Metallic corrosion in urban systems is costly and hazardous. A new method, DAMCN, using dual attention modules, has been developed to detect corrosion regions from digital images collected by unmanned aerial vehicles.
CBAM-keras: Sample code for CBAM module.
DAMCN-V1: Architecture of DAMCN-V1 model.
DAMCN-V2: Architecture of DAMCN-V2 model.
NEA-Dataset: NEA dataset and its annotations.
samples: Running scripts.