Infinite Nature: Perpetual View Generation of Natural Scenes from a Single Image
About
We introduce the problem of perpetual view generation - long-range generation of novel views corresponding to an arbitrarily long camera trajectory given a single image. This is a challenging problem that goes far beyond the capabilities of current view synthesis methods, which quickly degenerate when presented with large camera motions. Methods for video generation also have limited ability to produce long sequences and are often agnostic to scene geometry. We take a hybrid approach that integrates both geometry and image synthesis in an iterative `\emph{render}, \emph{refine} and \emph{repeat}' framework, allowing for long-range generation that cover large distances after hundreds of frames. Our approach can be trained from a set of monocular video sequences. We propose a dataset of aerial footage of coastal scenes, and compare our method with recent view synthesis and conditional video generation baselines, showing that it can generate plausible scenes for much longer time horizons over large camera trajectories compared to existing methods. Project page at https://infinite-nature.github.io/.
Related benchmarks
| Task | Dataset | Result | Rank | |
|---|---|---|---|---|
| Novel View Synthesis | ACID (test) | PSNR15.35 | 18 | |
| Scene Extrapolation | ACID (test) | FID48.27 | 15 | |
| Single-view Novel View Synthesis | DL3DV (Long-term (200th frame)) | PSNR8.98 | 13 | |
| Single-view Novel View Synthesis | RealEstate10K Long-term, 200th frame 84 (test) | PSNR10.07 | 13 | |
| Single-view Novel View Synthesis | RealEstate10K Short-term, 50th frame 84 (test) | PSNR14.12 | 13 | |
| Single-view Novel View Synthesis | DL3DV Short-term (50th frame) | PSNR10.05 | 13 | |
| Unbounded 3D scene generation | Large-scale Internet landscape image dataset 1.0 (test) | CE1.555 | 5 | |
| Novel View Synthesis | ACID (10 generated sequences) | PSNR19.94 | 3 | |
| Scene Extrapolation | ACID | Avg Points Reconstructed1.48e+6 | 3 |