FCMBench-Video：文档视频智能基准测试
FCMBench-Video: Benchmarking Document Video Intelligence
- AI Lab, Qifu Technology（启赋科技AI实验室）
- College of Future Information Technology, Fudan University（复旦大学未来信息学院）
- School of Future Technology, South China University of Technology（华南理工大学未来技术学院）
- Pazhou Lab（琶洲实验室）
机构由 AI 辅助整理，请以论文原文为准。
AI总结：
本文提出FCMBench-Video基准测试，用于评估文档视频理解中的文档感知、时间定位和证据推理能力，通过真实场景数据验证系统性能，揭示不同任务的敏感性和能力差异。
AI中文摘要：
文档理解是金融信贷审核、开卡和远程验证中的关键能力，其中决策准确性与证据可追溯性均至关重要。与静态文档图像相比，文档视频呈现时间冗余且顺序展开的证据流，需要跨帧证据整合，并保留与真实性敏感和反欺诈审核相关的采集过程提示。我们引入FCMBench-Video，一个用于文档视频智能的基准测试，评估文档感知、时间定位和证据推理能力。为了在隐私合规的前提下获得大规模真实数据，我们组织建设为原子级采集和组合工作流，记录可重用的单文档片段，应用受控降解，并组装长格式多文档视频。FCMBench-Video由495个原子视频组成1,200个长格式视频，配以11,322个专家标注的问题-答案实例，涵盖28种文档类型，时长在20秒至60秒之间，包含5,960个中文和5,362个英文实例。对九个最近的视频-大语言模型（Video-MLLMs）的评估显示，FCMBench-Video在系统和能力上提供了有意义的分离：计数是最敏感于时长的任务，跨文档验证和证据推理探测更高层次的证据整合，而视觉提示注入提供了互补的鲁棒性维度。总体得分分布广泛且近似钟形，表明该基准测试既不饱和也不被琐碎案例主导。这些结果使FCMBench-Video成为可重复使用的基准测试，用于跟踪视频-大语言模型在文档视频理解上的进展，并探测真实性敏感的信贷领域应用中的能力边界。
英文摘要：
Document understanding is a critical capability in financial credit review, onboarding, and remote verification, where both decision accuracy and evidence traceability matter. Compared with static document images, document videos present a temporally redundant and sequentially unfolding evidence stream, require evidence integration across frames, and preserve acquisition-process cues relevant to authenticity-sensitive and anti-fraud review. We introduce FCMBench-Video, a benchmark for document-video intelligence that evaluates document perception, temporal grounding, and evidence-grounded reasoning under realistic capture conditions. For privacy-compliant yet realistic data at scale, we organize construction as an atomic-acquisition and composition workflow that records reusable single-document clips, applies controlled degradations, and assembles long-form multi-document videos with prescribed temporal spans. FCMBench-Video is built from 495 atomic videos composed into 1,200 long-form videos paired with 11,322 expert-annotated question--answer instances, covering 28 document types over 20s--60s duration tiers and 5,960 Chinese / 5,362 English instances. Evaluations on nine recent Video-MLLMs show that FCMBench-Video provides meaningful separation across systems and capabilities: counting is the most duration-sensitive task, Cross-Document Validation and Evidence-Grounded Selection probe higher-level evidence integration, and Visual Prompt Injection provides a complementary robustness dimension. The overall score distribution is broad and approximately bell-shaped, indicating a benchmark that is neither saturated nor dominated by trivial cases. Together, these results position FCMBench-Video as a reproducible benchmark for tracking Video-MLLM progress on document-video understanding and probing capability boundaries in authenticity-sensitive credit-domain applications.
