Synthesizing a novel view from different angles is an interest in many fields like VR, video enhancement, games, etc. However, this process has some limitations when given an image of an object from an unrecognizable angle. The proposed method combines image-to-image translation with stable diffusion by giving text as an additional parameter
.png&w=256&q=75)
.png&w=640&q=75)