Infrared and visible image fusion using a guiding network to leverage perceptual similarity

Jun-Hyung Kim, Youngbae Hwang

Abstract: Highlights•A new joint loss function that considers visual quality and subsequent tasks.•A hybrid model to benefit from both convolutional neural network and Vision Transformer.•Preserving of visual significant information of both source images.•Applicability to multi-exposure and multi-focus image fusion.•Experimental evaluations of conduciveness to downstream tasks.