UniEditBench: 一种统一且成本效益高的图像和视频编辑基准，通过压缩的多模态大语言模型
UniEditBench: A Unified and Cost-Effective Benchmark for Image and Video Editing via Distilled MLLMs
AI总结：
UniEditBench提出统一的图像和视频编辑基准，支持基于重建和指令驱动的方法，通过压缩的多模态大语言模型提供多维评分，减少部署成本并提高评估效率。
摘要：
视觉编辑模型的评估在方法和模态之间仍碎片化。现有基准往往针对特定范式，使得跨范式比较困难，而视频编辑缺乏可靠的评估基准。此外，常用自动指标常与人类偏好不一致，而直接部署大型多模态模型（MLLMs）作为评估者会带来高昂的计算和财务成本。我们提出了UniEditBench，一种统一的图像和视频编辑基准，支持基于重建和指令驱动的方法，采用共享协议。UniEditBench包含九种图像操作（添加、移除、替换、更改、基于笔触的、提取、调整、计数、重新排列）和八种视频操作的结构化分类学，覆盖了具有挑战性的组合任务，如计数和空间重排。为了实现可扩展的评估，我们将高容量的MLLM评估者（Qwen3-VL-235B-A22B Instruct）压缩成轻量级的4B/8B评估者，提供结构忠实度、文本对齐、背景一致性、自然度和时间-空间一致性（针对视频）的多维评分。实验表明，压缩的评估者在与人类判断一致性和相对于教师模型的部署成本显著降低方面表现良好。UniEditBench提供了一个实用且可重复的协议，用于基准现代视觉编辑方法。我们的基准及相关奖励模型在https://github.com/wesar1/UniEditBench上公开可用。
英文摘要：
The evaluation of visual editing models remains fragmented across methods and modalities. Existing benchmarks are often tailored to specific paradigms, making fair cross-paradigm comparisons difficult, while video editing lacks reliable evaluation benchmarks. Furthermore, common automatic metrics often misalign with human preference, yet directly deploying large multimodal models (MLLMs) as evaluators incurs prohibitive computational and financial costs. We present UniEditBench, a unified benchmark for image and video editing that supports reconstruction-based and instruction-driven methods under a shared protocol. UniEditBench includes a structured taxonomy of nine image operations (Add, Remove, Replace, Change, Stroke-based, Extract, Adjust, Count, Reorder) and eight video operations, with coverage of challenging compositional tasks such as counting and spatial reordering. To enable scalable evaluation, we distill a high-capacity MLLM judge (Qwen3-VL-235B-A22B Instruct) into lightweight 4B/8B evaluators that provide multi-dimensional scoring over structural fidelity, text alignment, background consistency, naturalness, and temporal-spatial consistency (for videos). Experiments show that the distilled evaluators maintain strong agreement with human judgments and substantially reduce deployment cost relative to the teacher model. UniEditBench provides a practical and reproducible protocol for benchmarking modern visual editing methods. Our benchmark and the associated reward models are publicly available at https://github.com/wesar1/UniEditBench.
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。
