TCCFusion: An infrared and visible image fusion method based on transformer and cross correlation

Wei Tang, Fazhi He, Yu Liu

Abstract: Highlights•We propose an end-to-end infrared and visible image fusion method based on Transformer and cross correlation called TCCFusion.•We present a local-global parallel network to adequately preserve the complementary information.•We design a cross correlation loss to train TCCFusion in an unsupervised manner.•Extensive experiments demonstrate that the proposed method outperforms the state-of-the-art methods.