MetaView: Large-Baseline Synthesis from Single Image
Kwai-Kolors · hf · 2026-07-16
This work introduces MetaView, a diffusion-based monocular novel view synthesis framework that renders new views under large viewpoint variations from a single input image.
The core idea combines two elements:
- Implicit geometric modeling: Enhances flexibility, avoiding the constraints of overly heavy explicit reconstruction pipelines
- A few but critical explicit 3D cues:
- A feedforward geometry-aware network provides implicit geometric priors to constrain structural consistency
- Metric depth anchors the generation to real-world scales
The goal is to achieve both geometric consistency and precise controllability. Experiments show that MetaView outperforms existing methods and generalizes better in the challenging scenario of large-baseline single-image synthesis; the code is open-source.
Related event: Kuaishou Unveils MetaView for Single-Image 3D Synthesis(2 posts)→
More from Multimodal
- FastH3-Live hits 22fps: acceleration node benchmarks and the --vram-headroom trick — spartong945 · 2026-09-11
- Midjourney style code share: --sref 2912175708 — tisch_eins · 2026-09-11
- Astra storyboards plus Minimax H3 per-shot generation boost video success rates — Hailuo_AI · 2026-09-11
- MiniMax H3 MAX nails cooking anime clips: 15-second curry demo with prompts shared — Hailuo_AI · 2026-09-11
- MiniMax Music Production Toolkit 2.5 for ComfyUI adds full mastering chain — Vivid_Promise1700 · 2026-09-11
- New Node Finder for ComfyUI ranks fresh nodes by star velocity and recency — Luke2642 · 2026-09-11