Diff-2-in-1: Bridging Generation and Dense Perception with Diffusion Models

a) LR novel view (b) 3DSR trained on LR images (c) 3DSR trained with 3DSR (ours) Figure 1 . Qualitative comparison on the LLFF dataset with a downsampling factor of ×8 and upsampling of ×4. We train 3DGS using different super-resolved (SR) images and render novel views. The results demonstrate that our method yields fewer artifacts from 3D inconsistencies and better preserves overall structural integrity compared to the baselines.