Compress3D: a Compressed Latent Space for 3D Generation from a Single Image
This addresses the challenge of fast and data-efficient 3D generation for applications in graphics and AI, representing a strong specific gain rather than a foundational breakthrough.
The paper tackles the problem of efficiently generating high-quality 3D assets from a single image by introducing a triplane autoencoder with a 3D-aware cross-attention mechanism and training a diffusion model on the refined latent space, achieving generation in 7 seconds on a single A100 GPU and outperforming state-of-the-art methods with less training data and time.
3D generation has witnessed significant advancements, yet efficiently producing high-quality 3D assets from a single image remains challenging. In this paper, we present a triplane autoencoder, which encodes 3D models into a compact triplane latent space to effectively compress both the 3D geometry and texture information. Within the autoencoder framework, we introduce a 3D-aware cross-attention mechanism, which utilizes low-resolution latent representations to query features from a high-resolution 3D feature volume, thereby enhancing the representation capacity of the latent space. Subsequently, we train a diffusion model on this refined latent space. In contrast to solely relying on image embedding for 3D generation, our proposed method advocates for the simultaneous utilization of both image embedding and shape embedding as conditions. Specifically, the shape embedding is estimated via a diffusion prior model conditioned on the image embedding. Through comprehensive experiments, we demonstrate that our method outperforms state-of-the-art algorithms, achieving superior performance while requiring less training data and time. Our approach enables the generation of high-quality 3D assets in merely 7 seconds on a single A100 GPU.