Rethinking cluster-conditioned diffusion models for label-free image synthesis
This work addresses the challenge of label-free image generation for computer vision applications, but it is incremental as it builds on existing cluster-conditioning methods with empirical optimizations.
The paper tackles the problem of improving image synthesis quality in diffusion models by using cluster assignments as conditioning, showing that optimal cluster-conditioning achieves state-of-the-art performance with FID scores of 1.67 on CIFAR10 and 2.17 on CIFAR100, along with increased training sample efficiency.
Diffusion-based image generation models can enhance image quality when conditioned on ground truth labels. Here, we conduct a comprehensive experimental study on image-level conditioning for diffusion models using cluster assignments. We investigate how individual clustering determinants, such as the number of clusters and the clustering method, impact image synthesis across three different datasets. Given the optimal number of clusters with respect to image synthesis, we show that cluster-conditioning can achieve state-of-the-art performance, with an FID of 1.67 for CIFAR10 and 2.17 for CIFAR100, along with a strong increase in training sample efficiency. We further propose a novel empirical method to estimate an upper bound for the optimal number of clusters. Unlike existing approaches, we find no significant association between clustering performance and the corresponding cluster-conditional FID scores. The code is available at https://github.com/HHU-MMBS/cedm-official-wavc2025.