We address fine-grained visual reasoning in multimodal large language models (MLLMs), where key evidence may reside in tiny objects, cluttered regions, or subtle markings that are lost under a single global image encoding. We introduce TikArt (Thinking Aperture), an aperture-guided agent that casts multi-step vision-language reasoning as a decision process over regions of interest. TikArt follows a Think-Aperture-Observe loop, alternating between language generation and two aperture actions: Zoom extracts rectangular crops, while Segment invokes SAM2 to obtain mask-based crops for irregular targets. After every action, the model must produce an explicit observation, turning local visual cues into persistent linguistic memory. Built on Qwen3-VL-8B, TikArt optimizes its reasoning policy with AGRPO, a GRPO-style reinforcement learning algorithm with a two-stage curriculum: it warms up segmentation actions and then jointly optimizes visual math, fine-grained VQA, and segmentation, using rewards that couple task success with purposeful aperture use. Experiments on V*, HR-Bench-4K/8K, MME-RealWorld-Lite, MMStar, RefCOCO, and ReasonSeg show consistent gains over the backbone and yield interpretable aperture trajectories for high-resolution reasoning.

翻译：我们针对多模态大语言模型（MLLMs）中的细粒度视觉推理问题展开研究，该任务中的关键证据可能存在于微小物体、杂乱区域或细微标记中，而这些信息在单一的全局图像编码下容易丢失。我们提出了TikArt（思维孔径），一种孔径引导的智能体，它将多步视觉-语言推理过程建模为对感兴趣区域的决策过程。TikArt遵循“思维-孔径-观测”循环，交替进行语言生成与两种孔径操作：Zoom提取矩形裁剪区域，而Segment则调用SAM2以获得针对不规则目标的基于掩码的裁剪区域。每次操作后，模型必须生成明确的观测描述，从而将局部视觉线索转化为持久的语言记忆。基于Qwen3-VL-8B构建的TikArt，采用AGRPO（一种GRPO风格的强化学习算法）优化其推理策略，该算法包含两阶段课程学习：首先预热分割操作，随后联合优化视觉数学、细粒度视觉问答（VQA）和分割任务，所使用的奖励函数将任务成功与有目的的孔径使用相结合。在V*、HR-Bench-4K/8K、MME-RealWorld-Lite、MMStar、RefCOCO和ReasonSeg数据集上的实验表明，TikArt相较于骨干模型取得了持续的性能提升，并为高分辨率推理任务生成了可解释的孔径轨迹。