Relational representation learning network for cross-spectral image patch matching

Open Webpage

Chuang Yu, Yunpeng Liu, Jinmiao Zhao, Dou Quan, Zelin Shi, Xiangyu Yue

Published: 2026, Last Modified: 05 Nov 2025Inf. Fusion 2026EveryoneRevisionsBibTeXCC BY-SA 4.0
Abstract: Highlights•An innovative relational representation learning is proposed, which breaks the bottleneck of feature relation extraction in existing methods.•To further fully mine individual intrinsic features, a lightweight MGLA module is proposed and an efficient ALFE network is built.•A MLPP optimization strategy is proposed that can promote network optimization without introducing additional parameters or inference time.•Extensive experiments show that our RRL-Net achieves SOTA performance on multiple public cross-spectral datasets.