Efficient Covariance-Aware Vicinal Latent Diffusion for Continuous Conditional Image Generation

⚡ In one sentence AI-generated · check the original

Key findings

• Proposes CAViLD, an efficient latent diffusion framework combining density-adaptive vicinal learning, auxiliary semantic regularization, and NIQE-guided negative data augmentation.
• EMA-teacher internalization enables single-branch conditional inference, removing classifier-free guidance overhead.
• Covariance-aware progressive distillation supports sampling down to four NFEs, evaluated on RC-49, UTKFace, and Steering Angle.

📄 Abstract

Continuous Conditional Generative Modeling (CCGM) aims to synthesize high-dimensional data conditioned on continuously valued scalar labels. However, existing continuous diffusion models face severe data sparsity under imbalanced and long-tailed label distributions, which can lead to visual artifacts and condition inconsistency. Moreover, inference-time classifier-free guidance and dense pixel-space integration introduce substantial sampling overhead, limiting practical deployment. To address these challenges, we propose CAViLD, an efficient latent diffusion framework for continuous conditional generation. CAViLD first constructs a purified condition-aware latent space through density-adaptive vicinal learning, auxiliary semantic regularization, and NIQE-guided negative data augmentation. It then internalizes the conditional guidance direction through an EMA teacher, enabling single-branch conditional inference, and introduces covariance-aware progressive distillation for label-dependent probability-flow ODEs, supporting sampling down to four NFEs. Experiments on RC-49, UTKFace, and Steering Angle demonstrate a favorable trade-off among generation fidelity, continuous condition consistency, and sampling efficiency.