自由粒度层级视觉识别
Free-Grained Hierarchical Visual Recognition
把视觉语言模型的语义先验和半监督学习引入层级分类，用于处理训练标签粒度不统一的现实场景；其「自由粒度推理」让模型按置信度自适应选择预测深度，这一思路可迁移到任何需要输出粗细可变预测的视觉任务。
原文依据 we introduce free-grained hierarchical recognition, where training labels can appear at any level of a taxonomy
摘要译文
由模型翻译 · 术语统一 · 请对照原文核对关键结论层级图像识别旨在沿语义分类体系预测标签，通常假设每张图像都具备细粒度标注。然而，现实中的监督信号可能出现在任意层级：一张远景鸟类图像可能只被标注为「鸟」，而一张清晰图像则允许标注为「白头海雕」。为反映这一现实，我们提出自由粒度层级识别任务，其中训练标签可以出现在分类体系的任意层级，要求模型在部分监督与混合监督下仍能给出前后一致的预测。我们构建了具有不同标签粒度的基准数据集，并表明现有的层级方法在该设定下性能显著下降。为解决这一问题，我们提出了简单而有效的方法，分别利用 1）来自视觉语言模型的语义引导，以及 2）通过半监督学习获得的视觉结构。最后，我们研究了自由粒度推理，即模型自适应地选择预测深度，从而在细粒度预测不确定时给出可靠的粗粒度预测。我们的任务、数据集与方法共同为面向真实场景的层级识别迈出了实用的一步。
英文原文摘要 1115 字符
Hierarchical image recognition aims to predict labels across a semantic taxonomy, typically assuming fine-grained annotations for every image. However, real-world supervision may appear at any level: a distant bird may only be labeled as "Bird", while a clear image allows "Bald eagle". To reflect this reality, we introduce free-grained hierarchical recognition, where training labels can appear at any level of a taxonomy, requiring consistent predictions under partial and mixed supervision. We construct benchmark datasets with varying label granularity and show that existing hierarchical methods degrade significantly in this setting. To address this, we propose simple yet effective approaches that leverage 1) semantic guidance from vision-language models and 2) visual structure through semi-supervised learning. Finally, we study free-grained inference, where the model adaptively selects prediction depth, enabling reliable coarse predictions when fine-grained ones are uncertain. Together, our task, datasets, and methods provide a practical step toward hierarchical recognition in real-world scenarios.