How to recover the full image from randomly noisy patches instead of model-noised patches?

IIT-PAVIS / DiffAssemble

Official repository for "DiffAssemble: A Unified Graph-Diffusion Model for 2D and 3D Reassembly" accepted at CVPR2024

49 stars 4 forks source link

How to recover the full image from randomly noisy patches instead of model-noised patches? #5

Open eezaynah opened 3 weeks ago

eezaynah commented 3 weeks ago

Dear author, Thank you for your great work, but when I randomly put the scrambled image patches into the model for reasoning, I cannot achieve the recovery effect of the noise added to the image patches by the model. Is there any rule in the position and rotation noise added during training?