An unsupervised multi-focus image fusion method via dual-channel convolutional network and discriminator

Lixing Fang, Xiangxiang Wang, Junli Zhao, Zhenkuan Pan, Hui Li, Yi Li

Abstract: Highlights•A CNN-based end-to-end multi-focus image fusion framework is proposed.•The dual-channel cascade network pays more attention to brightness information.•The patch processing block is designed to generate the reconstructed image.•The dataset conforming to the defocus diffusion effect is synthesized.