Alias-Free Latent Diffusion Models:Improving Fractional Shift Equivariance of Diffusion Latent Space
Preprint
Abstract
Latent Diffusion Models (LDMs) are known to have an unstable generation process, where even small perturbations or shifts in the input noise can lead to significantly different outputs. This hinders their applicability in applications requiring consistent results. In this work, we redesign LDMs to enhance consistency by making them shift-equivariant. While introducing anti-aliasing operations can partially improve shift-equivariance, significant aliasing and inconsistency persist due to the unique challenges in LDMs, including 1) aliasing amplification during VAE training and multiple U-Net inferences, and 2) self-attention modules that inherently lack shift-equivariance. To address these issues, we redesign the attention modules to be shift-equivariant and propose an equivariance loss that effectively suppresses the frequency bandwidth of the features in the continuous domain. The resulting alias-free LDM (AF-LDM) achieves strong shift-equivariance and is also robust to irregular warping. Extensive experiments demonstrate that AF-LDM produces significantly more consistent results than vanilla LDM across various applications, including video editing and image-to-image translation. Code is available at: https://github.com/SingleZombie/AFLDMKeywords
All Related Versions
This publication has 6 references indexed in Scilit:
- Improving Shift Invariance in Convolutional Neural Networks with Translation Invariant Polyphase Sampling2025
- DreamMover: Leveraging the Prior of Diffusion Models for Image Interpolation with Large Motion2024
- StableNormal: Reducing Diffusion Variance for Stable and Sharp NormalACM Transactions on Graphics, 2024
- GeoWizard: Unleashing the Diffusion Priors for 3D Geometry Estimation from a Single Image2024
- Exploiting Diffusion Prior for Real-World Image Super-ResolutionInternational Journal of Computer Vision, 2024
- AID: Attention Interpolation of Text-to-Image Diffusion2024