SegMASt3R: Geometry Grounded Segment Matching
This addresses a challenging computer vision problem for applications like 3D instance mapping and object-relative navigation, with incremental improvements in performance.
The paper tackled wide-baseline segment matching under extreme viewpoint changes by leveraging 3D foundation models, achieving up to 30% improvement in AUPRC over state-of-the-art methods on datasets like ScanNet++ and Replica.
Segment matching is an important intermediate task in computer vision that establishes correspondences between semantically or geometrically coherent regions across images. Unlike keypoint matching, which focuses on localized features, segment matching captures structured regions, offering greater robustness to occlusions, lighting variations, and viewpoint changes. In this paper, we leverage the spatial understanding of 3D foundation models to tackle wide-baseline segment matching, a challenging setting involving extreme viewpoint shifts. We propose an architecture that uses the inductive bias of these 3D foundation models to match segments across image pairs with up to 180 degree view-point change rotation. Extensive experiments show that our approach outperforms state-of-the-art methods, including the SAM2 video propagator and local feature matching methods, by up to 30% on the AUPRC metric, on ScanNet++ and Replica datasets. We further demonstrate benefits of the proposed model on relevant downstream tasks, including 3D instance mapping and object-relative navigation. Project Page: https://segmast3r.github.io/