摘要：食物相似度是检索、推荐、食材替代及其他食品计算任务的基础，但食物表征通常仅通过身份保持（identity preservation）来进行评估。我们在身份识别之外进一步探讨食物相似度：不同的视觉表征如何将食物组织成关系邻域、这些邻域揭示或掩盖了哪些烹饪关系，以及这些结构在不同表征之间出现的一致性如何。印度菜肴提供了一个极具挑战性的实验场景——外观相似的食物可能属于完全不同的烹饪品类，而同一烹饪门类内的食物外观差异可能极为显著。基于扩展的 Khana 数据集以及涵盖卷积神经网络（CNN）、基于 Transformer、自监督学习和图文对齐模型的 7 种预训练视觉表征，我们评估了图像到图像检索以及图像到菜肴的邻域结构，深入考察了烹饪家族结构、对比案例以及跨表征一致性。我们还聚合了图像表征，以比较模型间的菜品级视觉相似度，并与相同菜品基于食谱衍生的相似度进行了对比。结果表明，图像级邻域在不同表征之间存在显著差异，包括出现了跨越烹饪家族边界的相似性。在菜品层面，聚合表征虽展现出局部的变异，但揭示出了共同的视觉结构，不过该结构与基于食谱的相似度关联较弱。这些发现表明，仅凭识别性能无法全面刻画关系结构：食物相似度取决于表征模型、模态和抽象层级，而非食物集合本身固有的静态属性。因此，理解这些结构是将多种食物信息模态整合为统一步调相似度概念的前提条件。

Abstract:Food similarity underlies retrieval, recommendation, substitution, and other food computing tasks, yet food representations are typically evaluated by identity preservation. We ask what happens to food similarity beyond identity: how different visual representations organize foods into relational neighborhoods, what culinary relationships these neighborhoods reveal or obscure, and how consistently these structures emerge across representations. Indian cuisine provides a challenging setting in which visually similar foods may have different culinary identities, while foods within a culinary family can vary substantially in appearance. Using an expanded Khana dataset and seven pretrained visual representations spanning convolutional, transformer-based, self-supervised, and image-text aligned models, we evaluate image-to-image retrieval and image-to-dish neighborhoods, examining culinary-family structure, contrast cases, and cross-representation agreement. We also aggregate image representations to compare dish-level visual similarity across models and against recipe-derived similarity for the same dishes. Image-level neighborhoods vary substantially across representations, including similarities across culinary-family boundaries. At the dish level, aggregation reveals common visual structure despite local variation, but this structure corresponds less strongly with recipe-derived similarity. These findings show that recognition performance alone cannot characterize relational structure: food similarity depends on representation, modality, and abstraction level rather than being a fixed property of the food collection. Understanding these structures is therefore a prerequisite to combining multiple food information modalities into a coherent notion of similarity.