Geometry conditioned method speeds up multi view image generation
NAMVIS: Next-Scale Autoregressive Multi-View Image Synthesis
Computer Vision and Pattern Recognition
Summary
Generating pictures of an object from a few photos taken at different angles is important but often slow when using certain AI techniques. The paper introduces NAMVIS, a new way to create these images faster without using the slow step-by-step cleaning process. This method uses the object's camera positions to guide making images from coarse to fine details quickly and accurately. The authors found NAMVIS produces better quality images more than three times faster than existing methods.
What this means in practice
- •For 3d content creators: Generate new views of 3D objects quickly and accurately from sparse input images without slow iterative methods.
- •For augmented reality developers: Create consistent multi-view imagery for AR scenes using geometry-aware models that improve speed and quality.
Authors
Ramil Khafizov, Ilya Statsenko, Ruslan Rakhimov, Artem Komarichev, Peter Wonka, Evgeny Burnaev
Abstract
Sparse-view novel view synthesis is a central problem in 3D content creation, but diffusion-based approaches remain limited by iterative denoising, making multi-view generation expensive at inference time. We introduce NAMVIS, a diffusion-free framework that reformulates multi-view image synthesis as geometry-conditioned next-scale autoregression. Instead of generating target views through repeated denoising, NAMVIS predicts discrete visual tokens through a small number of coarse-to-fine scale steps, while sampling all tokens within each scale and across target views in parallel. To anchor this generation process to explicit camera geometry, we propose Multi-scale Projective Pose Encoding, which injects source and target camera transformations into both target-view self-attention and source-to-target cross-attention at every resolution. NAMVIS further combines global conditioning with dense geometry-aware cross-attention, enabling the model to preserve source-view appearance while maintaining target-view consistency. Across Objaverse, GSO, and OmniObject3D, NAMVIS outperforms diffusion-based baselines in PSNR, SSIM, and LPIPS, while running over 3 times faster than the evaluated diffusion baselines under the same evaluation setting. These results suggest that geometry-conditioned next-scale autoregression is a promising and efficient alternative to diffusion for sparse-view multi-view synthesis. Additional qualitative results, videos, and resources are available at https://corl-team.github.io/namvis/