MonoVLM：基于视觉语言模型的单目三维视觉定位
MonoVLM: Monocular 3D Visual Grounding with Vision Language Models
把 GRPO 强化学习引入 VLM 的三阶段渐进训练，让语言模型从二维定位逐步过渡到三维结构理解与完整 3D 边界框回归，为「用 VLM 做几何感知任务」提供了一条可复用的训练范式。
原文依据 uses Group Relative Policy Optimization (GRPO) to gradually teach the model to first localize the described object
摘要译文
由模型翻译 · 术语统一 · 请对照原文核对关键结论视觉语言模型（VLM）在指令遵循与二维视觉理解方面已展现出卓越能力。然而，包括 GPT-5 在内的最优 VLM 在三维感知上仍存在困难，尤其是在单目三维视觉定位这类任务中。尽管专用的纯视觉模型在该领域表现优异，但它们往往缺乏 VLM 固有的丰富语义理解能力。为弥合这一差距，我们提出 MonoVLM，一种全新的三阶段训练框架，使 VLM 能够执行准确的单目三维定位。该方法的核心是一个渐进式训练过程，利用组相对策略优化（GRPO）逐步教会模型先定位所描述的物体，再理解其三维结构，最终准确地估计完整的三维边界框。大量实验表明，MonoVLM 显著优于现有 VLM，甚至超越了专用纯视觉模型的性能。我们通过广泛的对比与消融研究验证了该设计的有效性。
英文原文摘要 1053 字符
Vision-Language Models (VLMs) have demonstrated remarkable capabilities in instruction following and 2D visual understanding. However, state-of-the-art VLMs, including GPT-5, still struggle with 3D perception, particularly in tasks such as monocular 3D visual grounding. While specialized vision-only models excel in this domain, they often lack the rich semantic understanding inherent to VLMs. To bridge this gap, we propose \texttt MonoVLM , a novel triple-stage training framework that enables VLMs to perform accurate monocular 3D grounding. The core of our method is a progressive training process that uses Group Relative Policy Optimization (GRPO) to gradually teach the model to first localize the described object, then understand its 3D structure, and finally estimate the full 3D bounding box accurately. Comprehensive experiments show that \texttt MonoVLM models significantly outperform existing VLMs and even surpass the performance of specialized vision-only models. We validate our design via extensive comparisons and ablation studies.