Search papers, labs, and topics across Lattice.
The paper introduces Geometric Autoencoder (GAE), a novel latent space design for diffusion models that optimizes for semantic discriminability, reconstruction fidelity, and latent compactness. GAE leverages Vision Foundation Model priors to construct a low-dimensional semantic supervision target and employs latent normalization instead of KL-divergence for a stable latent manifold. Experiments on ImageNet-1K demonstrate that GAE achieves state-of-the-art gFID scores (1.82 at 80 epochs, 1.31 at 800 epochs) without Classifier-Free Guidance, outperforming existing methods.
Ditch the heuristic latent spaces: Geometric Autoencoders offer a principled way to inject VFM priors into diffusion models, yielding state-of-the-art image generation with better compression and semantic depth.
Latent diffusion models have established a new state-of-the-art in high-resolution visual generation. Integrating Vision Foundation Model priors improves generative efficiency, yet existing latent designs remain largely heuristic. These approaches often struggle to unify semantic discriminability, reconstruction fidelity, and latent compactness. In this paper, we propose Geometric Autoencoder (GAE), a principled framework that systematically addresses these challenges. By analyzing various alignment paradigms, GAE constructs an optimized low-dimensional semantic supervision target from VFMs to provide guidance for the autoencoder. Furthermore, we leverage latent normalization that replaces the restrictive KL-divergence of standard VAEs, enabling a more stable latent manifold specifically optimized for diffusion learning. To ensure robust reconstruction under high-intensity noise, GAE incorporates a dynamic noise sampling mechanism. Empirically, GAE achieves compelling performance on the ImageNet-1K $256 \times 256$ benchmark, reaching a gFID of 1.82 at only 80 epochs and 1.31 at 800 epochs without Classifier-Free Guidance, significantly surpassing existing state-of-the-art methods. Beyond generative quality, GAE establishes a superior equilibrium between compression, semantic depth and robust reconstruction stability. These results validate our design considerations, offering a promising paradigm for latent diffusion modeling. Code and models are publicly available at https://github.com/sii-research/GAE.
