Researchers have developed a new framework called ReDF to improve the accuracy of land-cover classification using hyperspectral image (HSI) and LiDAR data. This framework quantifies the reliability of features from each data source at specific locations, using this information to correct and fuse the data. The method involves a class prototype bank to estimate feature reliability and a reweighting module to adaptively combine contributions from each modality for final classification. Experiments on three datasets demonstrated that ReDF achieves higher overall accuracy and Kappa scores compared to existing methods. AI
IMPACT This research could improve the accuracy of remote sensing applications by enabling more precise land-cover classification.
RANK_REASON The item is an academic paper detailing a new method for data classification. [lever_c_demoted from research: ic=1 ai=0.7]
- alphaXiv
- arXivLabs
- CatalyzeX Code Finder for Papers
- CORE Recommender
- DagsHub
- Gotit.pub
- Hugging Face
- Hyperspectral Image (HSI)
- Influence Flower
- lidar
- Redfoo
- Reliability-Guided Feature Correction
- Reliability-Guided Feature Reweighting
- ScienceCast
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →