Paper Discussion: Phase-Preserving Diffusion for Structure-Aligned Generation

Paper Discussion: NeuralRemaster: Phase-Preserving Diffusion for Structure-Aligned Generation

Let's look at one of the recent papers on Diffusion based style transfer.

From the Abstract:
"Standard diffusion corrupts data using Gaussian noise whose Fourier coefficients have random magnitudes and random phases. While effective for unconditional or text-to-image generation, corrupting phase components destroys spatial structure, making it ill-suited for tasks requiring geometric consistency, such as re-rendering, simulation enhancement, and image-to-image translation. We introduce Phase-Preserving Diffusion ϕ-PD, a model-agnostic reformulation of the diffusion process that preserves input phase while randomizing magnitude, enabling structure-aligned generation without architectural changes or additional parameters."

----------------------
How to find us:
When you arrive at 44 Sydney Ave, Forrest, you can take the elevator up to the 3rd Floor. You should see the sign for the Trellis Data offices next to the stair case infront of you as you exit. Ring the intercom for someone to let you in.

Special thanks to:
1. Trellis Data for sponsoring this community event.