Synapse
⌘+K
Synapse
PulseExploreClubsResearchersJournals
Instagram
HomeClubsExplore
September 21, 2025˜The œinternational archives of the photogrammetry, remote sensing and spatial information sciences/International archives of the photogrammetry, remote sensing and spatial information sciencesOpen Access

VLM-Based Building Change Detection with CNN-Transformer

View Full Paper
Ask AI
Bookmark
Share

Authors

ZGZeinab GharibbafghiPRPeter Reinartz

Discussion

Loading...

Member takes

Overview

Hybrid framework enhances building change detection in satellite images, improving Recall and F1-Score.

Key Points

  • The proposed framework improves Recall by +3.98%, enhancing building change detection performance.
  • Using a vision-language model, the study generates semantic building masks without fine-tuning.
  • A lightweight CNN-Transformer architecture captures local and global context for more accurate detection.
  • Results on the LEVIR-CD dataset demonstrate significant performance enhancements over traditional baselines.

Cite This Study

Gharibbafghi et al. (2025) studied this question.

synapsesocial.com/papers/68d46fd431b076d99fa6a26fhttps://doi.org/10.5194/isprs-archives-xlviii-4-w16-2025-39-2025
View Full Paper
Ask AI
Bookmark
Share

Also Consider

Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context:

  1. 1The Application of Transformer Model in Building Information Modeling2025
  2. 2Enhanced Change Detection Method in Historical Districts: A Lightweight Visual Transformer Integration Model with Context-Aware Local Feature Augmentation2025
  3. 3Enhanced Change Detection Method in Historical Districts: A Lightweight Visual Transformer Integration Model with Context-Aware Local Feature Augmentation2025
  4. 4Cross-level and multiscale CNN-Transformer network for automatic building extraction from remote sensing imagery2024 · 11 citations
  5. 5A Transformer-Based Multi-Scale Semantic Extraction Change Detection Network for Building Change Application2025