图像模型能想象时间吗？ImageTime：通过时空一致性探究视觉世界建模的新基准
Can Image Models Imagine Time? ImageTime: A Novel Benchmark for Probing Visual World Modeling Through Spatiotemporal Consistency
AI总结：
提出ImageTime基准，通过四关键帧协议（初始状态、动作开始、过渡状态、最终状态）评估图像生成模型在时空一致性上的表现，揭示模型在维持连贯视觉世界状态方面的能力与不足。
摘要：
图像生成模型现在能够生成高质量的静态图像，但它们表示视觉世界随时间变化的能力仍然知之甚少。实际工作流程如故事板、逐步插图、参考引导编辑和视频预可视化要求模型在多个视觉状态之间保持身份、对象、空间关系和因果顺序。现有评估主要衡量单图像正确性、组合对齐或视频质量，而未明确图像模型是否能连贯地想象一个时间有序的过程。我们引入ImageTime，一个诊断基准，使用时空一致性作为图像生成中视觉世界建模的行为探针。给定一个动作指令，以及可选地指定初始状态的参考图像，模型必须生成一张包含四个有序关键状态的图像：初始状态、动作开始、过渡状态和最终状态。这个四关键帧协议比单图像生成在时间上要求更高，同时避免了密集视频动态的混淆。ImageTime通过渐进能力层次组织任务，并将每个场景分解为阶段状态谓词、跨帧时间约束和禁止的因果违规。GPT-5.5在结构化的VLM-as-judge协议下对所有生成的图像进行评分，产生可解释的能力分数、诊断子分数和失败标签。通过多家族基准测试，ImageTime揭示了当前图像生成系统在要求随时间维持连贯视觉世界状态时成功、失败和漂移的地方。
英文摘要：
Image generation models now produce high-quality static images, yet their ability to represent how a visual world changes over time remains poorly understood. Practical workflows such as storyboarding, step-by-step illustration, reference-guided editing, and video previsualization require models to preserve identities, objects, spatial relations, and causal order across multiple visual states. Existing evaluations largely measure single-image correctness, compositional alignment, or video quality, leaving open whether an image model can coherently imagine a temporally ordered process. We introduce ImageTime, a diagnostic benchmark that uses spatiotemporal consistency as a behavioral probe of visual world modeling in image generation. Given an action instruction, and optionally a reference image specifying the initial state, a model must generate one image containing four ordered key states: initial state, action onset, transition state, and final state. This four-keyframe protocol is more temporally demanding than single-image generation while avoiding the confounds of dense video dynamics. ImageTime organizes tasks with a progressive capability hierarchy and decomposes each scenario into stage-wise state predicates, cross-frame temporal constraints, and forbidden causal violations. GPT-5.5 scores all generated images under a structured VLM-as-judge protocol, producing interpretable capability scores, diagnostic subscores, and failure labels. Through multi-family benchmarking, ImageTime reveals where current image generation systems succeed, fail, and drift when asked to maintain coherent visual world states over time.
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。
