Paper Overview
Field: Computer Vision (CV) Authors: Yixing Lao, Xuyang Bai, Xiaoyang Wu, Nuoyuan Yan, Zixin Luo, Tian Fang, Jean-Daniel Nahmias, Yanghai Tsin, Shiwei Li, Hengshuang Zhao Published: 2026-03-26 arXiv: 2603.25745 Project page: https://yxlao.github.io/lgtm/
Abstract
Existing feed-forward 3D Gaussian Splatting methods predict pixel-aligned primitives, leading to a quadratic growth in primitive count as resolution increases. This fundamentally limits their scalability, making high-resolution synthesis such as 4K intractable.
We introduce LGTM (Less Gaussians, Texture More), a feed-forward framework that overcomes this resolution scaling barrier. By predicting compact Gaussian primitives coupled with per-primitive textures, LGTM decouples geometric complexity from rendering resolution. This approach enables high-fidelity 4K novel view synthesis without per-scene optimization, a capability previously out of reach for feed-forward methods, all while using significantly fewer Gaussian primitives.
Key Takeaways
- Problem: Pixel-aligned primitives in feed-forward Gaussian Splatting scale quadratically with resolution, blocking 4K synthesis.
- Solution: Compact Gaussian primitives with per-primitive textures decouple geometry from rendering resolution.
- Result: High-fidelity 4K novel view synthesis without per-scene optimization, with far fewer Gaussian primitives.
*Auto-collected on 2026-03-28*