International College

Title: The Method of Semantic Segmentation Improvement using RescueNet based PSPNet

Abstract: Due to climate change increasing rate of natural disasters worldwide can be observed. These catastrophes are causing severe damage to humanity and nature. As a result of the disasters, economic losses are steadily increasing, and the importance of fast and accurate auxiliary tools for rescue teams is growing in saving people's lives and eliminating financial losses. In this regard, by using deep learning algorithms, it is possible to assess the consequences of any disaster that can accurately understand the affected areas. Until now, several successful works have been done by several researchers. The following paper focuses on the experiment to enhance PSPNet model results on RescueNet datasets collected by UAV by modernizing configurations and datasets.