[Submitted on 10 Sep 2026 (v1), last revised 29 Oct 2026 (this version, v3)]
Abstract:Non-rigid point cloud registration depends on accurate correspondences under deformation and partial visibility. Existing coarse-to-fine methods use coarse matches to select regions for fine matching, but incorrect selection can remove valid correspondences before refinement. We introduce paper, a conditional transport framework that connects coarse correspondence estimation to high-resolution matching. A diffusion model first estimates a coarse matching matrix. We lift the complete estimate to a fine-level source matrix and refine its scores with a geometry-conditioned bridge. We consider two bridge formulations: a deterministic endpoint-parameterized flow and a stochastic paired Brownian bridge with endpoint and noise prediction. At each step, the current matrix provides an alignment for the next prediction. Both models reuse backbone features throughout inference. Experiments on 4DMatch and 4DLoMatch show improved correspondence and non-rigid registration accuracy. The models also transfer to CAPE and DeepDeform without target-domain training. Ablations examine the roles of candidate coverage, source initialization, and refinement budget.
Submission history
From: Qianliang Wu [view email]
[v1] Thu, 10 Sep 2026 12:39:58 UTC (2,493 KB)
[v2] Tue, 29 Sep 2026 02:51:20 UTC (2,492 KB)
[v3] Thu, 29 Oct 2026 09:54:52 UTC (2,608 KB)
