Multi-scale hierarchical feature fusion network for change detection

Hanhong Zheng, Mingyang Zhang, Maoguo Gong, Alex Kai Qin, Tongfei Liu, Fenlong Jiang

Abstract: Highlights•We propose a multi-scale hierarchical feature fusion network for change detection.•We invent a high-frequency enhancement block free from supervised parameters.•We propose a non-parametric attention mechanism, self-weighted attention block.•Comprehensive experiments have been conducted to validate the proposed method.