Is there a general adversarial network that can take multiple low-quality images of a subject to create a higher quality image of the subject?
SRGAN just takes a single low res image and makes it high res but I need something that can take multiple low quality (not necessarily low res) images to create a higher quality image.
By low quality, I mean you could have one image with a reflection or partially blurry or has a small obstruction that another image might compensate for. So, it could combine the lower quality images to create a higher quality image.
Is there any GAN that could do this?
