OccSplat is introduced, a feed-forward framework whose core contribution is Occlusion-Aware Gaussian Densification, explicitly targeting missing scene geometry rather than missing pixels, to improve novel-view synthesis while retaining feed-forward inference without per-scene optimization.
Abstract
While recent feed-forward 3D Gaussian Splatting methods enable efficient driving-scene reconstruction from surround-view images, their reliance on a single observation fundamentally hinders novel-view synthesis. Regions occluded in the input can become visible from novel viewpoints, leaving the reconstructed Gaussians under-constrained and incomplete. Moving beyond image-space inpainting for rendered views, we introduce OccSplat, a feed-forward framework whose core contribution is Occlusion-Aware Gaussian Densification, explicitly targeting missing scene geometry rather than missing pixels. By predicting new Gaussian primitives for unsupported regions, OccSplat completes the scene directly in the 3D representation. Specifically, OccSplat first identifies occluded regions that lack Gaussian support through Gaussian Occlusion Mapping. Gaussian Occlusion Completion then generates new primitives for these regions. The predicted primitives are integrated into the original representation to improve novel-view synthesis while retaining feed-forward inference without per-scene optimization. Experiments on the nuScenes benchmark demonstrate that OccSplat improves on the compared methods and yields particularly large gains in occluded regions, leading to more accurate and visually coherent driving-scene reconstruction.
Novel-view synthesis from a single image is a fundamentally ambiguous problem. As the camera moves away from the input viewpoint, previously hidden regions become visible, exposing missing geometry and holes in the reconstructed scene. Existing methods often rely on generative models to complete such regions. However,...
W. Jakubowska, Maciej Zięba, P. Spurek· arXiv.org· 0 citations
A few-shot 3DGS framework that leverages scene-adapted generative priors gated by multi-view geometric consistency to enhance the reconstruction quality for both interpolated and extrapolated views given limited observing viewpoints is proposed.
Hao-Yu Zhang, Shuai-Feng Zhi, Zhen-Hua Du et al.· IEEE Transactions on Visuali...· 0 citations
Reconstructing 3D scenes from sparse views without per-scene optimization remains highly challenging, especially for recovering accurate geometry and fine textures. While recent feedforward approaches leverage generalizable 3D Gaussian Splatting (3DGS) for scene generation, they typically assign one or multiple Gaussia...
Bing He, Jing-Nan Gao, Yu-Nuo Chen et al.· 0 citations
BudSplat is a compact feed-forward 3DGS framework that learns to keep the most valuable Gaussian primitives under a limited primitive count and has two novel designs, including a detail-aware importance score that is extracted from multi-level visual tokens, providing an explicit cue for regions where pruning is likely...
Mu-Yu Xu, Fangneng Zhan, Yu Wei et al.· 0 citations
Spackle is a lightweight residual learning framework that mitigates capacity competition without sacrificing efficiency on large-view-deviation NVS cases and achieves state-of-the-art performance on large-view-deviation cases.
PixGR is an efficient framework that integrates lightweight pixel-space generative modeling into feed-forward 3D Gaussian reconstruction to model scene uncertainty and maintains sub-second inference and remains competitive with substantially heavier pretrained generative models.
Ran-Ran Huang, Wen-Jing Bian, Wei-Xun Luo et al.· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.