pipette
ESEspañol

Energy-Conditioned Noise Schedule and Whitening for Spectral Diffusion

Bata Vasic, Bane Vasic

Preprint with a published version

In the authors' words

This paper introduces an energy-adaptive noise scheduling and whitening strategy for transform-domain diffusion models. Existing spectral diffusion methods account for the non-uniform statistics of transform coefficients through coefficient scaling, normalization, or frequency prioritization, while the forward diffusion noise schedule remains largely independent of the underlying spectral-energy distribution. We investigate whether the temporal evolution of the forward diffusion process should also follow the spectral organization of natural images. The proposed formulation combines global spectral whitening with energy-conditioned noise allocation that jointly modulates the injected noise according to the energy of individual transform coefficients and an image-dependent energy path over diffusion time. The resulting forward process preserves Gaussian transitions with closed-form marginals and remains compatible with standard DDPM and DDIM procedures without modifying the diffusion architecture. Experiments on CIFAR-10 demonstrate the contribution of the proposed energy-conditioned noise schedule and spectral whitening, reducing Fr\'echet Inception Distance from 142.48 for a compact DCTdiff U-Net variant to 100.45.

Main resultThe abstract does not state a limitation.

Appeared: Wednesday, October 7. arXiv. Preprint with a published version.

DOI: 10.2139/ssrn.7557215

Published version: 10.2139/ssrn.7557215

Authors' comment: 7 pages, 5 figures, 2 tables, Manuscript submitted for publication in Elsevier Pattern Recognition Letters