TWINGS: Thin Plate Splines Warp-aligned Initialization for Sparse-View Gaussian Splatting
About
Novel view synthesis from sparse-view inputs poses a significant challenge in 3D computer vision, particularly for achieving high-quality scene reconstructions with limited viewpoints. We introduce TWINGS, a framework that enhances 3D Gaussian Splatting (3DGS) by directly addressing point sparsity. We employ Thin Plate Splines (TPS), a smooth non-rigid deformation model that minimizes bending energy to estimate a globally coherent warp from control-point correspondences, to align backprojected points from estimated depth with triangulated 3D control points, yielding calibrated backprojected points. By sampling these calibrated points near the control points, TWINGS provides a fast and geometrically accurate initialization for 3DGS, ultimately improving structural detail preservation and color fidelity in reconstructed scenes. Extensive experiments on DTU, LLFF, and Mip-NeRF360 demonstrate that TWINGS consistently outperforms existing methods, delivering detailed and accurate reconstructions under sparse-view scenarios.
Related benchmarks
| Task | Dataset | Result | Rank | |
|---|---|---|---|---|
| Novel View Synthesis | LLFF 3-view | PSNR21.49 | 162 | |
| Novel View Synthesis | LLFF 6-view | PSNR24.97 | 129 | |
| Novel View Synthesis | LLFF 9-view | PSNR26.23 | 121 | |
| Novel View Synthesis | DTU 3-view | PSNR21.52 | 75 | |
| Novel View Synthesis | DTU 6-view | PSNR25.81 | 68 | |
| Novel View Synthesis | Mip-NeRF 360 12-view | PSNR20.35 | 53 | |
| Novel View Synthesis | DTU 9-view | PSNR28.22 | 41 | |
| Novel View Synthesis | Mip-NeRF 360 24-view | PSNR24.17 | 38 |