{
  "annotation_id": "402837a0-1a65-4ee0-bcb2-7eaa95073fef",
  "date_created": "2026-02-17T05:53:08.664000Z",
  "date_modified": "2026-02-17T05:53:08.664000Z",
  "file_hash": "f7a82c9d4c959edd8eb45fedf8387c7e3e36ede5ca56ddb40e578f77928233f7",
  "private": false,
  "record": {
    "abstract": "Most existing 3D referring expression segmentation (3DRES) methods rely on dense, high-quality point clouds, while real-world agents such as robots and mobile phones operate with only a few sparse RGB views and strict latency constraints. We introduce Multi-view 3D Referring Expression Segmentation (MV-3DRES), where the model must recover scene structure and segment the referred object directly from sparse multi-view images. Traditional two-stage pipelines, which first reconstruct a point cloud and then perform segmentation, often yield low-quality geometry, produce coarse or degraded target regions, and run slowly. We propose the Multimodal Visual Geometry Grounded Transformer (MVGGT), an efficient end-to-end framework that integrates language information into sparse-view geometric reasoning through a dual-branch design. Training in this setting exposes a critical optimization barrier, termed Foreground Gradient Dilution (FGD), where sparse 3D signals lead to weak supervision. To resolve this, we introduce Per-view No-target Suppression Optimization (PVSO), which provides stronger and more balanced gradients across views, enabling stable and efficient learning. To support consistent evaluation, we build MVRefer, a benchmark that defines standardized settings and metrics for MV-3DRES. Experiments show that MVGGT establishes the first strong baseline and achieves both high accuracy and fast inference, outperforming existing alternatives. Code and models are publicly available at https://mvggt.github.io.",
    "arxiv_id": "2601.06874",
    "authors": [
      "Changli Wu",
      "Haodong Wang",
      "Jiayi Ji",
      "Yutian Yao",
      "Chunsai Du",
      "Jihua Kang",
      "Yanwei Fu",
      "Liujuan Cao"
    ],
    "categories": [
      "cs.CV"
    ],
    "license": "http://creativecommons.org/licenses/by/4.0/",
    "title": "MVGGT: Multimodal Visual Geometry Grounded Transformer for Multiview 3D Referring Expression Segmentation",
    "url": "https://arxiv.org/abs/2601.06874",
    "version": "v1"
  },
  "schema_id": "dorsal/arxiv",
  "source": {
    "execution_id": "0db9afcd-a7a8-4a3f-affe-5a8f1c4384a9",
    "id": "arXiv Dataset",
    "type": "Model",
    "variant": "snapshot-2026-01-17",
    "version": "0.1.0"
  },
  "user_id": 1000002
}