ExpAlign：基于期望引导的视觉-语言对齐用于开放词汇接地
ExpAlign: Expectation-Guided Vision-Language Alignment for Open-Vocabulary Grounding
AI总结：
ExpAlign通过理论指导的多实例学习框架，实现开放词汇接地任务中隐式token和实例选择，提升检测和分割性能，尤其在长尾类别上表现突出。
摘要：
开放词汇接地需要在弱监督下实现准确的视觉-语言对齐，但现有方法要么依赖于缺乏细粒度表达能力的全局句子嵌入，要么引入带有显式监督的token级对齐或复杂的跨注意力设计。我们提出了ExpAlign，一个基于原理性多实例学习公式的视觉-语言对齐框架。ExpAlign引入了一个期望对齐头，通过基于注意力的软MIL池化处理token-区域相似性，能够在不额外标注的情况下实现隐式的token和实例选择。为进一步稳定对齐学习，我们开发了一种基于能量的多尺度一致性正则化方案，包括一个Top-K多正样本对比目标和一个从拉格朗日约束自由能最小化中衍生出的几何感知一致性目标。大量实验表明，ExpAlign在开放词汇检测和零样本实例分割方面始终优于其他方法，特别是在长尾类别上。最值得注意的是，它在LVIS minival分割上达到了36.2 AP_r，优于其他最先进的方法，在相同模型规模下，同时保持轻量级和推理高效。
英文摘要：
Open-vocabulary grounding requires accurate vision-language alignment under weak supervision, yet existing methods either rely on global sentence embeddings that lack fine-grained expressiveness or introduce token-level alignment with explicit supervision or heavy cross-attention designs. We propose ExpAlign, a theoretically grounded vision-language alignment framework built on a principled multiple instance learning formulation. ExpAlign introduces an Expectation Alignment Head that performs attention-based soft MIL pooling over token-region similarities, enabling implicit token and instance selection without additional annotations. To further stabilize alignment learning, we develop an energy-based multi-scale consistency regularization scheme, including a Top-K multi-positive contrastive objective and a Geometry-Aware Consistency Objective derived from a Lagrangian-constrained free-energy minimization. Extensive experiments show that ExpAlign consistently improves open-vocabulary detection and zero-shot instance segmentation, particularly on long-tail categories. Most notably, it achieves 36.2 AP$_r$ on the LVIS minival split, outperforming other state-of-the-art methods at comparable model scale, while remaining lightweight and inference-efficient.
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。
