Pseudo-color infrared and visible image fusion based on attention-dense network

Abstract In the existing infrared and visible image fusion algorithms, the texture details of the fused image are not clear, and the display of infrared information and texture details is unbalanced. In this paper, we come up with an image fusion method of pseudo-color infrared and visible images based on attention-dense network. Firstly, the gray infrared image is processed by pseudo-color, and then then it is fused with the color visible image. Secondly, during the training process, a generator network structure composed of convolutional layers and dense connected blocks with attention modules is designed. It focuses on the key information of source images and enhances the ability of the network to obtain the information of the source image. Finally, the content loss function is constructed by using infrared pixels, visible pixels, visible gradient and infrared gradient to keep the stability of infrared target and texture details in the fused image. The comparison experiments with five fusion methods are carried out. They show that the proposed fusion method is significantly improved compared with other fusion methods.