RINO: Rotation-Invariant Non-Rigid Correspondences

TL;DR

RINO introduces an unsupervised, rotation-invariant framework for dense 3D shape correspondence, effectively handling non-rigid deformations, partial data, and complex geometries without pre-alignment or handcrafted features.

Contribution

The paper presents RINONet, a novel feature extractor combining SO(3)-invariant learning with complex functional maps, enabling end-to-end shape matching that outperforms existing methods.

Findings

RINO achieves state-of-the-art results on challenging non-rigid matching tasks.

The method handles arbitrary poses, partiality, and noise effectively.

It unifies rigid and non-rigid shape matching in a data-driven framework.

Abstract

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.