arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2602.09809cs.CVcs.AI
收藏

SciFlow-Bench:通过逆解析评估结构感知的科学图表生成

SciFlow-Bench: Evaluating Structure-Aware Scientific Diagram Generation via Inverse Parsing

Peking University(北京大学);Shanghai Jiao Tong University(上海交通大学);Huawei Cloud BU(华为云业务部);Zhongguancun Academy(中关村学院);Beijing Key Laboratory of Data Intelligence and Security (Peking University)(北京数据智能与安全重点实验室(北京大学))

Tong Zhang, Honglin Lin, Zhou Liu, Chong Chen, Wentao Zhang

AI总结:

提出SciFlow-Bench基准,通过逆解析将生成的图表图像转换为结构化图进行比较,以结构可恢复性而非视觉相似性评估科学图表生成。

摘要:

科学图表传达显式的结构信息,然而现代文本到图像模型通常生成视觉上合理但结构错误的结果。现有基准要么依赖图像中心或主观指标,对结构不敏感,要么评估中间符号表示而非最终渲染图像,导致基于像素的图表生成研究不足。我们引入SciFlow-Bench,一个结构优先的基准,用于直接从像素级输出评估科学图表生成。基于真实科学PDF构建,SciFlow-Bench将每个源框架图与规范真值图配对,并在闭环往返协议下将模型作为黑盒图像生成器进行评估,该协议将生成的图表图像逆解析回结构化图以进行比较。该设计通过结构可恢复性而非仅视觉相似性进行强制评估,并由一个协调规划、感知和结构推理的分层多智能体系统实现。实验表明,保持结构正确性仍然是一个基本挑战,特别是对于具有复杂拓扑的图表,强调了结构感知评估的必要性。

英文摘要:

Scientific diagrams convey explicit structural information, yet modern text-to-image models often produce visually plausible but structurally incorrect results. Existing benchmarks either rely on image-centric or subjective metrics insensitive to structure, or evaluate intermediate symbolic representations rather than final rendered images, leaving pixel-based diagram generation underexplored. We introduce SciFlow-Bench, a structure-first benchmark for evaluating scientific diagram generation directly from pixel-level outputs. Built from real scientific PDFs, SciFlow-Bench pairs each source framework figure with a canonical ground-truth graph and evaluates models as black-box image generators under a closed-loop, round-trip protocol that inverse-parses generated diagram images back into structured graphs for comparison. This design enforces evaluation by structural recoverability rather than visual similarity alone, and is enabled by a hierarchical multi-agent system that coordinates planning, perception, and structural reasoning. Experiments show that preserving structural correctness remains a fundamental challenge, particularly for diagrams with complex topology, underscoring the need for structure-aware evaluation.

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。

↑