data enables crucial urban analysis and environmental monitoring applications. However, complex factors like illumination variance and occlusion make robust automated change interpretation challenging. We propose MaskChanger - a novel deep learning paradigm tailored for satellite image change detection. Our method adapts the segmentation-specialized Mask2Former architecture by incorporating Siamese networks to extract features separately from bi-temporal images, while retaining the original mask transformer decoder. To our knowledge, this is the first study in which change detection is converted from the existing per-pixel classification approach into a mask classification approach. Evaluated on the LEVIR-CD benchmark of over 600 very high-resolution image pairs exhibiting real-world rural and urban changes, MaskChanger achieves Fl-Score of 91.96%, outperforming prior transformer-based change detection approaches. © 2024 IEEE
  6. Keywords:
  7. Transformer ; Deep learning ; Image classification ; Image segmentation ; Remote sensing ; Classification approach ; Complex factors ; Environmental Monitoring ; Monitoring applications ; Multi-temporal remote sensing ; Remote sensing data ; Remote sensing images ; Urban analysis ; Change detection
  8. Source: Iranian Conference on Machine Vision and Image Processing, MVIP ; 2024 ; 21666776 (ISSN); 979-835035049-4 (ISBN)
  9. URL: https://ieeexplore.ieee.org/document/10491166