UrbanCDNet, a custom Siamese CNN, raises F1 to 0.7511 and IoU to 0.6014 on a Korean urban change-detection benchmark, with largest gains on sparse-change and high-photometric-gap subsets.
TINYCD: A (not so) deep learning model for change detection,
2 Pith papers cite this work. Polarity classification is still indexing.
2
Pith papers citing it
fields
cs.CV 2years
2026 2verdicts
UNVERDICTED 2representative citing papers
CoSA adds correlation-guided change gating with learnable residuals at multiple decoder scales to Siamese networks, yielding 1.5-2.6% F1 gains on four remote sensing change detection benchmarks.
citing papers explorer
-
UrbanCDNet: Appearance-Robust and Boundary-Aware Bitemporal Change Detection for Korean Urban Building Monitoring
UrbanCDNet, a custom Siamese CNN, raises F1 to 0.7511 and IoU to 0.6014 on a Korean urban change-detection benchmark, with largest gains on sparse-change and high-photometric-gap subsets.
-
CoSA: Correlation-Guided Change Attention with Learnable Residual Gating for Remote Sensing Change Detection
CoSA adds correlation-guided change gating with learnable residuals at multiple decoder scales to Siamese networks, yielding 1.5-2.6% F1 gains on four remote sensing change detection benchmarks.