RINO: Rotation-Invariant Non-Rigid Correspondences
About
Dense 3D shape correspondence remains a central challenge in computer vision and graphics as many deep learning approaches still rely on intermediate geometric features or handcrafted descriptors, limiting their effectiveness under non-isometric deformations, partial data, and non-manifold inputs. To overcome these issues, we introduce RINO, an unsupervised, rotation-invariant dense correspondence framework that effectively unifies rigid and non-rigid shape matching. The core of our method is the novel RINONet, a feature extractor that integrates vector-based SO(3)-invariant learning with orientation-aware complex functional maps to extract robust features directly from raw geometry. This allows for a fully end-to-end, data-driven approach that bypasses the need for shape pre-alignment or handcrafted features. Extensive experiments show unprecedented performance of RINO across challenging non-rigid matching tasks, including arbitrary poses, non-isometry, partiality, non-manifoldness, and noise.
Related benchmarks
| Task | Dataset | Result | Rank | |
|---|---|---|---|---|
| Non-isometric 3D shape matching | SMAL | Mean Geodesic Error4.6 | 58 | |
| Shape Matching | SHREC HOLES 2016 (test) | Average Geodesic Error12.09 | 26 | |
| Shape Matching | Faust | Geodesic Error1.6 | 21 | |
| Shape Matching | SMAL remeshed (test) | Mean Geodesic Error (x100)4.6 | 20 | |
| Non-isometric Shape Matching | DT4D (test) | Mean Geometric Error5.3 | 14 | |
| Raw scan matching | FSCAN (test) | Mean Geometric Error2.5 | 12 | |
| Non-rigid shape matching | SMAL SO(3) | mGeoErr4.6 | 7 | |
| Non-rigid shape matching | SMAL (I/I) | mGeoErr4.6 | 7 | |
| 3D shape matching | FAUST (test) | Geodesic Error (E)1.6 | 6 | |
| 3D shape matching | SCAPE (test) | E Error2 | 6 |