GradSplat is a feed-forward model that refines its own Gaussians by learned gradient steps inside its forward pass, with no target view, and has the best PSNR and SSIM among nine methods and keeps its accuracy from $6$ to $42$ input views.
Abstract
Feed-forward 3D Gaussian Splatting predicts a scene from a few images, but our predictor is less accurate than its own Gaussians optimized per scene on those images. This gap lies mostly in the shape of the Gaussians, their scale and rotation, since optimizing the shape alone recovers almost all of it and opacity and color recover less. Gradient steps on the input views correct this shape, and the model holds these views at inference. Log-anisotropy correction changes the axis ratios of each Gaussian in log space at constant volume, and in-forward gradient refinement renders the Gaussians into the input views and applies the gradient of the photometric loss with a learned recurrent update, first to opacity and color and then to the shape. GradSplat is a feed-forward model that refines its own Gaussians by learned gradient steps inside its forward pass, with no target view. On both complete six-view RealEstate10K protocols, it has the best PSNR and SSIM among nine methods and keeps its accuracy from $6$ to $42$ input views. Without fine-tuning, it is the most accurate on DL3DV and has the best PSNR and SSIM on ScanNet++. The code will be released at https://github.com/Jingong-Chen/GradSplat/.
DeltaSplat, a lightweight Gaussian refinement module for pose-free feed-forward 3DGS, improves its state-of-the-art backbone by 1.75 dB and surpassing even baselines supplied with ground-truth cameras; consistent gains hold across 6-24 views and all camera regimes.
Chanung Park, Seunghyeon Song, J. Lee et al.· 0 citations
3D Gaussian Splatting has become an effective representation for real-time novel view synthesis by optimizing explicit 3D Gaussian primitives from a set of posed images. A key component of this framework is adaptive density control, where Gaussians are cloned, split, or pruned during training to improve scene coverage...
Javlon Tursunov, G. Rozinaj, Vivek Dwivedi et al.· International Symposium ELMA...· 0 citations
Feed-forward Gaussian models have made novel view synthesis almost instantaneous. Feed-forward methods predict Gaussians directly, either in two stages or in a single merged pass. Using two separately trained models is attractive because each can specialize in its own objective: one for accurate pose estimation and the...
Feed-forward 3D Gaussian Splatting (3DGS) reconstructs a scene in a single forward pass, replacing per-scene optimization with a network trained across many scenes. Its quality, however, degrades sharply as the number of input images drops. The bottleneck is upstream of the reconstruction heads: from a few unposed view...
Minhyeok Lee, Jungho Lee, Minseok Kang et al.· 0 citations
Recent feed-forward 3D Gaussian Splatting (3DGS) methods have enabled effective multi-view 3D reconstruction by predicting Gaussian primitives in a single forward pass. However, existing approaches predominantly adopt a pixel-aligned paradigm that predicts one Gaussian per input pixel, coupling the Gaussian count to th...
Sheng Ye, Jiang-Ke Lin, Fu-Dong Wang et al.· 0 citations
Can a neural network learn to render 3D Gaussian splats on its own, with no rasterizer in the renderer? We present GaussianRenderFormer, a transformer that takes a set of Gaussians and a camera and outputs the image directly. Each Gaussian becomes one token in a view-independent encoder, and a view-dependent decoder tu...
S. Levy, Sagie Benaim· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.