Researchers have developed DMFNet, a novel dual-backbone multiscale fusion network designed for urban scene classification in remote sensing imagery. This framework addresses challenges in capturing complex feature interactions and learning robust representations by utilizing two pretrained backbones for diverse feature extraction. A multiscale fusion mechanism with residual feature propagation and a spatial attention module are incorporated to enhance feature interaction and highlight informative regions. Experiments on the AID dataset show DMFNet achieving 97.46% average accuracy, with ablative studies confirming the effectiveness of its components. AI
IMPACT Introduces a new deep learning architecture for improved remote sensing scene classification.
RANK_REASON This is a research paper detailing a new model and its performance on a benchmark dataset. [lever_c_demoted from research: ic=1 ai=1.0]
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →
