VisRL:通过强化推理实现意图驱动的视觉感知
VisRL: Intention-Driven Visual Perception via Reinforced Reasoning
AI总结:
针对意图驱动视觉感知中依赖昂贵区域标注限制可扩展性的问题,提出首个应用强化学习优化视觉推理全过程的框架VisRL,通过试错将关注区域选择作为内部决策,无需标注即超越强基线并实现跨模型泛化。
摘要:
视觉理解本质上是意图驱动的——人类会根据自身目标选择性地关注场景的不同区域。大型多模态模型(LMMs)的最新进展使得通过自然语言灵活表达此类意图成为可能,从而允许查询引导视觉推理过程。诸如Visual Chain-of-Thought之类的框架已经证明了纳入显式推理步骤的益处,即模型在回答查询前先预测一个关注区域。然而,现有方法严重依赖带有标注的中间边界框的监督训练,由于意图-区域对的组合爆炸,这严重限制了可扩展性。为了克服这一局限性,我们提出了VisRL,这是首个将强化学习(RL)应用于意图驱动视觉感知问题的框架。VisRL仅使用奖励信号来优化整个视觉推理过程。通过将中间关注区域的选择视为通过试错优化的内部决策,我们的方法消除了对昂贵区域标注的需求,同时更紧密地契合了人类学习感知世界的方式。在多个基准上的广泛实验表明,VisRL持续优于强基线,证明了其有效性以及在不同LMMs上的强大泛化能力。我们的代码可在https://github.com/zhangquanchen/VisRL获取。
英文摘要:
Visual understanding is inherently intention-driven - humans selectively focus on different regions of a scene based on their goals. Recent advances in large multimodal models (LMMs) enable flexible expression of such intentions through natural language, allowing queries to guide visual reasoning processes. Frameworks like Visual Chain-of-Thought have demonstrated the benefit of incorporating explicit reasoning steps, where the model predicts a focus region before answering a query. However, existing approaches rely heavily on supervised training with annotated intermediate bounding boxes, which severely limits scalability due to the combinatorial explosion of intention-region pairs. To overcome this limitation, we propose VisRL, the first framework that applies reinforcement learning (RL) to the problem of intention-driven visual perception. VisRL optimizes the entire visual reasoning process using only reward signals. By treating intermediate focus selection as an internal decision optimized through trial-and-error, our method eliminates the need for costly region annotations while aligning more closely with how humans learn to perceive the world. Extensive experiments across multiple benchmarks show that VisRL consistently outperforms strong baselines, demonstrating both its effectiveness and its strong generalization across different LMMs. Our code is available at https://github.com/zhangquanchen/VisRL.
机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。