EPFNet: Edge-Prototype Fusion Network Toward Few-Shot Semantic Segmentation for Aerial Remote-Sensing Images

Few-shot semantic segmentation is a technique that is receiving increasing attention. The aim of this approach is to enable models to segment objects with a few support images (usually 1, 5, 10, etc.). At present, few-shot semantic segmentation has made great progress in the field of natural scene images (NSIs), but these methods cannot be applied directly to the field of remote-sensing images (RSIs). To overcome this challenge, we propose a novel semantic segmentation network structure that integrates prototype information with global edge information to achieve more accurate prototype-matching results. In addition, we design a comprehensive weighted loss function to monitor the training process to help overcome the challenges. Results of the performance comparison with state-of-the-art few-shot semantic segmentation methods demonstrate the superiority of the proposed method.

Automated summary

This study presents a novel semantic segmentation network structure that integrates prototype information with global edge information to achieve more accurate prototype-matching results. In addition, a comprehensive weighted loss function is designed to monitor the training process and help overcome challenges. Results of performance comparison with existing few-shot semantic segmentation methods demonstrate the superiority of the proposed method.