SAM3R: Object-Centric 3D Mapping via Foundation-Model-Guided Data Association in Changing Scenes

Map-Det3D: Metric Feed-Forward 3D Reconstruction Prior for Multi-view 3D Object Detection from Streaming Inputs

Map-Det3D is an online multi-view 3D object detection model that brings detection directly into a 3D space reconstructed from RGB, suggesting that training reconstruction priors for detection is a practical route to stable metric 3D detection from monocular video.

D2M-Net: decoupled dual matching for few-shot 3D point-cloud semantic segmentation

MuViSeg: Multi-View Segment Correspondences from Dense Geometry Priors

This work proposes three learned matching heads: a LightGlue-style attention head with DoubleSoftmax scoring on frozen MASt3R descriptors; a DPT-style multi-scale fusion module that exposes layered spatial detail from the VGGT foundation model before pooling; and a multi-view extension that performs joint self-attention over segments drawn from several views at once, recovering transitive correspondences that strictly pairwise matchers cannot reach.

GORI: Image-Guided Selective 3D Object Re-Association for 3D Scene Graphs and Task Planning

3D scene graphs provide structured environmental representations that enable robots to perform language-grounded tasks such as navigation and manipulation. A key challenge in constructing 3D scene graphs is preserving consistent object identity. During robot navigation, newly observed 3D segments need to be associated with previously accumulated instances. Existing methods perform this association using geometric overlap and feature similarity, but these cues alone progressively fragment or merge object instances falsely. To address this challenge, we present GORI, an image-guided selective 3D object re-association framework that combines 2D multi-object tracking with selective 3D re-association. GORI employs 2D temporal tracking as the primary association mechanism and performs 3D re-association selectively to account for tracking discontinuity. To mitigate erroneous merges of spatially adjacent objects, GORI enforces a co-detection constraint that prevents merging 3D instances observed as distinct detections within the single image. The resulting 3D scene graph provides consistent object instances that serve as a grounding space for language-conditioned task planning. We evaluate GORI on HM3DSem dataset over existing 3D scene graph baselines, and assess its performance on real-world indoor scans. We demonstrate improved panoptic quality, F1 score, and average precision on HM3DSem; showing that improved object consistency supports more robust downstream task planning.

Mask-aware tri-modal learning for indoor 3D object detection

A mask-aware tri-modal framework that improves the quality of superpoint representations by retrieving a scene-level structural context from a pretrained PointSAM encoder to enhance object-centric evidence and predicting a soft mask weight to suppress unreliable superpoints.

Geometry-Grounded Unified 3D Perception for Autonomous Driving

Camera-based autonomous driving perception requires a shared representation that preserves metric 3D structure across synchronized multi-camera streams. However, existing image-based frameworks often rely on backbones pretrained for semantic recognition, and introduce 3D geometry through downstream task-specific modules. As a result, their shared representations may fail to preserve explicit metric geometry and consistent 3D scene structure. In this paper, we present a Geometry-grounded Unified 3D Perception (GeoUP) framework that adapts the reconstruction-oriented latent of VGGT to calibrated, streaming multi-camera driving scenes. GeoUP factorizes cross-image interaction into self, temporal, and view attention to capture structurally distinct temporal and cross-view correspondences. It further injects calibration-aware raymap encodings to provide metric scale and camera geometry. The resulting geometry-grounded latent is decoded for metric depth estimation, 3D object detection, and semantic occupancy prediction, corresponding to surface-, instance-, and volume-level readouts of the same 3D scene. Through joint multi-task and multi-dataset training, GeoUP effectively leverages heterogeneous annotations and generalizes across diverse sensor configurations and perception ranges. Extensive experiments on nuScenes, Argoverse 2, Waymo, KITTI, and DDAD demonstrate that GeoUP achieves SOTA performance across detection, occupancy, and depth estimation. These results validate the effectiveness of geometry-grounded representations for unified 3D driving perception.