DDRICFuse:An Infrared and Visible Image Fusion Network Based on Dual-branch Dense Residual And Infrared Compensation

Benefitting from the strong feature extraction capability of deep learning, infrared and visible image fusion has made a great progress in recent years. In this paper, we implement a dual-branch dense residual infrared and visible image fusion network based on auto-encoder. Specifically, the encoder has two branches that extract the shallow features and deep features of the image, respectively. The fusion layer adopts the residual block to fuse the two sets of features from the same branch of infrared and visible image to get fused features. The decoder is utilized to generate a fused image. To improve the overall performance of the fusion image, an infrared feature compensation network is added that can compensate salient radiation features of the infrared image. Experimental results show that our proposed method achieves reasonable performance compared with other state-of-the-art image fusion methods on structural similarity.