SPAD is a method that synthesizes 3D consistent views of the same object from a text prompt. It can generate many images from arbitrary camera viewpoints and achieves competitive results in terms of PSNR and SSIM.
Table of contents
Method: SPADSort: