Image fusion aims to combine information from various images form a single image that, in theory, incorporates all the essential elements from each of the initial images. The imaging system's limited extent of the field makes it challenging to retrieve all the useful information from a single image. Multi Focus Fusion (MFF-GAN), a generative adversarial network, is used in digital photography to merge images with very distinct focal points in order to reduce the Defocus Spread Effect (DSE) by creating focus maps where the foreground is appropriately greater than the related items. According to the differentiation of repeating blur, this framework introduces a flexible selection phase to determine if original cells concentrate or not. By retrieving and recreating information, our technology enables multi-focus picture fusion, which almost eliminates blurring and feature damage at the boundary. The present approaches that make use of precise concentrated pictures are known as deep learning techniques. Several programs, including Multi Focus Image Fusion, use deep learning.