SpatialBlock: Enhancing Spatial Intelligence in LVLMs via Synthetic Block-Stacking Problem

  • 类型：arxiv
  • 标识：2609.07064
  • 链接：https://arxiv.org/abs/2609.07064
  • 主分类：multimodal
  • 形态：method
  • TLDR：Large Vision-Language Models (LVLMs) have achieved strong performance on diverse visual tasks, yet their ability to reconstruct and reason about the 3D structure of the scene depicted in 2D images -- referred to as spatial intelligence -- remains limited. Existing approaches attempt to address this gap by using real-scene spatial question answering datasets that require dense geometric annotations. However, constructing such labels is costly, time-consuming, and often noisy due to reliance on external perception modules. In this work, we propose a novel paradigm inspired by human cognitive dev
  • 待LLM分类：否
  • 标题中文：SpatialBlock：通过合成积木堆叠问题增强 LVLM 的空间智能
  • TLDR中文：大视觉语言模型（LVLM）已在多种视觉任务上取得强劲表现，但其重建并推理二维图像中场景三维结构的能力（即空间智能）仍十分有限。现有方法试图通过使用需要密集几何标注的真实场景空间问答数据集来弥补这一差距。然而，此类标签的构建成本高、耗时长，且因依赖外部感知模块而常常带有噪声。本工作中，我们提出一种受人类认知发展启发的新范式……
  • 来源文件：
  • /inbox/tom/_candidates/2026-09-11-agent-rag-longcontext-candidates.json