FUFusion: Fuzzy Sets Theory for Infrared and Visible Image Fusion

Infrared and visible image fusion combines the high resolution, rich structure of visible images with the remarkable information of infrared images for a wide range of applications in tasks such as target detection and segmentation. However, the representation and retention of significant and non-significant structures in fused images is still a big challenge. To address this issue, we proposed a novel infrared and visible image fusion approach based on the theory of fuzzy sets. First, we proposed a novel filter that integrates the SV-bitonic filter into a least squares model. By leveraging both global and local image features, this new filter achieved edge preservation and smoothness while effectively decomposing the image into structure and base layers. Moreover, for the fusion of the structure layers, according to the salient structural characteristics, we proposed a feature extraction method based on fuzzy inference system. Additionally, the intuitionistic fuzzy sets similarity measure was utilized to extract details from the residual structure layer. Extensive experiments demonstrate that the proposed model outperforms the state-of-the-art methods on publicly available datasets. The code is available at https://github.com/JEI981214/FUFusion_PRCV .