Attention Based Semantic Segmentation on UAV Dataset for Natural Disaster Damage Assessment

The detrimental impacts of climate change include stronger and more\ndestructive hurricanes happening all over the world. Identifying different\ndamaged structures of an area including buildings and roads are vital since it\nhelps the rescue team to plan their efforts to minimize the damage caused by a\nnatural disaster. Semantic segmentation helps to identify different parts of an\nimage. We implement a novel self-attention based semantic segmentation model on\na high resolution UAV dataset and attain Mean IoU score of around 88% on the\ntest set. The result inspires to use self-attention schemes in natural disaster\ndamage assessment which will save human lives and reduce economic losses.\n