InfiniSplat: Implicit Gaussian Decoding for Large-Baseline Monocular View Synthesis
2026-08-03 • Computer Vision and Pattern Recognition
Computer Vision and Pattern Recognition
AI summaryⓘ
The authors developed InfiniSplat, a method that creates 3D scenes from just one picture without needing multiple views or extra tuning for each scene. Unlike previous methods that placed details rigidly based on pixel grids, InfiniSplat smartly positions elements along actual surface shapes by using depth information. This approach helps it keep the 3D scene structure more consistent when viewed from different angles. Tests showed that InfiniSplat works better than earlier single-image methods and can handle complex real-world scenes even without extra training.
3D Gaussian Splattingsingle-image 3D reconstructionpixel-aligned representationsurface-aligned representationgeometry-guided samplingimplicit decoderdepth estimationnovel view synthesiszero-shot generalizationHypersim dataset
Authors
Jiawei Wang, Hao Yu, Yongzhen Hu, Xinyi Yang, Tao Ni, Xin Zhan, Junbo Chen, Xiaowei Zhou, Ruizhen Hu, Sida Peng
Abstract
Single-image feed-forward 3D Gaussian Splatting (3DGS) aims to directly generate a renderable 3D scene representation from one input image, avoiding the cost of multi-view capture and per-scene optimization. However, existing methods are often constrained by a pixel-aligned representation, where Gaussians are predicted from fixed image-grid locations. Such pixel-aligned primitives can produce promising nearby-view renderings, but they remain weakly coupled to underlying scene surfaces and struggle to preserve coherent structures under large viewpoint shifts. We present InfiniSplat, a feed-forward single-image 3DGS framework that moves from a pixel-aligned representation toward a surface-aligned representation. InfiniSplat constructs this representation by first using geometry-guided sampling to place 2D supports according to depth-induced local surface structure, and then applying a query-conditioned implicit decoder to predict Gaussian attributes from the image features queried at these supports.By grounding support locations in geometry while decoupling Gaussian prediction from fixed pixel centers, InfiniSplat produces Gaussian layouts that better follow scene surfaces and reduce scattered primitives caused by grid discretization.Across multiple cross-dataset NVS evaluations, InfiniSplat achieves state-of-the-art performance compared with single-image feed-forward baselines, and demonstrates zero-shot generalization from Hypersim indoor synthetic training to complex open-world scenes.Project page: https://zju3dv.github.io/InfiniSplat.